The disease currently recognized as angioimmunoblastic T-cell lymphoma (AITL) was first described in the 1970s as a clinical syndrome characterized by generalized lymphadenopathy, hepatosplenomegaly, anaemia and hypergammaglobulinaemia (Frizzera et al, 1974; Lukes & Tindle, 1975; Lennert, 1979). The lymph node histology was observed to show a number of distinctive features such as the partial effacement of normal architecture by a polymorphic inflammatory infiltrate, including large blasts and marked vascular proliferation. Based on these histological appearances, the disease was initially referred to by a variety of terms, including immunoblastic lymphadenopathy (Lukes & Tindle, 1975), lymphogranulomatosis X (Lennert, 1979) and angioimmunoblastic lymphadenopathy with dysproteinaemia (AILD) (Frizzera et al, 1974). The latter term was accepted by most investigators and has come to define this clinical syndrome. AILD was initially thought to be an atypical lymphoid hyperplasia, a premalignant lesion, with a tendency to develop into a lymphoma rather than a frank neoplasm at onset. With the advent of immunophenotyping and molecular techniques, it became apparent that most cases of AILD contained monoclonal T-cell populations as well as clonal cytogenetic abnormalities, strongly suggesting that the majority of the patients were neoplastic from the onset. AITL was included in the Updated Kiel Classification for Lymphomas (Stansfeld et al, 1988) and is accepted as a distinct clinicopathological entity in the current WHO classification (Jaffe et al, 2001).
It is likely that earlier series, published before the availability of immunophenotyping and molecular tests, are diluted by a number of reactive proliferations. In this review, as far as possible, we have used the data from relatively recent reports published after the neoplastic nature of AITL was established.
Risk factors and aetiology
No consistent risk factors for the development of the disease have been reported and to date no aetiological agent has been identified. A significant proportion of patients have a history of drug use (Tobinai et al, 1988; Sasaki & Sumida, 2000), in particular, antibiotics. These are more likely to represent drugs prescribed to the patients because of systemic illness clinically mimicking an infectious process, rather than the primary cause of the disease.
A number of infectious diseases and agents have been reported to be associated with AITL, including bacterial infections such as tuberculosis (Rho et al, 1996) and fungal infections such as cryptococcus (Konig et al, 1991). Perhaps most intriguing is the relationship of AITL with a number lymphotropic viruses. The most significant of these is the Epstein–Barr virus (EBV), which is discussed separately below. The presence of other viruses, including human herpes virus (HHV)-6 (Luppi et al, 1993), HHV-8 (Luppi et al, 1996), human immunodeficiency virus (HIV) (Helm et al, 1990) and hepatitis C virus (Luppi & Torelli, 1996), have been reported in AITL. Other than EBV, the evidence for the role of a viral infection in the pathogenesis of AITL remains tenuous. HHV-6 was reported in seven out of 12 AITL patients identified by the polymerase chain reaction (PCR) (Luppi et al, 1993), but immunohistochemistry (Luppi et al, 1998) and in situ hybridization studies (Khan et al, 1993) showed that the expression of HHV-6 antigens was absent in the T cells, which is suggestive of a lack of a direct role for HHV-6. Occasional cases of AITL have also been described to have HHV-8 infection by PCR, but we (unpublished data) and others (Chadburn et al, 1997) failed to observe any evidence of HHV8 infection by PCR or immunohistochemistry in AITL, which, once again, argues against a causal relationship. Associations with other viruses are limited to a few case reports and are likely to be coincidences.
The lymph node biopsy of AITL is characterized by the partial effacement of the lymph node architecture by a polymorphic infiltrate, predominantly occupying the paracortical (interfollicular) areas. Usually the lymph node sinuses are preserved but the infiltrate leaks into the perinodal tissue. In the original descriptions of AITL, the absence of reactive hyperplastic B-cell follicles was considered to be a characteristic feature; it is now recognized that the architectural changes in AITL fall into three overlapping patterns (Ree et al, 1998, 1999; Attygalle et al, 2002a). In pattern I (20% of the patients), there is a preservation of the lymph node architecture. Hyperplastic B-cell follicles with poorly developed mantle zones and ill-defined borders are easily identifiable in the cortex of the lymph node. These merge into the expanded paracortex containing a polymorphic infiltrate of lymphocytes, transformed large lymphoid blasts, occasional multinucleate cells that are reminiscent of Reed–Sternberg cells (RS), plasma cells, macrophages and eosinophils within a prominent vascular network (Fig 1).
Pattern II (30% of the patients) is characterized by the loss of the normal architecture, except for the presence of occasional depleted follicles with concentrically arranged follicular dendritic cells (FDC). In some patients, FDC proliferation extending beyond the follicles can be seen. The remainder of the node shows a polymorphic infiltrate with increased numbers of transformed lymphoid blasts and vascular proliferation similar to that described for pattern I (Fig 1).
In pattern III (50% of the patients), the normal architecture is completely effaced and no B-cell follicles are present. Prominent irregular proliferation of FDCs can be seen in haematoxylin and eosin (H&E)-stained sections in most patients, and this is accompanied by extensive vascular proliferation and a polymorphic infiltrate similar to that seen in patterns I and II (Fig 1). Approximately half of the patients contain perivascular collections of atypical, medium–large-sized lymphoid cells with either clear or pale cytoplasm, whereas in other patients cytological features of malignancy may not be apparent. In a few cases, where consecutive biopsies from the same patient have been reviewed, an apparent transition from pattern 1 to pattern III as the tumour progresses has been noted, suggesting that pattern III patients represent advanced disease (Ree et al, 1998; Attygalle et al, 2002a).
The reproducibility of the diagnosis based on histological appearances appears to be high, at least amongst expert haematopathologists (Rudiger et al, 2002). In most patients, histological examination has to be supplemented by additional ancillary tests. The histological diagnosis may be problematic in a small number of patients as the morphology and phenotypic features can overlap with a variety of reactive and neoplastic conditions, such as reactive lymphadenopathies, multicentric Castleman's disease, diffuse large B-cell lymphoma and classical Hodgkin's lymphoma (Attygalle et al, 2002a).
Although generalized lymphadenopathy is the main presenting sign and the diagnosis of AITL rests on histological examination of the lymph node, many patients have evidence of extranodal involvement at the time of diagnosis. The most frequently involved extranodal sites include the bone marrow, spleen, skin and lungs. The histological appearances in these sites are usually non-specific but mimic some of the features described in the lymph node, including increased vascularity and a polymorphic inflammatory infiltrate with or without clear cells (Seehafer et al, 1980; Ghani & Krause, 1985; Brown et al, 2001). Cytological features of malignancy can rarely be identified, and tumour involvement can only be shown by immunohistochemistry and molecular clonality analysis (Martel et al, 2000; Murakami et al, 2001).
Immunohistochemistry shows the expansion of the interfollicular areas by a diffuse infiltrate of CD3+ T cells. In most patients, CD4+ T cells dominate but there is usually an intermixed population of CD8+ T cells. The B-cell markers CD20 and CD79a highlight the residual follicle centre and mantle zone B cells as well as many of the large transformed blasts and RS-like cells in the interfollicular areas. In some instances, these can be numerous, mimicking a large B-cell lymphoma or classical Hodgkin's lymphoma, though they are typically polytypic for light chain expression. As described by histology, one of the most distinctive features of AITL is the proliferation of FDC which is best appreciated with immunostaining for the FDC markers CD21, CD23 or CD35 (Leung et al, 1993; Raymond et al, 1997; Jones et al, 1998; Bagdi et al, 2001). In early cases (pattern 1), the FDC proliferation can be minimal and limited to a few follicles. In more advanced cases (Pattern II and III), the FDC are seen to extend into the interfollicular area and wrap around the arborizing vessels (Attygalle et al, 2002a) (Fig 2).
Until recently, the assessment of the immunophenotype of AITL tumour cells was difficult, as in many patients the neoplastic cells were not readily identifiable and no specific markers were available to recognize the tumour cells. However, studies using immunohistochemistry, single-cell microdissection and molecular analysis have now shown that the neoplastic tumour cells of AITL can be recognized by the aberrant expression of CD10 (Attygalle et al, 2002a). CD10, also known as neutral endopeptidase (NEP), is a cell surface metallopeptidase normally expressed by a subset of lymphocytes, and epithelial and stromal cells. It belongs to a family of membrane peptidases that includes, among others, the structurally related leucocyte-associated molecules CD26 and CD13. Physiologically CD10/NEP reduces the peptide concentration available for receptor binding and regulates a number of cell functions (Antczak et al, 2001a, b; Turner et al, 2001). In haematopathology, CD10 has been extensively used for the diagnosis of lymphomas because of its restricted expression by the precursor B-cell and follicle centre B-cell compartments (Chu et al, 2000; Dogan et al, 2000). In both lung (Shipp et al, 1991) and prostate cancer (Papandreou et al, 1998), CD10 has been shown to regulate tumour survival in vivo by decreasing extracellular neuropeptide concentrations and inhibiting certain signal transduction pathways (Sumitomo et al, 2000, 2001). In lymphocytes, its expression is largely restricted to B-cell compartments with a typically high rate of apoptosis (Chu et al, 2000). CD10 is not expressed by normal peripheral T cells or nodal peripheral T-cell lymphomas, but is induced in T cells undergoing apoptosis during HIV infection in both in vivo and in vitro experiments (Cutrona et al, 1999). This close association with CD10 expression and apoptosis suggests that CD10 plays a role in the regulation of lymphocyte survival.
In AITL, it appears that most, if not all, tumour cells express CD10, and CD10-positive T cells account for only a fraction of the infiltrate, varying from around 5% of all cells in pattern II patients to up to 30% of all cells in pattern III patients (Fig 2). Interestingly in early cases, the CD10-positive cells have low-grade cytology and a relatively low proliferation fraction. CD10-positive T cells are intimately related to the residual reactive B-cell follicles and the expanded FDC meshwork, some being located within the follicle centres and others in the area surrounding the follicles. As the tumour progresses (patterns II and III), the tumour cells spill into the interfollicular area but retain the intimate association with the FDC meshwork. This suggests that the FDC microenvironment may be important for tumour growth. Moreover, in many patients, the neoplastic cells stain for CD4, CD57 and B-cell lymphoma-6 (Bcl-6), markers expressed by normal follicle centre T cells, suggesting that the tumour may have derived from this cell population (Ree et al, 1999; de Leval et al, 2001; Yuan et al, 2002).
The neoplastic T cells in extranodal dissemination may retain the expression of CD10 and this may aid diagnosis (Attygalle et al, 2002b).
There appears to be a substantial immune activation in the peripheral blood and lymph nodes of AITL patients compared with the reactive lymph nodes or other peripheral T-cell lymphomas. The evidence for this includes increased levels of serum soluble interleukin 2 (IL-2) receptor, CD30 and CD8 molecules (Pizzolo et al, 1990), and also the expression of an array of cytokines such as tumour necrosis factor alpha, lymphotoxin, IL-1 beta, IL-2, IL-4, IL-6, IL-13 and interferon gamma (Takeshita et al, 1993; Foss et al, 1995; Ohshima et al, 2000).
Recent studies examining the expression of T-cell activation markers have shown an increased expression of the TNF receptor family member CD134 (OX40) (Jones et al, 1999), chemokine receptor CXCR3 (Jones et al, 2000), and CD69, which is a marker of early T-cell activation (Dorfman & Shahsafaei, 2002). These molecules are preferentially associated with the T-helper (Th)-1 phenotype in normal T cells, suggesting that Th-1-type differentiation, characterized by IL-2 and interferon gamma production, is a feature of AITL.
In contrast to this marked immune activation, the functional studies performed on T cells recovered from lymph nodes and peripheral blood of AITL patients have shown defective T-cell responses supporting an underlying immunodeficiency. The abnormalities reported include a reduction of the absolute number of circulating T cells, inversion of the CD4/CD8 ratio, high percentages of activated T cells (CD8+/HLA-DR+), defective T-cell response in vitro to the phytohaemagglutinin (PHA) mitogen, and minimal helper and enhanced in vitro suppressor functions (Pizzolo et al, 1987).
Vascular endothelial growth factor (VEGF), one of the main angiogenic cytokines in human solid tumours, is expressed at high levels by the stromal cells in AITL (Foss et al, 1997). It has been hypothesized that it may be responsible for the vascular proliferation observed in AITL. Interestingly, the increased VEGF signal appears to correlate with the number of mast cells infiltrating the tumour, suggesting that mast cell activation may also play a role in the pathogenesis (Fukushima et al, 2001).
A characteristic feature of AITL, seen in over 95% of all patients, is the presence of increased numbers of EBV-infected cells compared with both normal lymph nodes and peripheral T-cell lymphomas. Initial studies suggested the EBV-infected cells might be within both the T-cell and B-cell population (Anagnostopoulos et al, 1992). In contrast, studies using double immunohistochemistry, in situ hybridization and microdissection have shown that virtually all cells infected by EBV are B cells, and that EBV infection is unlikely to play a primary role the lymphomagenesis of AITL (Weiss et al, 1992; Ohshima et al, 1994; Brauninger et al, 2001). The EBV-infected cells have a variable cytology, with some having an immunoblast-like and others having an RS-like appearance. They account for most of the large B cells present in the interfollicular zone of the AITL lymph nodes.
The EBV protein expression pattern in AITL B cells is usually consistent with latency (Anagnostopoulos et al, 1995; Zettl et al, 2002). The causes for the expansion of EBV-infected cells are not known. It is likely that an underlying immunodeficiency with reduced cytotoxic activity, and the presence of growth factors favouring the outgrowth of EBV-infected cells play roles.
Before the availability of molecular tools that can demonstrate the presence of expanded clones of T and B cells, the neoplastic nature of AITL was not fully appreciated and the disease was considered to be a premalignant state from which high-grade, large cell lymphomas occasionally arose. When Southern blotting technology became available, Weiss and colleagues were the first to investigate the clonality of lymphocytes in AITL (Weiss et al, 1986). They demonstrated the presence of clonal T-cell populations, not only in patients with histologically apparent lymphoma, but also in patients diagnosed as AILD who lacked the cytological features of malignancy. This was followed by a number of similar publications confirming the presence of a monoclonal T-cell population in virtually all patients with AILD/AITL (Lipford et al, 1987; Feller et al, 1988; Tobinai et al, 1988; Ree et al, 1998; Smith et al, 2000; Willenbrock et al, 2001; Attygalle et al, 2002a). The results of clonality analyses from the large series are summarized in Table I. One of the unusual observations to be made during these studies was the presence of an expanded monoclonal B-cell population in a significant minority of patients. This has led to the suggestion that AITL is a mixture of T-cell lymphomas and B-cell lymphomas, but not a single clinicopathological entity. However, the current evidence suggests that these patients also fall within the framework of AITL. These patients typically exhibit increased numbers of EBV-infected B-large cells, and it is thought that the B-cell clone detected by molecular analysis of whole lymph node extracts lie within this population. It is likely that this is an EBV-driven lymphoproliferation that occurs secondary to the immunodeficiency associated with the underlying AITL, perhaps similarly to other EBV-driven lymphoproliferations that are associated with immunosuppression (Zettl et al, 2002). Not surprisingly, a subset of AITL patients go on to develop full-scale EBV-associated, diffuse large B-cell lymphomas (Nathwani et al, 1978; Abruzzo et al, 1993; Knecht et al, 1995) or Burkitt's lymphoma (Mazur et al, 1979).
The studies on the genetic changes occurring in AITL have been hampered by a number of factors, including the relative rarity of the tumour and dilution of tumour cells by a large number of reactive cells. Most of our knowledge on genetic changes comes from cytogenetic-based studies. These are summarized in Table II.
Table II. Cytogenetic alterations in AITL.
Patients with cytogenetic alterations
Patients with clonal cytogenetic alterations
Cytogenetic abnormalities frequently encountered
All studies have used metaphase cytogenetics; Schlegelberger et al (1994) have used both metaphase and interphase cytogenetics. + 3, + 5, + 15, + 19, + 21, + 22 and + X refer to trisomy of chromosomes 3, 5, 15, 19, 21, 22 and X respectively.
Approximately 90% of AITL patients have cytogenetic alterations observed by the use of combined metaphase and interphase cytogenetics (Kaneko et al, 1982, 1988; Godde-Salz et al, 1987; Frizzera et al, 1989; Cosimi et al, 1990; Schlegelberger et al, 1990a, 1994). Clonal chromosomal aberrations are seen in approximately 70% of the patients. Trisomy 3, trisomy 5 and gain of an X chromosome are the most frequent recurrent abnormalities seen in AITL, but these are also present in other peripheral T-cell lymphomas. Schlegelberger and colleagues showed that half the patients harboured cytogenetically unrelated clones, which is a unique phenomenon that is exceptional in other lymphomas (Schlegelberger et al, 1990a, 1994). This was hypothesized to be consistent with the stepwise development of chromosomal aberrations in AITL. The steps being the appearance of chromosomal abnormalities in different cells because of genetic instability and the impaired elimination of aberrant cells due to the immune defect, followed by the establishment of chromosomally aberrant clones, and finally a cytogenetically detectable level of monoclonal proliferation. However, at the first step, T-cell receptor gene rearrangement shows clonal T-cell proliferation, indicating that the abnormal tumour clone is present at the onset, although perhaps not detectable by cytogenetic methods that have a low sensitivity.
The presence of T-cell clones and the EBV-driven expansion of B-cell clones raises the question as to which cells harbour the aberrant karyotype. In one patient, using a method involving the simultaneous demonstration of immunophenotype and karyotype, it was shown that the aberrant mitoses with trisomy 3 were CD3 positive and, therefore, T cells (Schlegelberger et al, 1990b). It is possible that the non-clonal aberrations seen in AITL originate from EBV-infected B cells that are likely to have an in-vitro growth advantage.
The lower proportion of aberrant cells reported on interphase cytogenetics compared with metaphase cytogenetics is consistent with the morphology of AITL, with a few neoplastic cells amidst an abundance of reactive cells (Schlegelberger et al, 1994).
In a study assessing the significance of cytogenetics on clinical outcome, only the presence of complex aberrant clones was determined to be an independent prognostic factor, and trisomy 3 had no effect on survival (Schlegelberger et al, 1996).
Genes known to be critical in lymphomagenesis have rarely been studied in AITL. Petit and colleagues examined p53 expression and mutations, and report that both are rare, suggesting that this pathway is not altered in the majority of AITL patients (Petit et al, 2001). The other gene that has been examined is BCL-6. Rearrangements and/or mutations of the 5′ non-coding region of the BCL-6 gene play a role in the development of diffuse large B-cell lymphomas. Despite the expression of BCL-6 in most patients with AITL (Ree et al, 1999; Attygalle et al, 2002a; Yuan et al, 2002), no mutations were detected in the 5′ non-coding region of the gene (Kerl et al, 2001).
Non-Hodgkin's lymphoma is the seventh commonest type of cancer in adults in the United Kingdom (UK), with over 8600 new patients diagnosed each year. Ten per cent of these are peripheral T-cell lymphomas, and AITL accounts for approximately 2% of all non-Hodgkin's lymphomas (Rudiger et al, 2002), with an estimated 150–200 new patients diagnosed in the UK each year. Our experience in a tertiary referral centre suggests that AITL patients are being significantly under diagnosed, indicating that the real incidence may be much higher (Ree et al, 1998; Attygalle et al, 2002a).
AITL typically presents with systemic illness, characterized by B symptoms and generalized lymphadenopathy, often mimicking an infectious process. The majority of patients show hepatosplenomegaly and pruritis, and a skin rash is also seen in a half of patients. The reported frequency of common presenting symptoms and signs observed in AITL are summarized in Table III (Tobinai et al, 1988; Siegert et al, 1995; Pautier et al, 1999).
Laboratory investigations often show the presence of anaemia and occasionally pancytopenia. Typically, there is hypergammaglobulinaemia, and both the lactate dehydrogenase (LDH) and the erythrocyte sedimentation rate (ESR) are often elevated. A significant proportion of patients have circulating autoantibodies, including a positive Coomb's test, cold agglutinins, cryoglobulins and circulating immune complexes. The most common laboratory findings and their frequencies are shown in Table IV (Tobinai et al, 1988; Siegert et al, 1995; Pautier et al, 1999).
The clinical syndrome of AITL overlaps with a wide range of inflammatory and neoplastic processes, and the changes in peripheral blood and on bone marrow examination are usually non-specific. Fine needle aspiration of the enlarged lymph node may be helpful, but is rarely diagnostic because cytological appearances can be within normal limits and architectural features cannot be obtained. For the same reasons, a needle core also has limited value. The diagnosis of AITL can only be achieved by biopsy and histological examination of one of the enlarged lymph nodes, where characteristic morphological features can be best appreciated.
Publications regarding the outcome and clinical management of AITL are limited because of the rarity of the disease. Most of the information is based on retrospective studies, small patient numbers and a limited number of case reports.
The clinical outcome of the AITL remains dismal, with a median survival of less than 36 months and a 5-year survival of around 30–35% (Siegert et al, 1992, 1995; Pautier et al, 1999). Most patients die of infectious complications rather than tumour load, suggesting that an underlying immunodeficiency significantly contributes to the AITL-associated mortality.
Both single agent and combination chemotherapeutic regimens such as CHOP (cyclophosphamide, hydroxydaunomycin, Oncovin, prednisone), CVP (cyclophosphamide, vincristine, prednisone), VAP (vincristine, asparaginase, prednisone), steroids with or without cyclophosphamide, and high-dose methylprednisolone, prednisone with or without COPBLAM (cyclophosphamide, Oncovin, prednisone, bleomycin, Adriamycin, Matulane) or IMVP-16 (ifosfamide, methotrexate, VP-16) have been reported (Colbert et al, 1982; Awidi et al, 1983; Siegert et al, 1992, 1995; Pautier et al, 1999). Although a complete remission rate of 50% can be achieved with combination chemotherapy, relapse rates remain high. Overall, combination chemotherapy appears to be superior to steroids alone (Pautier et al, 1999).
Other therapeutic approaches, including low-dose methotrexate together with steroids (Gerlando et al, 2000), fludarabine (Ong et al, 1996; Hast et al, 1999; Tsatalas et al, 2001) and 2-chlorodeoxyadenosine (Sallah & Bernard, 1996) can be beneficial, but again studies are based on a small number of patients, which does not allow statistically significant conclusions. Interferon-alpha has been used for consolidation–maintenance therapy following conventional treatment to prolong chemotherapy-induced remissions as a result of its differentiating, immunomodulating and antiproliferative effects (Feremans & Khodadadi, 1987; Meuthen et al, 1990; Hast & Gustafsson, 1991; Schwarzmeier et al, 1991; Siegert et al, 1991; Pautier et al, 1999). In the majority of the patients, the remission duration is variable but is not longer than that observed with conventional treatments. Cyclosporin A, which has a suppressive effect on the immune system, most notably on T cells, but also has a direct cytotoxic/apoptosis-inducing effect on lymphocytes, has also been given (Murayama et al, 1992; Advani et al, 1997; Takemori et al, 1999). Its combined effects on neoplastic T cells may play an important role in the achievement of remission, but once again studies are limited to a few case reports. Thalidomide has been used as antiangiogenetic agent in a few patients, either following relapse or in refractory AITL, with promising results (Strupp et al, 2002). There are little published data regarding the effectiveness of high-dose chemotherapy followed by blood stem cell transplantation in this group of lymphoma patients. Two patients treated after relapse with this approach responded favourably (Pautier et al, 1999).
AITL is a systemic disease characterized by the monoclonal proliferation of T cells expressing CD3 and CD4. In most patients, the tumour cells are greatly outnumbered by numerous relative cells. This raises a number of difficulties, regarding both diagnosis and the investigation of the biological characteristics of the tumour cells, and, not surprisingly, the clinical outcome of these patients remains bleak, despite improvements in the management of other aggressive lymphomas. In this respect, the emergence of the aberrant expression of CD10 as marker of neoplastic T cells of AITL will be important. This antigen provides an objective phenotypic marker for diagnosis for the first time, as well as a tool to investigate the biological properties of AITL tumour cells.
Our future success in dealing with AITL will depend on the progress we make in understanding the biology of the disease and in establishing international collaborations to test biological discoveries in large clinical trials.