Lymphoma of the gastrointestinal tract is the most common extranodual tumor, accounting for 4%－20% of all non-Hodgkin's lymphomas.［1,2］ Approximately 9% gastrointestinal lymphomas are intestinal T-cell lymphomas (ITCLs),［3］ which were described as a form of malignant histiocytosis of the intestine.［4］ Recent studies have suggested that ITCL is a heterogeneous lymphoid neoplastic group with variable clinical and pathological features.［5］ However, enteropathy-type intestinal T-cell lymphoma (ETCL) defined by the WHO classification is rare in oriental countries.［6］ Patients with ITCL in these countries always have neither a history of malabsorption nor proven celiac disease, but have been associated with Epstein-Barr virus (EBV).［7-19］ Up to the present, the difference between ETCL and ITCL in Chinese has not yet been reported. In this retrospective study, the clinicopathologic features, immunophenotype, the expression of cytotoxic molecule (T-cell-restricted intracellular antigen-1, TIA1), T-cell receptor (TCR)-γ gene rearrangement and EBV latent infection in 42 ITCL patients without celiac disease in Chinese were observed to investigate the clinicopathologic features of ITCL and its association with EBV.
Forty-seven patients were retrieved from pathologic files from July 1980 to January 2001 of the Department of Pathology of West China Hospital of Sichuan University. Of the 47 patients, 24 were reported previously by us.［8］ Excluding 2 patients without typical lesions and 3 patients without paraffin blocks, 42 patients were finally studied with specimens, including operative specimens from 37 patients and biopsied specimens from 5 patients.
Histological and immunophenotypic diagnosis
Four μm thick sections from paraffin-embeded tissue blocks were stained with haematoxylin and eosin. Primary ITCLs were defined by the updated Kiel classification and the criteria of Chott.［20］ All clinical data were reviewed. Immunophenotyping was performed on paraffin sections by the LsAB(SP) method. A panel of primary antibodies to T-cell differentiated antigens (CD45RO, CD3ε, CD4, CD8) (dilution 1∶200; Dakopatts, Denmark), to B-cell differentiated antigen (CD20; dilution 1∶200; Dakopatts, Denmark), to histiocyte associated antigen (CD68; dilution 1∶200; Dakopatts, Denmark), to NK-cell associated antigen (CD56; dilution 1∶50; Monosan, the Netherlands), to cytotoxic granular protein (TIA-1; dilution 1∶200; Culter, USA), and to the products of EBV genome (LMP-1; dilution 1∶50; Dakopatts, Denmark) were used. The SP kit and basic DAB detection system (Dakopatts, Denmark) were used for the rest of the procedures. Reactive lymphoid tissue and a case of Hodgkin lymphoma previously demonstrated to be positive for LMP-1 were used as positive control. For negative control, the primary antibody was substituted with PBS.
Polymerase chain reaction (PCR)
Paraffin tissue DNA was prepared by the tissue DNA extraction and purification kit (DneasyTM Tissue Kit，Qiagene). With β-globulin gene being amplified as control, three pairs of TVG/TJX、Vγ11/Vγ101/Jγ12 and Vγ11/Vγ101/Jp12 primers (Gibco-BRL, USA) were used to amplify the rearranged TCR-γ gene.［21,22］ The sequence and reannealing temperature were shown in Table . A total of 0.5 μmmol/L of each oligonucleotide primer was used per 25 μl reaction. The PCR reaction mixture also contained 0.5 μg template DNA, 1.25 U of Taq DNA polymerase, 220 μmmol/L of dNTP and 1×buffer (supplied with 15 mmol/L MgCl2). The samples were denatured for 6 minutes at 95℃, and subjected to 40 cycles of amplification in a DNA thermal cycler (Perkin-Elmer). One cycle represented primer extension for 1 minute at 72℃, denaturation for 1 minute at 95℃, and reannealing for 1 minute followed by primer extension as mentioned above. Amplified DNA was analyzed by electeophoresis in 8% polyacrylamide gels and visualized by staining with ethidium bromide. The gels were subsequently displayed by the gel displaying system (Gibco-BRL, USA). Human reactive lymphoid tissue DNA and human T cell lymphoma cell line Peer DNA were used as polyclonal and monoclonal control, respectively. Each experiment included reactions without DNA to identify any contamination.
In situ hybridization (ISH)
In situ hybridization was carried out with EBERs (EBER1 and EBER2) oligonucleotide probes labelled with fluorescein isothiocyanates (FITC) (Y017, Dakopatts, Denmark) or digoxin (donated by Prof. ZHOU Xiao-ge, Friendship Hospital, Beijing). For the former, rabbit anti-FITC antibody conjugated with alkaline phosphatase (AP) (Dakopatts, Denmark) was used to combine with the probe, while NBT/BCIP was used as a substrate. After hybridized by the digoxin-labelled probe, paraffin sections were stained by the LsAB method (described former) with mouse anti-digoxin antibody (Dakopatts, Denmark) being used as the primary antibody. EBV positivity located in the nucleus. The former, royal blue color and the latter, buffy color. The number of positive cells was visually estimated at <1, 1 to 25, 25 to 100, and >100 per medium power field (mpf) by a 10× ocular lens and a 20× objective. Therefore, the positivity patterns could be divided into four types: single-cell-, scatter-, cluster- and diffuse-type. The case in which positive reaction presented as cluster- or diffuse-type was regarded as EBV-associated lymphoma.［23］ A case of lung metastatic NK/T-cell lymphoma previously demonstrated to harbor EBV was used as positive control. For negative control, the probe was substituted with diethyl pyrocarbonate (DEPC).
Overall survival time was measured from the date of diagnosis to the date of death or the last follow-up. The Kaplan-Meier method was used to generate overall survival curves, which were compared for statistical difference with the log-rank test (SPSS 10.0). A P value of less than 0.05 was considered statistically significant.
The tumors were localized from the duodenum to the rectum, while the most common sites were the ileocecum and colon. Less commonly, the ileum and rectum were involved. The tumors in more than one site were found in 22 patients (52.4%). The tumors as ulcers were seen in 35 patients (83.3%), and the ulcers varied in shape or value. Mesentery lymphadenopathy was observed in 14 patients (33.3%). Histologically, 42 ITCLs were classified into pleomorphic median-sized and large cells (n=38, 90.5%) ( Fig. 1 ), monomorphic median-sized cells (n=2, 4.8%), and pleomorphic small cells (n=2, 4.8%). The lesions showed ulceration with frequent angiocentricity and angioinvasion (28/42, 66.7%), associated with zonal necrosis (27/42, 64.3%). Most commonly, they were a mixture of median-sized cells and large cells. The cells had irregular nuclei with granular chromatin. Their nucleoli were usually inconspicuous or small with moderate and pale cytoplasm. Mitotic changes were easily found. Features of lympho-epithelial lesion (LEL) and reactive histiocytosis with phagocytosis were observed in most patients. Except villous atrophy was observed in 4 patients, the other histological features of enteropathy such as increase of intraepithelial lymphocytes and crypt hypertrophy were not found in any patients.
In the 42 patients with ITCL revealed CD45RO positive ( Fig. 2 )， 28 (66.7%) expressed CD3ε, 10 (23.8%) expressed CD8，7 (16.7%) expressed CD4, 12 (28.6%) expressed CD56, and 39 (92.9%) expressed TIA-1, but none expressed CD20 and CD68. Half of the 42 patients presented the same immunophenotype CD45RO+CD4―-CD8-CD56-. The expression frequency of LMP-1 was 38.1% (16/42). Scattered LMP-1 positive tumor cells were presented.
TCR-γ gene rearrangement detection
Positive bands with the β-globulin PCR (110 bp) were obtained in all patients with ITCL, indicating that adequate DNA was present in the 42 patients. In these 42 patients, 31 (73.8%) exhibited specific bands of TCR-γ gene rearrangement ( Fig. 3 ), in which 27 (64.3%), 27 (64.3%) and 25 (59.5%) were detected by TVG/TJX, Vγ11/Vγ101/Jγ12, and Vγ11/Vγ101/Jp12 primers, respectively.
In situ hybridization for EBER1/2
EBER1/2+ITCLs were identified in 36 patients or 85.7% of the 42 patients with ITCLs. The EBER1/2 positivity was located in the nucleus and appeared as cluster-type (21 patients) or diffuse-type (15 patients) ( Fig. 4 ). The distribution of EBER1/2 positive cells coincided with that of tumor cells.
The patients of this series included 33 men and 9 women with a mean age of 30.4 years (range 10－60 years). The male to female ratio was 3.67∶1.
Their most common symptoms were abdominal pain, hemotochezia, fever, weight loss, and bowel perforation with frequencies of 76.3%, 64.3%, 69.1%, 57.1% and 42.9%, respectively. None of the patients were immunosuppressed at the time of diagnosis. All the patients had neither malabsorption nor proven celiac disease before presentation. Data of treatment were obtained for 39 patients (92.9%)， in whom 10 received bowel resection and combined chemotherapy (CHOP or COP or VCP), one received bowel resection, combined chemotherapy and radiotherapy, 26 received bowel resection only, and 2 abandoned therapy. Follow-up information was obtained for 30 patients (76.9%). One patient who died at the operation was excluded. Six patients who had tumor in one site survived within 156 months (13 years). Twenty-three patients died and 4 patients relapsed within 24.3 months after diagnosis. The overall median survival was 3 months. Kaplan-Meier survival curves showed that the 1- and 2-year survival rates were 30% and 22%, respectively ( Fig. 5 ). Survival curves were plotted for single lesion or multi lesions according to whether there were hemotocheziab, fever, bowel perforation, lymph node metastasis, tumor cells expressed CD4, CD8, CD56, TCR-γ gene rearrangement, and bowel resection combined with chemotherapy or radiotherapy. The curves were compared by the log-rank test. No statistical difference was found except TCR-γ gene rearrangement and different therapy (P=0.0078, P=0.0250, respectively).
Among peripheral T-cell lymphomas, enteropathy-type intestinal T-cell lymphoma (ETCL) is regarded as an entity in the updated WHO classification.［6］ However, ITCLs in oriental countries are very different from those in clinicopathological features and etiology.［7-19］
Clinically, ETCLs in western countries occur with equal frequency in men and women aged on average from 56 to 60.1 years. About 49% of the patients have a history of celiac disease, which is localized mostly in the jejunum. The patients often complain of abdominal pain, weight loss, and intestinal obstruction. Only 3% of the patients present with hemotochezia. Pathologically, ulcers or thickened plaques can be seen in 44% of the patients. Most patients (70%－81%) show pleomorphic median-sized or large cells. Tumor cells are CD4-CD8-CD56- in 63% patients.［5,18,20］
In China, patients with ITCL are commonly seen in young men with fever and hemotochezia. The lesions are mostly localized at the ileocecum and colon. Tumor cells could express CD4, CD8, CD56 and TIA1. Histological features of celiac disease such as increase of intraepithelial lymphocytes and crypt hypertrophy are not observed in any patients.As to the prognosis, ETCL has a poor prognosis with a mortality rate of 84.0% and a overall median survival of 4.0 months. The 1- and 5-year survival rates are 38.7% and 19.7%, respectively. The 1- and 5-year survival rates in patients without symptoms are 19.4% and 3.2%, respectively.［24］ Although a patient can achieve a remission after 6 courses of CHOP chemotherapy in oriental countries,［25］ 6/7 patients may die within 3 months.［11］ Li et al［7］ showed that the 5-year survival rate was 13% and the overall median survival time was 9 months. In the 30 patients who were followed up, 23 died and 4 relapsed within 24.3 months after the diagnosis. The overall median survival time was 3 months. The 1- amd 2-year survival rates were 30% and 22% respectively, suggesting the poor prognosis of ITCL in China. Hsiao et al［11］ reported that clinical stage and bowel perforation are significant factors for prognosis. We found that the outcome was better in patients with TCR-γ gene rearrangement than in those without TCR-γ gene rearrangement and that in the patients who received bowel resection combined with chemotherapy or radiotherapy than in those underwent bowel resection only. A high frequency of TIA-1 in ITCL, however, indicates that the aggressive course and poor prognosis are attributable to the cytotoxic function of tumor cells.
The positive frequency of EBV in ETCL in western countries was as low as 6.3%－36%,［5,6］ whereas EBER1/2 in ITCL was detected in 4/7 cases,［11］ 6/11 cases［17］ in Taiwan and Japan respectively, EBER1/2+ in gastrointestinal lymphoma in 9/13 cases［15］ and EBER1/2+ in intestinal lymphoma in 6/6 cases［16］ in Korea. In 43 cases of primary gastrointestinal T-cell lymphoma in Australia, 17/29 were detected the EBV gene, and in 5/38 cases positive for EBER-ISH, 2 were positive for LMP-1.［18］ In 20/24 (83.3%) ITCLs positive for EBER1/2-ISH, four were positive for LMP1.［8］ In our series, EBV infection was detected in 41 (97.6%) of the 42 patients. That EBER1/2 positivity appeared as cluster-type (21 patients) or diffuse-type (15) indicated that all 36 cases of EBER1/2+ were EBV-associated lymphoma. Kanavaros et al［26］ found that EBV infection probably took place before the clonal proliferation of tumor cells in EBV-associated lymphoma, and the virus plays a central role in start-up and/or promotion in pathogenesis of the tumor. The fact awaits further study of the clonality of EBV in ITCL.
LMP-1 is a trans-membrane protein, closely associated with the intermediate filament vimentin of the cytoskeletal network. LMP-1 has properties characterized by an oncogenic protein which induces morphological transformatin in human keratinocytes and inhibits differentiation of epithelial cells.［27］ In our series, 16 (38.10%) of the 42 patients presented LMP-1 positive reaction as was reported by de Bruin［21］ and Pan (2/5, 2/4).［18,28］ The discrepancy of detection between LMP-1 and EBER1/2 may be due to the low frequency of LMP-1 expression, different EBV subtypes or different EBV infection patterns. We found that LMP-1 positive tumor cells were focally scattered. Whether the biological properties of different tumor cell subclones result in the diversity of LMP-1 expression is unclear. The interaction between LMP-1 expression and the biological status of host cell works is uncertain. In this interaction, the imbalance of EBV copy at a low level and immune reaction, hyperplasia of EBV, host cell transformation, and the pathogenesis of ITCL are worth investigating.
To our knowledge, not only ETCL ,but also extranodal NK/T-cell lymphoma of nasal type could involve the intestine.［6］ Clinically, the latter mainly involves middle-aged men with major symptoms of bleeding, ulcer and bone destroy.［29］ Follow-up data information from Tapan disclosed that extranodal NK/T-cell lymphoma of nasal type has a poor prognosis with a median survival of 3 months.［30］ These features resemble to those of the cases of this series. Pathologically, pleomorphic median-sized or large cells were predominant. Most lesions in this series showed ucleration with frequent angiocentricity and zonal necrosis. Features of lympho-epithelial lesion (LEL) and reactive histiocytosis with phagocytosis were frequently observed, and were similar to the pathological changes of extranodal NK/T-cell lymphoma of nasal type.［29］ Comparison of the immunophenotype and expression of TIA1 showed that their difference only lies in the expressions of CD8 and CD56.［31-33］ Another important feature of ITCL is the frequency of TCR-γ gene rearrangement (59.52%－64.29%). Thus, a small proportion of ITCLs have a cell linage of cytotoxic T cell or NK cell. The positive frequency of EBER and LMP1 in extranodal NK/T-cell lymphoma of nasal type is 83.78%－93% and 26.32%, respectively,［31,32］ and 85.71% and 38.10% in ITCL. Hence ITCL is a heterogeneous lymphoid neoplastic group, and a small proportion of primary ITCLs in China and extranodal NK/T-cell lymphoma of nasal type belong to the same spectrum.
In short, intestinal T-cell lymphoma ( ITCL) without celiac disease in Chinese is a special EBV-associated entity group. Their peculiarity lies in two aspects: unusual clinicopathological features such as predilection people, age, site, histological changes and expression of CD56 and TIA-1; and relation to Epstein-Barr virus.
Acknowledgement ： We are grateful to Dr. WEI Yu-quan for his analysis of PCR product and Prof. ZHOU Xiao-ge for providing EBERs oligonucleotide probes labeled with digoxin.
1.Greiner TC, Medeiros LJ, Jaffe ES. Non-Hodgkin's lymphoma. Cancer 1995;75(suppl):370-380.
2.Crump M, Gospodarowicz M, Shepherd FA, et al. Lymphoma of the gastrointestinal tract. Semin Oncol 1999；26:324-337.
3.Franssila KO, Jaser N, Sivula A. Gastrointestinal non-Hodgkin's lymphoma. A population-based clinico-pathological study of 111 adult cases with a follow-up of 10-15 years. APMIS 1993;101:631-641.
4.Isaacson PG, Spencer J, Connolly CE, et al. Malignant histiocytosis of the intestine: a T-cell lymphoma. Lancet 1985;11:688-691.
5.Domizio P, Owen RA, Sheperd NA, et al. Primary lymphoma of the small intestine. A clinicopathological study of 119 cases. Am J Surg Pathol 1993;17:429-442.
6.Jaffe ES, Harris NL, Stein H, et al. World Health Organization classification of tumor: Pathology and genetics of tumors of haematopoietic and lymphoid tissues. International agency for research on cancer (IARC) Press, Lyon 2001;274-279.
7.Li G, Ouyang Q, Liu K, et al. Primary non-Hodgkin's lymphoma of the intestine: a morphological, immunohistochemical and clinical study of 31 Chinese cases. Histopathology 1994;25:113-121.
8.Ren X, Liu W, Li G, et al. Epstein-Barr virus infection and expression of T-cell intracellular antigen-1 (TIA-1) in intestinal T-cell lymphoma. Chin J Pathol (Chin) 1999;28:348-351.
9.Chan AC, Ho JW, Chiang AK, et al. Phenotypic and cytotoxic characteristics of peripheral T-cell and NK-cell lymphomas in relation to Epstein-Barr virus association. Histopathology 1999;34:16-24.
10.Chott A, Haedicke W, Mosberger I, et al. Most CD56+ intestinal lymphomas are CD8+CD5- T-cell lymphomas of monomorphic small to medium size histology. Am J Pathol 1998;153:1483-1490.
11.Hsiao CH, Lee WI, Chang SL, et al. Angiocentric T-cell lymphoma of the intestine: a distinct etiology of ischemic bowel disease. Gastroenterology 1996;110:985-990.
12.Lee SS, Jang JJ, Cho KJ, et al. Epstein-Barr virus-associated primary gastrointestinal lymphoma in non-immunocompromised patients in Korea. Histopathology 1997;30:234-242.
13.Kim YS, Paik SR, Kim HK, et al. Epstein-Barr virus and CD21 expression in gastrointestinal tumors. Pathol Res Pract 1998;194:705-711.
14.Yang WI, Cho MS, Tomita Y, et al. Epstein-Barr virus and gastrointestinal lymphomas in Korea. Yonsei Med J 1998;39:268-276.
15.Lee YS, Kang CS, Kim BK, et al. Extranasal T/NK-cell lymphoma presenting as intestinal diverticulum. J Korean Med Sci 2000;15:229-232.
16.Huh J, Cho K, Heo DS, et al. Detection of Epstein-Barr virus in Korean peripheral T-cell lymphoma. Am J Hematol 1999;60:205-214.
17.Katoh A, Ohshima K, Kanda M, et al. Gastrointestinal T cell lymphoma: predominant cytotoxic phenotypes, including alpha/beta, gamma/delda T cell and natural killer cells. Leuk Lymphoma 2000;39:97-111.
18.de Bruin PC, Jiwa NM, Oudejans JJ, et al. Epstein-barr virus in primary gastrointestinal T cell lymphomas: association with gluten-sensitive enteropathy, pathological features, and immunophenotype. Am J Pathol 1995;146:861-867.
19.Quintanilla-Martinez L, Lome-Maldonado C, Ott G, et al. Primary non-Hodgkin's lymphoma of the intestine: high prevalence of Epstein-Barr virus in Mexican lymphomas as compared with European cases. Blood 1997;89:644-651.
20.Chott A, Vesely M, Simonitsch I, et al. Classification of intestinal T-cell neoplasms and their differential diagnosis. Am J Clin Pathol 1999;111(1 Suppl 1):S68-S74.
21.Benhattar J, Delacretaz F, Martin P, et al. Improved polymerase chain reaction detection of clonal T-cell lymphoid neoplasms. Diagn Mol Pathol 1995;4:108-112.
22.McCarthy KP, Sloane JP, Kabarowski JH, et al. A simplified method of detection of clonal rearrangement of the T-cell receptor-γchain gene. Diagn Mol Pathol 1992;1:173-179.
23.Weiss LM, Jaffe ES, Liu XF, et al. Detection and localization of Epstein-Barr viral genomees in angio-immunoblastic lymphadenopathy and angio-immunoblastic lymphadenopathy-like lymphoma. Blood 1992;79:1789-1795.
24.Gale J, Simmonds PD, Mead GM, et al. Enteropathy-type intestinal T-cell lymphoma: clinical features and treatment of 31 patients in a single center. J Clin Oncol 2000;18:795-803.
25.Widjaja A, Wagner S, Mix H, et al. Malabsorption with progressive weight loss and multiple intestinal ulcers in a patient with T-cell lymphoma. Z Gastroenterol 1999;37:611-614.
26.Kanavaros P, Briere J, Emile J-F, et al. Epstein-Barr virus in T and natural killer(NK) cell non-Hodgkin's lymphomas. Leukemia 1996;10 Suppl 2:S84-S87.
27.Gregory CD, Dive C, Henderson S, et al. Activation of Epstein-Barr virus latent genes protects human B cells from death by apoptosis. Nature 1991;349:612-614.
28.Pan L, Diss TC, Peng H, et al. Epstein-Barr virus in enteropathy-associated T-cell lymphoma (ETCL). J Pathol 1993;170:137-143.
29.Lorenzen J, Liu WP, Li GD, et al. Nasal T/NK cell lymphoma: a clicicopathologic study of 30 west Chinese patients with special reference to proliferation and apoptosis. Leuk Lymphoma 1996;23:593-602.
30.Harabuchi Y, Imai S, Wakashima J, et al. Nasal T-cell lymphoma causally associated with Epstein-Barr virus: clinicopathologic, phenotypic, and genotypic studies. Cancer 1996;77:2137-2149.
31.Liu W, van Gorp J, Li G. Epstein-Barr virus infection in midline malignant reticulosis. Chin J Oncol (Chin) 1997;19:49-52.
32.Peh SC, Quen QW. Nasal and nasal-type natural killer (NK)/T-cell lymphoma: immunophenotype and Epstein-Barr virus (EBV) association. Med J Malaysia 2003;58:196-204.
33.Liu W, Chan AC, Ho FC. Expression of cytotoxic-granule-associated protein TIA-1 in nasal NK/T cell lymphomas and lymphoid hyperplasia. Chin J Pathol (Chin) 1998;27:247-250.