Michael Nalesnik, MD, Anthony J. Demetris, MD, John J. Fung, MD, PhD, Parmjeet Randhawa, MD, Adriana Zeevi, PhD
· Defining Characteristics of PTLD
· Incidence and Risk Factors
o Type of Allograft
o Epstein-Barr Virus Infection
o Differences Between Adults and Children
o Underlying Disease
o Intensity of Immunosuppression
o Immunosuppression Era and Time of Published Data
· Clinical Presentation
o Time to Onset
o Clinical Manifestations
· Making a Clinical Diagnosis
o Histology and Clonality
o Classification Systems
· Treatment Strategies
o Reduction of Immunosuppression
o Systemic Antiviral Therapy
o Cytokine Therapy
o Radiotherapy and Chemotherapy
o Cell Therapy
Posttransplantation lymphoproliferative disorders (PTLD), one of the most serious complications occurring after transplantation, have been recognized as a complication of organ and cell transplantation for more than 30 years. Initial reports of these disorders raised questions that have only been partially answered at present. This article will review the clinical features of PTLD and discuss the underlying pathology of these growths. Current methods of treatment will be considered, and areas of current investigation will be highlighted.
PTLD refers not to a single disease but to a syndrome that includes a wide range of abnormal hyperplastic (inflammatory or reactive) and neoplastic lymphocyte growths, ranging from a benign self-limited form of lymphoproliferation to an aggressive, widely disseminated disease.[2-6] Approximately 90% of these growths are of B-cell origin, and 90% to 95% contain the Epstein-Barr virus (EBV). The outer boundaries in regard to the types of lesions that can be categorized as a form of PTLD are not rigorously defined. However, a consistent group of lesions constitutes the bulk of these abnormal growths, and such lesions have been recognized by all the major classification systems to date[2,4-6] (see Classification Systems" section). Patients with PTLD appear to have different histologic findings, a more aggressive clinical course, less likelihood of responding to conventional treatments for lymphoma, and poorer outcomes when compared with immunocompetent hosts who develop malignant lymphomas. This article will use the unqualified term PTLD to refer to clonal EBV-containing lymphoid tumors. When hyperplasias or EBV-negative lymphoid tumors are discussed, they will be specifically referred to as such.
Posttransplantation lymphoproliferative disorders are different from lymphoproliferative disorders that occur in the general population. Although relatively uncommon, the risk of developing lymphoma after transplantation has been reported to be 28 to 49 times greater than that in the general population. The main repository of data on posttransplantation cancer is the Cincinnati Transplant Tumor Registry (CTTR) which has collected data on more than 6,000 patients. According to the CTTR, PTLD accounts for 16% of cancers in transplant recipients compared with 5% in the general population. However, these data are heavily skewed toward kidney transplant recipients.
Although the incidence of PTLD has been reported to be as high as 65% after primary and 30% after reactivation EBV infection, overall frequency figures range from 1% to 10%. Most estimates are based on relatively small transplant series from individual institutions. Migliazzi and colleagues assessed long-term survival in 198 consecutive pediatric liver transplant recipients. At a median follow-up of 28 months, the incidence of PTLD was 7%. Although the overall incidence of PTLD may appear to be low, in pediatric patients it is the predominant neoplasm encountered. In an analysis of tumors in 512 patients in the CTTR, PTLD comprised 52% of all tumors. There was a disproportionately high incidence of PTLD among nonrenal allograft recipients compared with renal allograft recipients (81% vs 31%) in this group of patients.
The frequency of PTLD differs in response to many variables, including allograft type, presence or absence of EBV infection before the time of transplantation (EBV seropositive or seronegative status), adult vs pediatric population, underlying disease, intensity of immunosuppression, and the year(s) included during the study interval for any particular series of patients.
The pathogenesis of PTLD is multifactorial and complex. Important host factors include impaired immune surveillance. Chronic antigenic stimulation from the allograft may also play a role. Additional risk factors include the specific organ(s) transplanted; the type, intensity, and duration of maintenance immunosuppression; and the occurrence and severity of acute rejection. Epstein-Barr virus infection (primary or reactivation), however, is the main etiologic factor.
Estimates of PTLD frequency in recipients of different types of allografts are as follows: kidney, 1% to 4%; liver, 2%; heart, 2% to 10%; heart and lung, 5% to 9%; and intestine, 19%.[14-16] In the case of bone marrow recipients, the frequency is 1% to 2% excepting cases of mismatched T-cell-depleted allografts, for which the frequency has been historically as high as 24%. Innovations such as removal of B cells from the marrow allograft have reduced and in some series eliminated this complication. Patients who receive allogeneic hematopoietic stem cell transplants also have an approximate 1% risk of developing PTLD.
A significant role of EBV in PTLD is underscored by the increased frequency of PTLD in transplant recipients who are EBV seronegative at the time of operation. These patients will invariably be exposed to EBV at some future posttransplantation date, thereby undergoing their primary infection while in an immunosuppressed state. At present, there is no available anti-EBV vaccine to prevent infection.
Ho and colleagues reported on a series of pediatric kidney transplant recipients. Patients who were EBV seronegative at time of transplantation had a 10% frequency of PTLD compared with patients who were EBV positive at the time of transplantation, who had a 0% frequency of PTLD. At the same institution, PTLD frequencies in adults who were EBV seronegative and EBV seropositive were 4.9% and 1.6%, respectively. Despite these figures, Harwood and colleagues were unable to document any increase in PTLD in a series of EBV-seronegative pediatric thoracic transplant patients who received organs from seropositive donors. They concluded that EBV matching was not justified in this population.
The series of Ho and colleagues also highlights differences in the frequency of PTLD based on patient age at the time of transplantation. It is our experience that pediatric patients have a higher frequency of PTLD in general than do adult patients receiving similar allografts. In one such study from our institution, Shapiro and colleagues reported a 10.1% PTLD frequency in pediatric kidney transplant recipients compared with a 1.2% frequency in the adult renal transplant population. Eighty-six percent of pediatric cases and 50% of adult cases involved a transplant from an EBV-seropositive donor to an EBV-seronegative recipient, contrasting with the experience of Harwood and colleagues. Thus, at least part of the difference in frequency between adults and children may be explained by the higher proportion of EBV-seronegative patients in the pediatric as opposed to the adult population.
Several reports have suggested that underlying disease may represent a risk factor for PTLD. For example, patients who receive a bone marrow transplant (BMT) or hematopoietic stem cell transplant for treatment of underlying immunodeficiency diseases or for chronic myelocytic leukemia have been reported to be at higher risk for PTLD than those who receive a BMT for other reasons. Shpilberg and colleagues suggest that, in liver transplant patients, underlying autoimmune disorders such as autoimmune hepatitis or primary biliary cirrhosis may predispose to PTLD. An even more striking association was reported in one series of patients who underwent liver transplantation for treatment of Langerhans cell histiocytosis. In this group, two thirds of patients developed PTLD. Underlying hepatitis C virus (HCV) infection was also found to be associated with a 10.5% frequency of PTLD in one series, whereas liver transplant for other diseases was associated with a 1.7% frequency. Although patients with HCV were noted to have a higher requirement for immunosuppression with antilymphocyte antibodies, the authors observed that an increased risk remained even after this variable had been accounted for.
PTLD has been documented in three transplant immunosuppression eras: conventional (precyclosporine), cyclosporine, and postcyclosporine. The level of immunosuppression (ie, intensity, type, and amount) is an independent risk factor for PTLD. Swinnen and colleagues first reported an increased frequency of posttransplantation lymphoid tumors in conjunction with the use of monoclonal anti-CD3 antibody (muromonab OKT3) used as an antirejection agent. In that study, a cumulative anti-CD3 dose of greater than 75 mg was associated with a 38% frequency of PTLD, whereas less cumulative doses were associated with a frequency of only 6%. The higher incidence of PTLD in the high-dose OKT3 group was attributed to overimmunosuppression. However, the possibility that a cytokine release syndrome associated with this antibody may play a role has not been ruled out.
Ciancio and colleagues reported on the incidence of PTLD under different immunosuppressant regimens during an 18-year period. They noted a recent increase in the incidence of PTLD with the advent of newer immunosuppressive agents. By contrast, the use of mycophenolate mofetil in a steroid-free immunosuppressive protocol with concomitant acyclovir therapy was associated with a lower incidence of primary and reactivation EBV infection and PTLD.
Interpretation of reports of PTLD frequency must take into account the time the data were published. As has been noted, PTLD has been documented in three transplant immunosuppression eras. In initial reports following introduction of new immunosuppressants, there is a tendency to report relatively high frequency of PTLD. These frequencies usually decline, however, as more patients are treated and additional experience gained, thereby avoiding toxic side effects related to the learning curve associated with the use of new agents.
Other innovations may affect PTLD frequency and may be reflected by differences in reports from different treatment eras. In our experience, PTLD frequency in the intestinal transplant population decreased around the mid-1990s, coinciding with the introduction of a laboratory assay that allowed detection of EBV in the peripheral blood. Also during this time, we introduced the use of supplemental donor bone marrow infusion in this population in an effort to enhance microchimerism. A separate study in liver transplant recipients also noted a trend toward a lower incidence of PTLD in patients who received supplemental donor bone marrow cells. Although the effect of this manipulation on the frequency of PTLD remains unproved at present, this observation does show that the frequency of PTLD is in dynamic flux as newer forms of posttransplantation management are used.
Additional factors of possible correlates of an increased risk for the development of PTLD are under study. These include the phenotype of lymphoid cells at the time of transplantation, cytomegalovirus infection, and the presence of constitutional polymorphisms in cytokine genes.
Posttransplantation lymphoproliferative disorders may present as symptoms from localized or systemic involvement or in the asymptomatic patient incidental to other clinical or radiographic findings. An unexplained infectious syndrome in a transplant recipient should raise the suspicion of a PTLD.
PTLD can present as early as less than a month to as late as several years after transplantation. In general, however, PTLD is remarkable for a short posttransplantation time of onset. The time of onset is shorter in cyclosporine- and tacrolimus-treated patients than in the precyclosporine era. In our experience, approximately 47% of cases occur within 6 months, 62% within 1 year, and 90% within 5 years following transplantation. A small trickle of cases occurs thereafter. The latest case in our series occurred 13 years after liver transplantation, and the earliest bona fide case occurred 3 weeks after transplantation. In the Cincinnati Transplant Tumor Registry of Penn, the latest case occurred 25 years after transplantation.
PTLDs that do not contain EBV tend to arise at a later time than those that do contain the virus. In one series, 50% of EBV-positive PTLDs arose by 6 months following transplantation, whereas the 50% mark for occurrence of EBV-negative PTLDs was not reached until 5 years after transplantation. PTLDs of T-cell origin are uncommon and may also arise later in the posttransplantation course, but a case of a monoclonal T-cell tumor arising 2 months after transplantation has been described. A definitive statement regarding time of onset of this tumor subset awaits the accrual of a larger number of cases.
The clinical presentation of PTLD is heterogeneous but tends to fall within several well-defined, albeit overlapping, forms. These include (1) an infectious mononucleosis-like syndrome with or without generalized lymphadenopathy, (2) one or more nodal or extranodal tumors, and (3) a fulminant and disseminated presentation with sepsis.
A mononucleosis syndrome may occur early after transplantation, particularly in association with a primary EBV infection. This presentation is particularly common in the pediatric population, and indeed, in some cases it is infectious mononucleosis. Otolaryngologic symptoms and findings are often the first manifestation of PTLD in children. Patients may present with tonsillitis, tonsillar necrosis, lymphadenitis, sinusitis, and otitis media. There is a tendency for more severe upper airway symptoms, including airway obstruction. It should also be noted that the underlying process in these cases, ie, infectious mononucleosis vs frank tumorous PTLD, cannot always be inferred from the clinical picture alone.
A PTLD that occurs later is more likely to be circumscribed anatomically and to be associated with a more gradual clinical course. In this situation, extranodal disease with visceral involvement is common with gastrointestinal, pulmonary, or central nervous system (CNS) symptoms. Lymphadenopathy is painless, and atypical lymphocytes may or may not be present in the white blood cell differential count.
Most patients with PTLD present with at least 1 tumor. About two thirds of these tumors are extranodal, and about one third are nodal. There is a tendency to involve specific sites. The gastrointestinal tract is involved in about 26% of cases and CNS in about 27% of cases.[37,38] The allograft can also be involved. In this case, the frequency of involvement varies according to the specific type of allograft. PTLDs that arise in lung or intestinal transplant recipients involve those allografts in up to 80% of cases. The reason for this is not known. However, it is interesting that the lung and bowel are transplanted with a large indigenous lymphoid population. PTLDs that occur in patients receiving other types of allografts, such as liver and kidney, involve the allograft in about one third of cases. In contrast, the transplanted heart is only rarely involved with these tumors.
Symptoms of PTLD are related to the site of tumor growth. Gastrointestinal tumors can cause abdominal pain with hemorrhage and may perforate and lead to acute abdomen. Central nervous system tumors cause symptoms secondary to local necrosis and tumor mass effect. However, PTLD can occur at any site. For example, isolated skin involvement has been noted,[41,42] and gallbladder involvement has been observed in one case as well. We have observed PTLD arising in the bile duct near the site of prior anastomosis in a liver transplant patient. A fulminant form of PTLD is an uncommon presentation, occurring in approximately 1% of cases. These patients present with a septic picture and may also have lymphadenopathy or frank tumors. They can also have involvement of serous surfaces and can develop pleural effusions or ascites from which tumor cells can be recovered.
The diagnosis of PTLD requires an awareness of the myriad appearances of this syndrome. Isolated or systemic lymphadenopathy or "lumps and bumps" that suddenly appear should include PTLD in the differential diagnosis. Abdominal pain, particularly with evidence of intestinal bleeding, raises the possibility of PTLD in the GI tract. In one pediatric series, diarrhea and/or gastrointestinal bleeding in the presence of active EBV infection was associated with PTLD in 43% of cases. Persistent headaches or CNS symptoms suggest localization to the brain. Upper respiratory tract infections that may be associated with lymphadenopathy or that do not resolve after a course of antibiotics should raise a suspicion of PTLD.
Localization of the dysfunction directs the diagnostic evaluation and may lead to discovery of PTLD. Endoscopic evaluation may disclose large or small ulceronodular lesions that reflect PTLD in the organ. In the case of pulmonary involvement, multiple nodular densities may be seen on x-ray.
Diagnosis is by biopsy, although cytologic examination may in some cases help to establish a diagnosis. It must be remembered that PTLD may involve the allograft in a primary fashion and also that it may occur in an infiltrative, rather than tumorous, form. Thus, every allograft biopsy specimen should evoke at least passing consideration by the examiner as to whether the cellular infiltrate seen may reflect PTLD. In summary, PTLD enters into a wide variety of differential diagnoses encountered in the posttransplantation patient, and constant vigilance is required to detect these tumors at an early stage.
Several laboratory assays have applicability in suggesting or supporting the diagnosis of PTLD. Badley and colleagues demonstrated monoclonal gammopathy in 71% of transplant recipients with and in 27% of transplant recipients without PTLD. A separate study showed that PTLD developed in 9% of all transplant recipients who had monoclonal gammopathy.
Epstein-Barr viral serologic testing may be used to evaluate the presence of recent or remote infection and thus may provide indirect information relevant to the diagnostic workup for PTLD. However, a diagnosis of EBV infection, active or remote, is not synonymous with a diagnosis of PTLD. For example, one study of EBV-seronegative pediatric liver transplant recipients showed an 80% conversion rate to seropositivity within the first 3 months after transplantation. Of these patients, approximately 85% were asymptomatic and only 15% developed PTLD.
Of the various serologic assays for EBV infection, IgM antiviral capsid antigen (IgM-VCA) is particularly useful in detecting active infection. In one study, IgM-VCA antiviral capsid antigen level was elevated an average of 5 days after a detectable rise in circulating EBV genomes shown by polymerase chain reaction (PCR) assay. Quantitative estimation of the number of EBV genomes in the peripheral blood by use of the PCR assay provides a more useful correlate of the EBV infection types most likely to be associated with PTLD.
This technique was applied following the observation that patients with PTLD had early and spontaneous outgrowth of virus when peripheral blood cells were cultured in vitro. Such outgrowth does not occur in "normal" EBV-positive patients. It was subsequently shown that patients with PTLD had elevated numbers of circulating viral genomes. Hanasono and colleaguesshowed that normal EBV-positive patients had less than 2,000 viral genomes per microgram of blood cell DNA, whereas the number of genomes was increased 10- to 100-fold in patients with PTLD. Rowe and colleagues found an increased risk of PTLD when the number of circulating EBV genomes exceeded 500/105 peripheral blood lymphocytes. Furthermore, regression of PTLD was associated with a decrease in the number of circulating viral genomes, indicating that this parameter also served as a useful means of monitoring therapy.
Allograft biopsy specimens should be examined carefully to exclude early PTLD before beginning antirejection treatment, because increasing immunosuppression is contraindicated in PTLD. Biopsy of lymph nodes and sites of extranodal involvement (ie, organs, bone marrow) is the definitive diagnostic test. Biopsy specimens are analyzed for massive infiltration of B cells and plasma cells and may be subjected to in situ hybridization or immunostaining for detection of EBV. In situ hybridization for immunoglobulin light chains or molecular analyses on a separate specimen submitted specifically for that purpose may provide information regarding the clonality of the lesion.
Lymphoproliferative lesions are currently classified according to histologic parameters. Histologic findings refer to the microscopic appearance and characteristics of the tissue. Polymorphic lesions contain a proliferation of cells with varied morphologic structure, whereas monomorphic PTLDs generally contain a uniform population of cells. With the rapid progress in molecular diagnostic techniques, including DNA array technology, it is likely that the classic approach will soon be supplemented or superseded by more comprehensive molecular approaches.
When examined under the microscope, PTLDs may show a heterogeneous or homogeneous appearance. These are termed polymorphic and monomorphic, respectively. Closely related terms are used to describe the clonal composition of these tumors. Some PTLDs arise from a number of different cells and are termed polyclonal. Others ultimately arise from a single cell and are termed monoclonal. Occasionally, a PTLD may arise from several (around 3 to 5) cells that each give rise to numerous daughter cells. Such lesions are termed oligoclonal or multiclonal. If a person has several monoclonal PTLDs at the same time, each tumor may have arisen from 1 common ancestor cell, or, in other cases, each tumor may contain a unique clone.
The relationship between histology and clonality is complex, but some generalizations can be made. "Early" lesions, such as infectious mononucleosis or plasma cell hyperplasia, are usually polyclonal or may contain a minor clone in a polyclonal background. Most, and some would say all, polymorphic PTLDs are monoclonal. Many would equate this, for all practical purposes, as evidence of a neoplastic process. All monomorphic PTLDs to date have also been monoclonal when this has been evaluated.
Determination of clonality may be approached by several means, usually by Southern blot or polymerase chain reaction. Immunostaining for immunoglobulin light chains on tissue sections can be performed, but this is decidedly less sensitive and must be rigorously controlled. In situ hybridization for immunoglobulin light chains is a better choice on conventionally processed tissue specimens. However, it must be remembered that these latter 2 studies will only be capable of detecting B-cell clonality, and the clonality of T-cell PTLD cannot be assessed except by molecular means. Evaluation of oncogene or tumor suppressor gene abnormalities is primarily based on a molecular approach at present, and these studies are generally independent from the clonal analyses.
The initial PTLD categorization was proposed 20 years ago and recognized 4 categories of posttransplantation lymphoid growths: normal reactive hyperplasia, true lymphomas originally called immunoblastic sarcomas, and two forms of intermediate tumorlike lesions known as polymorphic diffuse B-cell hyperplasia and polymorphic B-cell lymphoma. The term polymorphic referred to the heterogeneous appearances of lymphocytes in these growths. These latter two categories were applied because it was originally felt that polymorphic-reactive hyperplasias could be reliably differentiated from polymorphic lymphomas on the basis of histology alone. Experience with additional cases showed that this was not the case, however.
Furthermore, forms of PTLD that resembled different forms of "true" lymphomas other than immunoblastic sarcomas began to appear. Since the lymphomatous PTLDs had in common a fairly uniform cellular appearance, in contrast to the heterogeneous appearance of the polymorphic PTLDs, the term monomorphic was applied. It was subsequently proposed that PTLD be subdivided into 2 major categories, namely polymorphic PTLD, including polymorphic diffuse B-cell hyperplasia and polymorphic B-cell lymphoma, and monomorphic PTLD, including immunoblastic sarcoma and other forms of lymphoma. At the same time, because of advanced diagnostic techniques that allowed for increased sensitivity in the detection of EBV in tissues, it was realized that subtler forms of EBV-associated lymphoproliferations also existed. Some of these resembled infectious mononucleosis, and others showed a bland overgrowth of unremarkable plasma cells without destruction of underlying tissue.
Intermediate lesions with features of more than one PTLD were also observed, and this caused a further proliferation of terms. Knowles and colleagues formally incorporated several histologically reactive proliferations under the umbrella term PTLD and segregated these from polymorphic and lymphomatous PTLD lesions. A later effort by the American Society of Hematopathology further integrated the various terms under one working system. This formulation was instituted in 1997 and recognized three main categories together with a miscellaneous category. These included early lesions, such as posttransplantation infectious mononucleosis and plasma cell hyperplasia, that occur within the PTLD spectrum but represent benign growths. The polymorphic category of PTLD was retained, and a monomorphic category was used to subsume all the various B-, T-, and natural killer (NK) lymphocyte neoplasms observed to date. The individual lymphomatous growths were in turn categorized according to the terms of the revised European and American lymphoma classification. Miscellaneous lesions considered as forms of PTLD included Hodgkin-like lymphomas and cases indistinguishable from plasmacytoma and multiple myeloma. The terms were applicable to EBV-positive and EBV-negative lesions. Clinical relevance of this grading system was shown in a retrospective study of pediatric liver transplant patients. A separate group later suggested that the diagnosis of PTLD also contains a suffix to designate the presence or absence of EBV and the nature of the clonal status of the lesion to facilitate interinstitutional comparisons of treatment results.
The stage of PTLD represents the extent of the disease: local vs disseminated and nodal vs organ involvement. In approximately 50% of cases, multiple organs or sites are involved at the time of presentation. The lymph nodes and GI tract are the 2 most common sites. No formal system of PTLD staging exists, and it is suggested that the standard Ann Arbor classification with Cotswold modification be used when possible in reporting cases. Rare cases of posttransplantation leukemias exist, but the relationship of these to the spectrum of PTLD has not been established at this time.
A brief discussion of the pathophysiology of PTLD will help to provide a rationale for the various treatment modalities used in this disorder. EBV infection can occur as a primary disease or as a reactivation of previous infection. Cellular immunity is important in the regulation of growth of EBV-infected lymphocytes. Untreated EBV infection, whether a primary or reactivation infection in the T-cell-depleted host, leads to B-cell proliferation.
Primary disease in normal hosts occurs as a result of salivary contact with an infected person and presents as a self-limiting infectious mononucleosis. The host develops permanent antibodies, but the virus lays dormant for life. EBV seroconversion occurs early in life in the majority of the population. By adulthood, nearly 90% of the population carries EBV as a persistent, latent infection. Primary disease in the transplant recipient is transmitted from an EBV seropositive person (who may or may not be the actual organ donor) to an EBV seronegative recipient. Essentially all seronegative patients will seroconvert after transplantation. Reactivation infection occurs in the immunocompromised host who is seropositive at the time of transplantation because of EBV exposure in childhood or adolescence.
Most PTLDs arise from B cells that are infected by EBV. The EBV infects B cells by interaction of the viral surface glycoprotein gp350/200 with the cell surface protein CD21 of B cells. The virus is then endocytosed and brought into the cell nucleus, where it circularizes and equilibrates into one of several forms of latent infection.
Factors that favor one state over another state of latency are not known at present. One form of latency (latency III) is associated with the production of at least 9 viral proteins, including LMP1, LMP2a and 2b, EBNAs 1, 2, 3a, 3b, 3c, and LP. LMP1 is the main oncogenic protein of EBV. The reader is referred to other sources for a detailed account of the actions of these proteins. In aggregate, these proteins protect the B cell from death by apoptosis and cause cell proliferation. Normally, such proliferation is inhibited by host immune responses, specifically T cells. If T-cell function is compromised, such as in the case of posttransplantation immunosuppression, the B cells may be able to escape immune control. Increased numbers of circulating EBV-infected cells routinely occur in immunosuppressed transplant patients. Interestingly, these cells are not actively proliferating but represent expansion of the memory B-cell population. These B cells carry EBV in a form of latency that is antigenically silent and presumably do not attract the attention of the T-cell surveillance mechanism.
By contrast, the B cells that are found in PTLD are not resting cells but are actively proliferating cells that carry the virus in a form of latency that expresses the proteins referred to above. The events that trigger this proliferation are unknown. However, some of the viral proteins that contribute to the abnormal growth of B cells also provide immunologic targets for the host T-cell immune surveillance mechanism. Thus, the fact that B cells are able to grow despite the presence of these viral targets suggests that either the B cell is able to suppress the host immune defenses or that the immune surveillance itself is faulty. It is possible that different factors may lead to one or both of these common final pathways in an individual patient.
Posttransplantation lymphoproliferative disorders, in particular polymorphic forms, have substantial numbers of T cells and macrophages within them. This may represent an abortive host response, although the possibility that it may actually be contributing to the lesion has not been excluded. Indeed, the intratumoral cytokine environment of PTLD is TH2-like, an environment that is associated with support of B-cell growth. Several centers are currently studying the nature of the locally infiltrating immune cells in PTLD.
Other factors, including the accumulation of cellular genetic alterations, also appear to be at play in some PTLDs. One study has suggested that mutations in the bcl-6 gene may help differentiate those PTLDs that will and will not respond to conservative therapy. Additional studies are required to support this hypothesis. Alterations in oncogenes such as c-myc or Ras have also been documented in a small number of cases.[5,66]
The host immune response that leads to remission of PTLD is thought to center on the antiviral effect of the CD8+ cytotoxic T cell, which is the major anti-EBV effector cell in the normal host. Khatri and colleagues have demonstrated expansion of this subset of T cells coincident with PTLD regression in the clinical setting, supporting this concept. Other immune cells, such as macrophages, may also have a role in the control of PTLD and are under investigation in this regard.[68,69]
The origin of EBV-negative PTLDs remains obscure. It has been suggested that these tumors may simply represent sporadic lymphomas arising in patients who happen to be immunosuppressed. We have speculated that some cases may reflect a constant stimulation of lymphocytes related to the chimeric posttransplantation status. Finally, in other tumor systems, a "hit and run" hypothesis for EBV has been proposed.[71-73] In this scenario, EBV induces preneoplastic cell injury, then is disposed of by the cell when full neoplasia ensues and the virus becomes a liability by slowing growth or subjecting the cell to immune attack. The experimental support for this concept is in the early developmental stage, and the full breadth of this hypothesis remains to be established.
Early diagnosis and use of appropriate therapies are essential to the successful treatment and management of PTLDs. There is no universal approach, however. Treatment should be tailored to the specific form of disease present in the individual patient. Most centers follow a graded approach,[57,74] with the initial intervention influenced by the extent of disease and the degree of acute illness of the patient. This results in a diversity of modifications and makes comparison of uniform therapies difficult.
The most immediate treatment measure is to reduce the level of immunosuppression. Reduction of immunosuppression was first proposed by Starzl and colleagues.[75,76] Penn estimates that this intervention alone is sufficient for 31% of cases, although other series have placed this figure closer to 50%.[77,78] One group has suggested that the presence or absence of bcl-6 mutations in PTLD predicts response or nonresponse to this therapeutic approach. The general applicability of this observation requires larger numbers of cases, however. It should also be noted that some EBV-negative PTLDs, including T-cell tumors, may also respond to reduced immunosuppression. One obvious downside to this form of therapy is the risk of allograft rejection. For that reason, other approaches, both immunologic and nonimmunologic, have also been attempted. Additional therapeutic modalities are indicated for patients who do not respond to reduction of immunosuppression.
Systemic treatment is necessary for patients with extensive involvement. High-dose acyclovir, an inhibitor of EBV DNA replication, has met with some success in eliminating polyclonal oropharyngeal EBV lesions (lesions that had not progressed past the proliferative phase). However, latently infected lymphocytes are not affected by this treatment. The efficacy of antiviral therapy for treating PTLD has not been firmly established at this time.
Treatment with cytokines is a logical extension of the immunologic approach to therapy and attempts to stimulate the host immune system to reject the PTLD. Interferon alpha is the cytokine most commonly used. Faro compiled the published reports of patients treated with interferon alpha and concluded that this modality may succeed when no response has been seen with reduction of immunosuppression. However, the risk of rejection is also present with the use of this agent.
Radiotherapy has also been used and may have some benefit in the treatment of CNS tumors and for control of localized processes elsewhere. Reports on results of treatment of PTLD with conventional chemotherapeutic agents are conflicting.[8,9] Some centers report complete regression of disease, whereas others report low success and high mortality rates with this form of treatment. Early (circa 1970-1980) efforts at use of antilymphoma chemotherapy could not overcome problems inherent with the immunodeficient status of the patients, and high mortality rates resulted. More recently, modified approaches have been used, with regimens used to both treat the tumor and maintain an immunosuppressed state to preserve the allograft.[30,82,83]
Cell therapy represents a recent innovation in the treatment of EBV-related malignancies. The therapeutic use of anti-EBV immune effector cells arose from observations of the graft-vs-leukemia effect observed in BMT recipients. In those cases, donor lymphocytes with reactivity against recipient cells could also show reactivity against residual host leukemic cells. Much effort has been expended in separating this graft-vs-leukemia from the graft-vs-host reactivity and has succeeded to some extent. Indeed, in BMT recipients with relapsed chronic myelogenous leukemia, donor lymphocyte infusions from the original bone marrow donor can result in remissions in 60% to 80% of cases. This approach was applied to BMT recipients with PTLD and led to disease remission in several cases. However, toxic side effects included changes that suggested reactivity of the donor cells against the host.
A separate approach to cell therapy for PTLD was used by other groups.[78,86-88] Donor T cells were stimulated with EBV-positive B cells, and those cells with anti-EBV reactivity were collected and transfused into recipients. Heslop and colleagues and Rooney and colleagues have shown that this approach provides effective prophylaxis against PTLD in a BMT population. This form of therapy can also be used to treat established PTLD, although there are several limitations at present. First, there is a time delay needed to grow and stimulate the required number of cells. This could be overcome by preparing cells at time of transplantation in high-risk patients or by preparing banks of HLA-matched cells for use as required. Second, the use of EBV-stimulated effector cells will not target viral proteins that are not expressed in the patient's tumor virus if a mutation has occurred in the latter. Third, this approach would not be expected to be effective in cases that did not contain the virus. Other means of targeting the tumor cells would be necessary in these cases. We and others have had some success in using autologous lymphokine-activated killer cells against EBV-positive PTLD. We have not seen activity of these cells against PTLDs not containing the virus.
An alternate strategy for therapeutically targeting PTLD is the use of anti-B-cell antibodies. Monoclonal anti-B-cell antibodies have been used against standard non-Hodgkin lymphomas, and similar antibodies have been applied to PTLD in small numbers in the past with encouraging results. Benkerrou and colleagues reported a 58% response rate in an early report of PTLD treatment by means of monoclonal antibodies. Senderowicz and colleagues reported on the successful use of an anti-B-cell (CD22)-conjugated immunotoxin to treat a PTLD that was refractory to prior chemotherapy. The role of antibody therapy for PTLD needs to be defined, preferably by a multicenter study. Anti-EBV antibodies may also provide some activity against the virus or virus-containing PTLD, but this possibility has not yet been conclusively shown.
Overall response and survival rates are difficult to compare because of the wide range of PTLD forms and therapies. Furthermore, crude rather than actuarial survival rates are often reported. Within these limitations, a review of relevant literature shows responses that tend to vary according to histology and stage. In two separate series,[5,93] mortality from polymorphic and monomorphic PTLDs ranged from 0% to 20% and 67% to 87%, respectively. PTLDs with abnormalities of oncogenes or tumor suppressor genes would fit within the monomorphic category, and these abnormalities appear to augur a worse prognosis.[5,66]
In a separate compiled series, it was observed that 44% of PTLD survivors had involvement of only one organ, and involvement of 3 or more organs occurred in 57% of fatal cases. Dror and colleagues considered thrombocytopenia and neutropenia to represent negative prognostic indicators and PTLD histology and stage to be marginally significant in their series. An absence of stage effect on survival was also reported in a retrospective review of 27 pediatric patients. In this series, mortality was more closely related to the underlying procedure, with BMT and heart transplant recipients having higher (100% and 55%, respectively) mortality rates than liver and kidney transplant recipients (0% and 0%, respectively). Gross and colleagues reported a 92% mortality in PTLD arising in recipients of allogeneic hematopoietic stem cell transplants. In their series, the only responders seen were among those patients treated with interferon alpha. In a separate pediatric liver transplant series, 4 patients with B-cell lymphoma and 1 with B-cell leukemia were successfully treated with reduced immunosuppression and high-dose acyclovir alone (2/5) or with this treatment followed by chemotherapy (3/5).
The heterogeneity of these reports exemplifies the variable results seen with different treatment regimens among different centers and argues for standardized multicenter therapeutic trials against this disease. In our own series of 256 patients with PTLD, the overall 2-year actuarial survival is 90%, and the overall actuarial 5-year survival is 77%. This reflects the particular mixture of patients with PTLD seen and therapies used at our institution. Therefore, direct comparison to the results of other centers may not be applicable. A detailed analysis of this series will be the subject of upcoming clinical reports.
In conclusion, we have learned much about PTLD, but much cooperative work remains to be done before this disease can be conquered. EBV positive PTLD has served as a model of virus-associated lymphomagenesis, and the lessons learned here may have applicability in other tumor systems as well. As EBV-positive lymphomas are prevented or successfully treated, our attention will turn to EBV-negative tumors and other forms of neoplasia that continue to plague transplant recipients. For the present, however, awareness of PTLD and an aggressive stance toward the diagnosis of this disease, together with graded therapy using the best means available for the individual disease, remain our best form of defense against this theoretically fascinating yet deadly enemy.