Main

Adult T-cell leukemia/lymphoma (ATLL) is a lymphoproliferative disorder (LPD) caused by human T-cell lymphotropic virus type-1 (HTLV-1) infection, which is highly endemic, especially in southern Japan and the Caribbean basin.1 In general, ATLL shows rapid progression, drug resistance and poor prognosis.1 ATLL cells are believed to be derived from post-thymic helper T cells with a phenotype of CD4+ and CD25+. However, ATLL cells are not regarded as having a helper function; on the contrary, they have demonstrated a strong immunosuppressive activity in vitro.2 In fact, ATLL is involved in opportunistic infections more frequently than other lymphoma,3 such as Pneumocystis carinii pneumonitis4 and Strongyloidiasis.5 Furthermore, impaired cell-mediated immunity in this disease has been well documented.6

Recent reports have characterized regulatory T cells, known as Tregs, in terms of coexpression of CD4 and CD25.7 Upon T-cell receptor stimulation, Tregs potently suppress the activation/proliferation of other CD4+ T cells and CD8+ T cells in an antigen-nonspecific manner8 and in vivo usually serve to protect the host against the development of autoimmunity.9, 10, 11 But in some malignant neoplasms, for example, pancreatic and breast carcinoma12 and B-cell lymphoma,13 Tregs suppress the reaction of lymphocytes to tumor antigens and induce malignant neoplasm progression. Recent studies have also shown that Tregs help malaria parasites avoid host immunity.14 These findings indicate that Tregs sometimes account for unfavorable immunodeficiency and it is possible that abnormal Treg proliferation may also induce an unfavorable immunosuppressive state.

FOXP3 represents a new T-cell marker considered to be more specific for Treg cells than other markers such as CD4 and CD25.15, 16, 17, 18 Gene transfer of FOXP3 converts naïve CD4+CD25− T cells into a regulatory phenotype in both mice and humans, thus making this molecule a functionally important marker of this Treg-cell population.17, 18, 19

After the existence of Tregs with a distinct phenotype similar to that of ATLL cells had been substantiated, we evaluated ATLL cells in terms of Treg cells and demonstrated FOXP3 expression in some ATLL cells, a finding supported by some other reports.20, 21, 22, 23 Furthermore, recent studies have found that one of the functions of FOXP3(+) ATLL cells is a Treg-like suppressive activity.23, 24, 25, 26 Roncador et al22 reported that FOXP3 was a specific marker of ATLL in peripheral T-cell lymphoma.22 We can therefore assume that some ATLL cells express FOXP3 and may be derived from Tregs. However, FOXP3-negative (−) ATLL cases are regularly identified but the frequency of these cases has not been well established, and no studies have attempted to analyze the pathological and clinical implications of FOXP3 expression in ATLL. In the study presented here, we compared FOXP3(+) ATLL and FOXP3(−) ATLL both clinically and pathologically to establish the importance of FOXP3 expression in ATLL cases.

Materials and methods

Biological Materials

The study population consisted of 110 cases retrieved from the archives of the Department of Pathology, Fukuoka University (1992–2005), to which 59 cases were added from the archives of the Department of Pathology, Kurume University (2005–2006). All samples were from lymph nodes and were obtained with informed consent and with the approval of the doctors of the participating hospitals. Serological tests had proved that the peripheral blood of all patients was positive for anti-adult T-cell leukemia/lymphoma-associated antigen. Parts of the tissue samples were formalin-fixed, paraffin-embedded and then stained with the hematoxylin–eosin method. Two hematopathological specialists (KO and MK) reviewed all cases.

Clinical information was obtained by reviewing the tumor registry records and/or the patients' medical charts.

Morphological Variants of ATLL

For this study, we established three morphological variants mainly by following the new WHO classification.1 First, we divided the cases into diffuse pleomorphic cell variant, which is the most common variant, and anaplastic large cell variant (ALCL).1, 27, 28, 29 The pleomorphic cell variant was then subdivided into (1) pleomorphic cell variant, medium-sized cells predominant (pleo medium), and (2) pleomorphic cell variant, large cells predominant (pleo large). Both pleo medium and pleo large types show diffuse proliferation of atypical cells with irregular nuclei and coarse chromatin, but the pleo medium type is mainly (>50%) composed of medium-sized cells (Figure 1a) and the pleo large type mainly (>50%) large cells (Figure 1b). The so-called small-cell variant of ATLL1 was included in the pleo medium type. The ALCL shows diffuse proliferation of atypical large lymphoid cells with oval or ‘kidney-shaped’ nuclei and abundant cytoplasm (Figure 1c), and lymphoma cells are usually positive for CD30 (Figure 1d). In some of these, a cohesive pattern was detected, indicating the presence of ‘anaplastic large cell lymphoma’. Two Hodgkin-like ATLL30 cases and one AILT-like ATLL31 case were excluded from this study because they are extremely rare variants.

Figure 1
figure 1figure 1

(a–d) Morphological variants of ATLL. (a, b) The majority of cases showed diffuse proliferation of pleomorphic lymphoma cells. (a) Some cases were mainly composed of small- to medium-sized atypical lymphoid cells whereas (b) large cells were evident in some cases. (c) An ALCL composed of large lymphoma cells with ‘horse-shoe’ or ‘kidney-like’ nuclei and abundant cytoplasm. Multinucleated giant cells also scattered. (d) ALCL. Large lymphoma cells, including multinucleated giant cells, are positive for CD30. (e) EBV was detected by in situ hybridization. This figure shows scattered EBV-infected cells. EBER in situ hybridization. (f , g) FOXP3 expression in ATLL. (f) About 70% of lymphoma cells were positive for FOXP3. (g) Almost all lymphoma cells were negative for FOXP3. Two centrally located small cells, positive for FOXP3, were considered to be non-lymphoma cells because they lacked nuclear atypia. Original magnification for all figures: × 400.

Full size image

Evaluation of Immunodeficient State

We asked the participating clinicians whether any patients suffered from infection during the course of the disease. Almost all of them experienced some infectious state, for example, fever, elevated inflammation-associated protein (C-reactive protein, etc) and upper respiratory tract infection. However, to detect a ‘severe and definitive’ infectious state as an indication of an immunosuppressive state, we excluded cases with an inflammatory/infectious state during neutropenia (neutrophils in peripheral blood: <500/μl) after chemotherapy or during the end stage (in this study, during 3 weeks before death).

Immunostaining

Paraffin sections from each of the samples were immunostained with monoclonal antibodies against CD4 (Nichirei, Tokyo, Japan), CCR4 (PharMingen, San Diego, CA, USA), CD25 (Novocastra, Newcastle, UK), CD30 (DakoCytomation, Kyoto, Japan) and FOXP3 (eBioscience, San Diego, CA, USA), using heat-mediated antigen retrieval. A tissue sample was considered positive if more than 30% of the tumor cells were positive. Frozen sections were used for staining with the antibody against CCR8 (Genentech, Minneapolis, MN, USA).

In Situ Hybridization

Epstein–Barr virus (EBV)-encoded small RNA (EBER) was detected by in situ hybridization, using the EBER-PNA probe kit (DakoCytomation) according to the manufacturer's instructions. The evaluation was positive if positive cells were evident and accounted for more than 1% of the scattered cells (Figure 1e) with reference to the grading reported by Dupuis et al.32

Southern Blot Analysis and Cytogenetic Analysis

The remaining parts of the tissue samples were frozen and used for DNA isolation and gene analysis. The monoclonal or oligoclonal integration of HTLV-1 pro-viral DNA was examined by means of digestion with EcoRI, as described previously,33 and all cases (169 cases) were found to have pro-viral integration. The usual G-banding method was used for cytogenetic analysis. Karyotypes were described according to the International System for Human Cytogenetics Nomenclature (1995). Numbers of chromosomal breaks from structural abnormalities were calculated as the sum of all breaks whose derivation by band could be identified in the stemline karyotype. Marker chromosomes were not included in the analysis of numbers of chromosomal breaks. Multiple rearrangements of the same break site observed in the same patient were recorded separately.

Statistical Analysis

Student's t-test and χ2 test were used for comparisons of clinical and pathological findings among different groups. Survival curves were calculated with the Kaplan–Meyer method. The log rank test was used for survival analysis. A P-value less than 0.05 was considered to denote a statistically significant difference.

Results

FOXP3 Expression and Histological Findings

Of the total 169 cases, 147 (87%) were classified as pleomorphic cell variant and 22 (13%) as ALCL (Figure 1c and d).1, 27, 28, 29 Of the former, 71 cases (48%) were classified as pleo medium (Figure 1a) and 76 (52%) as pleo large (Figure 1b). Sixteen cases of ALCL were analyzed with regard to CD30 expression and almost all of the cases (15/16) showed positivity. Among 169 cases, 60 (36%) were positive for FOXP3 (Figure 1f and g). Of the FOXP3(+) cases, 36 were included in the pleo medium type, accounting for 60%, a significantly higher percentage than for the other 2 types. Moreover, only one case of ALCL (2%) was positive for FOXP3, which was significantly lower than for the other two types.

Immunohistochemical Analysis

Immunohistochemical analysis was performed in all cases (169 cases) in CCR4 (Table 1). Only 96 cases, however, were additionally evaluated in CD4, CD25 and CD30 because of the limitation in the number of unstained sections. The lymphoma cells were mainly positive for CD4 and CD25. There was no significant difference between FOXP3(+) ATLL and FOXP3(−) ATLL in CD4 and CD25. The CD4+CD25+ phenotype was detected in 80 cases (83%), and 16 cases (17%) did not show CD4+CD25+ phenotype. But there was no significant difference in FOXP3-positive rate between the two groups (33/91 and 5/20; P=0.57). FOXP3(−) cases showed relatively higher positivity for CD30 due to the fact that almost all cases of ALCL showed CD30 positivity. We stained ATLL cases with the chemokine receptors CCR4 and CCR8, which have been identified as Treg-specific molecules.34 Almost all cases of ATLL (95%) were positive for CCR4, while only about 34% were positive for CCR8. CCR4 was relatively more frequently expressed in FOXP3(+) than in FOXP3(−) type, although not significantly so (P=0.12), while there was no difference in CCR8 frequency between the two types (P=0.66).

Table 1 Pathological features and FOXP3 expression in ATLL
Full size table

EBV Infection

EBER in situ hybridization was used to analyze 169 cases (60 cases of FOXP3(+) and 109 of FOXP3(−)) (Table 1 and Figure 1). Cells with definitive EBV infection (EBER-positive cells) were detected in 35 cases (21%), while 38% (23/60) of FOXP3(+) ATLL cases and only 11% (12/109) of FOXP3(−) ATLL cases showed EBV-infected cells. Finally, EBV-infected cells were significantly more frequently detected in the tissue of FOXP3(+) ATLL than in that of FOXP3(−) ATLL cases (P<0.0001).

Clinical Features

The clinical findings are summarized in Table 2. The 96 cases enrolled in our study comprised 34 FOXP3(+) and 62 FOXP3(−) cases, of which 40 cases were lymphomatous type and 56 acute type. There were 42 females and 54 males aged 27–90 years with a mean age of 63 years. The median age of FOXP3(+) cases was relatively lower than that of FOXP3(−) cases (58 and 64 years, respectively). The ratio of patients over age 60 years was relatively smaller for FOXP3(+) (47%) than for FOXP3(−) cases (68%) (P=0.09), and FOXP3(+) cases showed female predominance but FOXP3(−) cases did not (59 vs 35%; P<0.05). However, other factors (clinical stage, clinical type (lymphoma and acute type), serum LDH and calcium elevation and extranodal site involvement including peripheral blood and bone marrow) did not show any significant differences between FOXP3(+) and FOXP3(−) cases. Therapy based on clinical data was administered to 94 patients, all of whom received chemotherapy (CHOP-like regimen) and 20 cases received radiation therapy. Nine cases received stem cell transplantation (SCT). Five patients received allogeneic SCT and four patients received autologous SCT. Among the nine cases who received SCT, five were FOXP3(+). Relatively more FOXP3(+) than FOXP3(−) cases received SCT (P<0.05). Overall survival shown in Figure 2 demonstrates no significant difference between the two groups.

Table 2 Clinical features of FOXP3(+) ATLL and FOXP3(−) ATLL
Full size table
Figure 2
figure 2

Overall survival (OS) of FOXP3(+) and FOXP3(−) ATLL (n=96). There was no significant difference in OS.

Full size image

Immunodeficient Cases

During the clinical course, 10 of the 96 cases showed definitive immunodeficient status, as shown in Table 3, with a significantly higher probability for FOXP3(+) cases (8/34; 24%) than for FOXP3(−) cases (2/62; 3%) (P<0.005). Only one patient received autologous SCT, but other nine cases were treated with only chemotherapy without SCT. This suggests that immunodeficient status in these 10 patients was not induced by intensive therapy.

Table 3 Ten cases suspected as immunosuppressive state
Full size table

Cytogenetic Analysis

All 74 patients, comprising 26 FOXP3(+) and 48 FOXP3(−) cases, showed abnormal karyotypes (Table 4; the Table containing the detailed cytogenetic findings can be obtained on request). Aneuploidy was observed in 18 FOXP3(+) cases (16 hyperdiploid, 1 hypodiploid, 1 nearly tetraploid) and in 44 FOXP3(−) cases (28 hyperdiploid, 10 hypodiploid, 5 nearly triploid, 1 nearly tetraploid). Aneuploidy was thus found significantly more frequently in the FOXP3(−) than in the FOXP3(+) type (P<0.05). No specific recurrent abnormality was detected in either FOXP3(+) or FOXP3(−) cases. Simple numerical abnormality without structural abnormality was observed in 22 FOXP3(+) cases and 43 FOXP3(−) cases. The average number of numerical abnormalities was significantly higher in FOXP3(−) cases (6.8) than in FOXP3(+) cases (2.8) (P<0.0001). Structural abnormalities were detected in 71 cases, while only 3 cases were free from structural abnormalities, and all of them were negative for FOXP3. The average number of breaks in FOXP3(+) and FOXP3(−) cases was not significantly different (7.3 vs 8.6, P=0.35). We focused on abnormalities on Xp11–q13, where the FOXP3 gene is located, but only two cases, one FOXP3(+) and one FOXP3(−), showed amplification of Xp11–q13 and only one FOXP3(−) case showed translocation of Xp11–q13 (data not shown).

Table 4 Cytogenetic features and FOXP3 expression in ATLL
Full size table

Discussion

This study shows for the first time that FOXP3 expression in ATLL is associated with a clinically and pathologically immunodeficient state. About 36% of the ATLL cases enrolled in this study were positive for FOXP3, indicating that ATLL can be divided into two types: FOXP3(+) and FOXP3(−). The percentage of FOXP3(+) type was consistent with previously reported ratios22, 23, 24 and confirms the heterogeneity of ATLL.

Various reports have described the phenotypic features of Tregs except for CD4 and CD25. These include cytotoxic T-lymphocyte-associated antigen 4, glucocorticoid-induced tumor necrosis factor receptor family-related gene and CD103 (αE integrin).35, 36, 37 Recently, Iellem et al34 reported that CCR4 and CCR8 were relatively specifically expressed on Tregs, and they speculated that these chemokine receptors may guide Treg cells to sites of antigen presentation in secondary lymphoid tissues and inflamed areas to attenuate T-cell activation.34 CCR4 and CCR8 expression on ATLL cells has also been reported in association with skin invasion38 or autocrine antiapoptotic stimulation of ATLL cells.39 We could not detect any correlation between CCR8 expression and FOXP3 expression, but FOXP3(+) cases expressed CCR4 relatively more frequently, although not significantly so, than did FOXP3(−) cases and this may reflect the nature of Tregs in FOXP3(+) ATLL. Ishida et al21 found that the CD4+CD25+CCR4+ phenotype is selectively associated with normal Tregs.

EBV-positive LPDs are often associated with an immunodeficient state. AIDS-related LPD, post-transplant LPD and senile EBV-positive LPD are representative immunodeficient states related to EBV-positive LPD.40, 41 In healthy carriers, EBV-positive cell proliferation is well regulated by EBV-specific cytotoxic T lymphocytes (CTLs). In the immunodeficient state, however, EBV-infected cells proliferate with a decline in EBV-specific cytotoxic T-cell functions.42 Actually, regression of post-transplant lymphoproliferative disorder may occur in some cases if immunosuppression is reduced or the patient has been infused with EBV-specific CTLs generated from the original allogeneic donor.43, 44 In Hodgkin's lymphoma, Hodgkin and Reed–Sternberg (HRS) cells are often infected with the EBV, and regulatory T-cell infiltration around HRS cells coincides with the suppression of latent membrane antigen-specific CD8+ T-cell function and helps HRS cells to survive.45, 46 It is well known that EBV-infected cells, the majority of them B cells, emerge in peripheral T-cell lymphoma, including ATLL.1, 32, 47, 48 With reference to the new WHO classification, Kikuchi et al1 described the expansion of EBV-infected cells as a secondary event subsequent to underlying immunodeficiency in ATLL patients. Katsuki et al6 previously reported that T-cell control of EBV-infected B cells was impaired in healthy HTLV-1 carriers. The higher frequency of EBV-positive cells in FOXP3(+) than in FOXP3(−) cases indicates that FOXP3-positive lymphoma cells may have a stronger suppressive effect on the T-cell control of EBV-positive cells and thus induce their proliferation.

Our cytogenetic analysis did not detect specific recurrent abnormalities, similar to previous studies.49, 50 However, FOXP3(−) cases included more cases with aneuploidy and showed more numerical chromosome abnormalities than FOXP3(+) cases. In serially studied ATL patients, only normal or simple abnormal karyotypes were detected at diagnosis, while complex karyotypes appeared later at relapse.51, 52 These findings are compatible with the hypothesis of multistep carcinogenesis in ATLL.53 Therefore, the occurrence of relatively simpler cytogenetic abnormalities in FOXP3(+) cases may indicate a relatively earlier stage in multistep carcinogenesis than in FOXP3(−) cases. This implies that FOXP3 expression may be lost with disease progression, and it is explicable for the different FOXP3 expression among each of the cases.

Clinically, there was no significant difference in prognosis between FOXP3(+) ATLL and FOXP3(−) ATLL. Roncador et al22 reported that FOXP3(+) ATLL showed a more aggressive clinical course than FOXP3(−) ATLL, but that the difference was not significant. While our study analyzed 96 ATLL cases, many more than those enrolled in previous studies, no significant difference in overall survival rate between the two types of ATLL could be identified. FOXP3(+) ATLL cells may suppress tumor immunity and promote tumor growth, but other factors, especially relatively younger age and more cases with SCT than in the case of FOXP3(−), may compensate for some disadvantages (eg, immunodeficient status) and improve the prognosis for FOXP3(+) ATLL.

The mechanisms of immunosuppression in ATLL have been the subject of many discussions. The immunodeficient state is also seen in HTLV-1 carriers and patients with HAM/TSP without malignancies and is associated with infections caused by various fungi, viruses and protozoa.4, 6, 54 In one experiment, inoculation of rabbits with a cell line infected with HTLV-1 resulted in thymic atrophy, which was found to be associated with infiltration of HTLV-1-infected cells into the thymus.55 Impaired production of naïve T lymphocytes in HTLV-1-infected individuals was reported recently.56 The results of our study indicate that ATLL cells may function as Treg-like cells and induce the immunosuppressive state described above, especially in FOXP3(+) cases.

In conclusion, FOXP3 expression was found to be associated with an immunosuppressive clinical state, proliferation of EBV-positive cells in tissue, relatively simple cytogenetic changes and smaller lymphoma cells, and especially with an exclusively anaplastic morphology. These findings disclose that FOXP3 expression is meaningful rather than aberrant and reflects specific biological functions in ATLL.