Expression of survivin and of antigen detected by a novel monoclonal antibody, T332, is associated with outcome of diffuse large B-cell lymphoma and its subtypes
Although diffuse large B-cell lymphoma (DLBCL) is the most common type of non-Hodgkin lymphoma, it is both clinically and morphologically heterogenous. The present study investigates the significance of survivin and a novel monoclonal antibody (MAb), T332,immunohistochemically for predicting the prognoses of DLBCL and its subtypes classified as germinal center B-cell-like type (GCB) and non-GCB type (NGCB) based on the expression profiles of CD10, bcl-6, and MUM1. A total of 60 cases of DLBCL (GCB, n = 22; NGCB, n = 38) were examined for the expression of survivin and T332 antigen. Survivin+ DLBCL had a signifi- cantly worse prognosis (P = 0.01) than survivin– cases, as already reported, while survivin+ GCB or NGCB tended to have poor prognoses (P = 0.06 and 0.07, respectively). How- ever, T332+ DLBCL and NGCB had significantly more unfa- vorable prognoses than T332– cases (P = 0.01 and 0.02, respectively) while there was no significant survival differ- ence between the T332+ and T332– groups of GCB (P = 0.11). Interestingly DLBCL coexpressing survivin and T332 (n = 13) had a significantly worse prognosis (P = 0.009) than the remaining single positive and double negative cases (n = 31). In conclusion, survivin and the novel MAb, T332, might be a good predictor of DLBCL and its subtypes.
Key words: DLBCL, germinal center B cell, non-germinal center B cell, predictor, prognosis, survivin, T332
The current WHO classification for lymphomas is principally based on the understanding of the pathogenesis of each neoplasm and is a significant indicator to determine appro- priate therapy and to predict prognosis.1 However, diffuse large B-cell lymphoma (DLBCL), which causes approximately 33.3% of all 3194 lymphoma cases in Japan,2 is still defined by morphological findings such as cell size and growth pattern, without phenotypic and genotypic informa- tion. Thus, DLBCL is likely to include more than one disease entity, as suggested by the marked variability of clinical pre- sentation and response to treatment. Indeed, the heteroge- neity of DLBCL was recently demonstrated using a cDNA array,3,4 showing that DLBCL could be subdivided at least into germinal center (GC) B-cell-like, activated B-cell type, and type 3.4 However, such a cDNA array is not yet applicable to routine examinations in many medical facilities. Thus, it is necessary to develop alternative method(s) to classify the subtypes of DLBCL. Meanwhile, it was reported that routine immunohistochemistry (IHC) was able to classify DLBCL into GC B-cell like (GCB), and non-GC B-cell type (NGCB), and the GCB phenotype had favorable prognosis than NGCB.5
The purpose of the present study was to investigate the utility of survivin and our novel monoclonal antibody (MAb), T332, in IHC to predict the prognoses of patients with DLBCL and its subtypes, because both antigens are expressed in dark-zone B-cells of GC and are considered to be associated with the outcome of DLBCL and its subtypes.
MATERIALS AND METHODS
Case selection and histological diagnosis
Sixty patients participated in the present study, which was based on a prospective, consecutive entry design, between October 1995 and December 2001, and were diagnosed as having DLBCL according to the Working Formulation.6 There- after these 60 patients were reclassified based on the new WHO classification.1 Follicular large cell lymphoma, Richter’s syndrome, and blastic transformation of marginal zone B-cell lymphoma were excluded because the aim of the present study was to clarify the outcome of de novo DLBCL. Patients under 70 years old were assigned to receive eight cycles of cyclophosphamide, doxorubicin, vincristine, and predniso- lone (CHOP) or pirarubicin, cyclophosphamide, vincristine, and prednisolone (THP-COP therapy7), while patients aged 70 and over were assigned to receive six cycles of THP-COP therapy. The CHOP and THP-COP chemotherapy cycles were repeated at 14 day intervals in patients aged under 70, while the THP-COP chemotherapy cycles were repeated at 21 day intervals in elderly patients aged 70 and over, using granulocyte colony-stimulating factor. Patients with a bulky tumor mass received radiotherapy ranging from 30 Gy to 40 Gy after chemotherapy. Patients who relapsed or had disease progression after CHOP or THP-COP, and patients who were resistant to CHOP or THP-COP received methyl- prednisolone, ifosfamide, methotrexate, etoposide, and car- boplatin (the P-IMVP-16/CBDCA regimen8), as a second-line regimen. Some patients under 70 with refractory or relapsed non-Hodgkin lymphoma (NHL) who responded to P-IMVP- 16/CBDCA received high-dose chemotherapy followed by autologous hematopoietic stem cell transplantation.
Production of a novel MAb
A novel mouse MAb (clone, T332: isotype, IgG2a) was pro- duced using the standard hybridoma constructing method.9 Briefly, BALB/c mouse (Japan Charles River, Yokohama, Japan) was immunized three times with human tonsilar lym- phoid cells, then the splenocytes were hybridized with mouse myeloma cell, NS-1 (American Type Culture collections, Manassas, VA, USA), using polyethylene glycol method. The culture supernatants from each culture well were immunohis- tochemically screened on cryosections of tonsil to select the antibody that would specifically react with germinal center B cells, and the MAb, T332 was established by the limiting dilution method.
Immunohistochemical analysis
The tissues obtained were snap-frozen in liquid nitrogen and/ or were fixed in 10% formalin fixative. The latter samples were embedded in paraffin within the next 24 h. The frozen tissues were stored at -80∞C until use. The phenotypic profile of each lymphoma case was clarified by IHC using the avidin–biotin– enzyme complex (ABC) method or catalyzed signal amplifi- cation (CSA) system (DakoCytomation, Glostrup, Denmark) on paraffin or cryostat sections as indicated in the manufac- turer’s instructions. Briefly, first and second antibodies were reacted for 1 h in a humidifier, followed by ABC in the ABC method. For the CSA system, amplification reagent then streptavidin–peroxidase in staining kit were reacted after the
second biotinylated antibody. Phosphate-buffered physiolog- ical saline (PBS) was used for the intervening wash. Double immunofluorescence was carried out using fluorescein isothiocyanate (FITC)-labeled CD3 and biotinylated MAb, T332 followed by phycoerythrin-labeled streptavidin (BD PharMingen, San Diego, CA, USA). As a negative control, primary antibody was added to appropriately diluted normal mouse or rabbit sera. The primary MAb used were as follows: FITC-labeled CD3 (BD PharMingen), CD8, CD20, CD21, CD30, CD43, CD45RO, bcl-2, Ki-67, MUM1 (DakoCytomation), CD5, CD10, cyclin D1 (MBL, Nagoya, Japan), survivin (Santa Cruz Biotechnology, Santa Cruz, CA, USA), TdT, bcl-6, CD4, CD56, p80 (Novocastra, Newcastle upon Tyne, UK), and rabbit anti-CD3 (DakoCytomation). The following reagents used for ABC method were biotinylated goat anti- mouse immunoglobulin (Ig), and biotinylated swine antirabbit Ig as secondary antibody, and Strept ABComplex/horseradish peroxidase (HRP) (DakoCytomation). In the case of bcl-6 and survivin, the CSA system was applied. Finally the reactions were visualized by 3,3¢-diaminobenzidine tetrahydrochloride (DAB) Liquid System for HRP (DakoCytomation). Each case was determined as positive when the definite positive reac- tions were observed in >30% of tumor cells.5,10 In the case of survivin, nuclear staining was defined as positive.
Subtypes of diffuse large B-cell lymphoma
The 60 cases of DLBCL were subdivided into GCB (n = 22) and NGCB (n = 38) according to the phenotypic profiles as proposed by Hans et al. using CD10, bcl-6, MUM1 (analysis after the diagnosis of DLBCL was estblished).5 Briefly, CD10+ cases (n = 17) and CD10– BCL-6+ MUM1– cases (n = 5) were determined as GCB, while the remainder of the cases (n = 38) were classified as NGCB.
Statistical analysis
We compared the expression of survivin and antigen detected by MAb, T332 and clinical parameters such as overall survival (OS), age, performance status (PS), clinical stage (CS), serum lactate dehydrogenase activity (LDH) and original site of lymphoma (site). OS was calculated from the onset of treatment to the time of last follow up or death from any cause. StatView (ver. 5.0, SAS Institute, Cary, NC, USA) was used for statistical calculations and P < 0.05 was taken as significant. All follow-up data were updated on December 2003. Comparisons among groups and clinical parameters such as age, LDH, PS, and CS and site were carried out by unpaired t-test. Survival analysis was based on the Kaplan– Meier method and the significance of differences was esti- mated using log–rank test. RESULTS Tissue distribution of T332 antigen The MAb, T332 demonstrated restricted reaction on cryosec- tions with germinal center B cells (Fig. 1a) in lymph nodes, tonsil and spleen although small B-cells in the mantle zone of lymphoid follicle and marginal zone of spleen were not stained. The T cells in the GC were not stained, as seen in Fig. 1a. The positive reaction to the MAb T332 was seen in the cytoplasm but not on the cell surface by flow cytometric analysis (data not shown). The formalin-fixed and paraffin- embedded tissue sections failed to have a positive reaction to the MAb T332 even after applying various antigen-retrieval treatments. Expression of T332 antigen and survivin Both T332 antigen (Fig. 1a) and survivin (Fig. 1d) showed various degrees of staining. Thus, the cases with positive cytoplasmic staining by the Mab T332 or the cases with nuclear staining pattern of survivin in >30% of lymphoma cells were defined as positive (Fig. 1b,e, respectively). Neg- ative cases had only a few positive cells (Fig. 1c,f). The reactivities of the MAb T332 and survivin are summarized in Table 1. Expression of T332 antigen was seen in 36.3% (16/44), and survivin in 60% (36/60). Among the subtypes of DLBCL, T332 antigen was detected in 37.5% (6/16) of GCB, and in 35.7% (10/28) of NGCB. Similarly survivin was posi- tive in 59.1% (13/22) of GCB and in 60.5% (23/38) of NGCB. The expressions of T332 antigen and survivin did not correlate with each other, thus 29.5% (13/44) were T332 antigen and survivin double positive (DP), while 6.8% (3/44) were T332+survivin–, 31.8% (14/44) were T332–survivin+, and 31.8% (14/44) were T332–survivin– double negative (DN).
Clinical features
The cases expressing survivin had a significantly worse prognosis (P = 0.01) in DLBCL overall (Fig. 2a). The expres- sion of survivin in subtypes of DLBCL did not have a signif- icant difference, but tended to show unfavorable prognoses in both GCB (P = 0.06) and NGCB (P = 0.07) (Fig. 2b,c, respectively). Patients having expression of T332 antigen had a significantly poorer prognosis (P = 0.01) in DLBCL overall (Fig. 3a). Among the subtypes of DLBCL, T332+ NGCB also had a significantly worse prognosis than T332– cases (P = 0.02) (Fig. 3c), while GCB did not demonstrate any significant difference (P = 0.11; Fig. 3b). Finally, a signif- icantly poorer prognosis (P = 0.009) was observed in T332 antigen and survivin DP cases compared to that in the remaining DN and single positive patients (Fig. 4).
There was no significant correlation between the results of IHC with expression of either T332 antigen or survivin, and various clinical parameters (age, LDH, PS, CS, site (Table 2).
DISCUSSION
DLBCL is the most common type of NHL in Japan2 as well as in Western countries,11 thus it is important to know the outcome predictors. Recent gene-expression profiling analy- ses disclosed favorable prognosis of GCB subtype compared to NGCB subtype.3,4 This results is understandable because the GCB is immunohistochemically defined by the expression of CD10 and bcl-6 without MUM1,5,12–14 although there was a minor difference regarding the expression of MUM1,15 and the expression of CD1016 and bcl-617,18 has already been demonstrated as an improved predictor of DLBCL. The reason for better prognosis in GCB is not fully clarified but may partly be attributed to the increased apoptotic status.14 How- ever, the emergence of tumors may depend on the balance between apoptosis and cell proliferation, thus the present study were constructed to clarify the meaning of apoptotic status (ms in preparation) and cell growth activity of DLBCL and its subtypes.
In the present study, 60% of DLBCL were defined as sur- vivin positive and this figure was similar to that obtained in a previous study.19
In addition, the present study found that survivin was associated with unfavorable outcome of DLBCL and was likely to be a predictor of poor prognosis in GCB and NGCB subtypes, and this was compatible with other studies.19,20 Survivin is an inhibitor of apoptosis protein expressed in the G2/M phase of the cell cycle in a cycle- regulated manner,21 and immunohistochemically detected in lymphoma as well as in many cancers.22 Gene expression profiling analysis demonstrated four biologic groups of pre- dictive genes such as the proliferation signature, signatures relating to the immune response to the tumor cells, signature encoding components of the extracellular matrix and connec- tive tissue growth factor, and it was concluded that the pro- liferation signature was the best predictor of an adverse outcome.3,4 Thus the unfavorable prognosis of survivin expression is predicted by the proliferation signature as well as by proliferating cell-marker Ki-6723 or anti-apoptotic pro- tein bcl-2.24 The expression of survivin has been studied in lymphomas other than DLBCL and has been found to be associated with poor prognosis.24,25
Similarly the expression of T332 antigen was associated with unfavorable prognosis of DLBCL or NGCB subtype, although the expression of antigen detected by the MAb T332 was not significant in the GCB subtype. The character- ization of T332 antigen is not completed but it is thought that the MAb T332 might be reacting with molecules relating to cell cycle or cell proliferation because it reacted with dark- zone B-cells in GC and mitotic figures. If this is true, then the worse prognosis for DLBCL coexpressing both survivin and antigen detected by MAb T332 is understandable. We believe that further characterization of T332 antigen may contribute to the understanding of the outcome and histogenesis BI-3802 of DLBCL.