|
|
| |||||||||||||||||||||||||||||||||||||||
Background About 40 percent of non-Hodgkin's lymphomas are diffuse lymphomas with a large-cell component (DLLC). Current therapy can induce a long-term remission in half the patients with DLLC, but more intensive treatment has the potential to improve outcome, particularly in patients at high risk for treatment failure. Clinical and cytogenetic markers can identify subgroups at high or low risk. Rearrangement of a novel candidate proto-oncogene, bcl-6, is a possible prognostic indicator in DLLC.
Methods We performed Southern blot hybridization to detect bcl-6 and bcl-2 gene rearrangement in samples of lymphoma from 102 patients with B-cell DLLC. The results were correlated with the patients' histologic features, age, disease stage, tumor sites and bulk of disease, serum lactate dehydrogenase level, and treatment outcome.
Results Rearranged bcl-6 was found in 23 cases, and rearranged bcl-2 in 21 cases. Nineteen of the patients with rearranged bcl-6 had extranodal DLLC, two had primary splenic lymphomas, and only one had bone marrow involvement. Thirty-six months after diagnosis, the proportion with freedom from progression of disease was projected to be 82 percent (95 percent confidence interval, 66 to 98 percent) among the patients with rearranged bcl-6, as compared with 56 percent (95 percent confidence interval, 43 to 70 percent) for the patients with germ-line bcl-6 and bcl-2 and 31 percent (95 percent confidence interval, 8 to 53 percent) for the patients with rearranged bcl-2. The status of the bcl-6 gene was an independent prognostic marker of survival and freedom from disease progression in a multivariate model and added predictive value to established prognostic signs.
Conclusions Rearrangement of the bcl-6 gene correlated with a favorable clinical outcome in DLLC and may thus serve as a prognostic marker in patients with this form of malignant lymphoma.
Cytogenetic studies and molecular analysis of proto-oncogenes and tumor-suppressor genes have provided insights into the pathogenesis of non-Hodgkin's lymphoma and have determined diagnostic and prognostic markers8,9. Examples include rearrangements of the bcl-2 gene on chromosome band 18q21, observed in up to 85 percent of follicular lymphomas; rearrangement of the bcl-1 gene on chromosome band 11q13 in non-Hodgkin's lymphoma with intermediate differentiation; and perturbation of the myc gene in Burkitt's lymphoma8,9. No single genetic abnormality has been consistently associated with diffuse large-cell lymphoma, but rearrangement of bcl-2 occurs in 20 to 30 percent of patients and it correlates with decreased overall or disease-free survival10,11. Chromosomal translocations, including translocations of the myc proto-oncogene, have also been detected in DLLC, but they are not as important to prognosis as other recurring chromosomal abnormalities8,12.
Recently, we and others identified a novel candidate proto-oncogene, bcl-6, which has structural similarities to a class of transcription factors that participate in the control of cell proliferation and differentiation and organ formation13,14,15,16,17,18. We have also shown that bcl-6 rearrangements occur in a large proportion of cases of DLLC (about 30 percent), making this the first genetic lesion of specific occurrence in these lymphomas19. This report addresses the clinical and prognostic relevance of bcl-6 rearrangement in DLLC. We found that bcl-6 rearrangement correlated with extranodal presentation and a favorable clinical outcome in DLLC. In concert with other clinical features, this molecular marker may be a useful prognostic indicator at the time of diagnosis.
Methods
We studied 102 lymphomas from patients with DLLC evaluated at the time of diagnosis. In all these cases there was clonal rearrangement of an immunoglobulin heavy-chain gene. Each lymphoma thus originated from a neoplastic B cell. The 102 cases involved 102 of 229 patients with DLLC whose lymphomas were serially ascertained over a nine-year period. The other 127 patients were excluded because they had been studied at relapse, they had T-cell DLLC, or no DNA was available for analysis. A hematopathologist classified the lymphomas as being of diffuse large-cell cleaved, noncleaved, immunoblastic, or mixed subtype, according to the international working formulation20. Cytogenetic analysis of each specimen was attempted as previously described21. To detect bcl-6 rearrangements, DNA was digested with BamHI and XbaI and subjected to Southern blot analysis, with a 4-kb SacI-SacI fragment of the bcl-6 gene used as the probe18. Specimens that did not yield cells in metaphase for karyotypic analysis were analyzed for rearrangement of the MBR and MCR breakpoint regions of the bcl-2 gene, as previously described11. Aggregate descriptions of 47 of the cases in the current series were included in our earlier reports of cytogenetic abnormalities in DLLC11,12,13. A detailed molecular analysis of specimens from six patients with bcl-6 rearrangement (Cases 352, 1098, 1254, 1403, 1444, and 1445) has been reported separately18.
Clinical data on each case were compiled as previously described22. Disease stage was assessed according to the modified Ann Arbor criteria23. To evaluate the number of extranodal sites of disease, the extent of involvement of these sites was documented radiographically or pathologically. In the prognostic analysis, bone marrow involvement was scored as an extranodal site, but not the spleen, in accord with the international prognostic index5.
Clinical end points, including complete response to treatment and freedom from disease progression, were defined as previously described3. After genetic analysis of the lymphoma, 93 patients received chemotherapy; 9 patients with early-stage disease were treated by surgical resection, radiation therapy, or both. All patients were treated with curative intent. Their chemotherapy regimens were classified as first-generation chemotherapy,3,4,24,25 second-generation chemotherapy,3,4,26 or third-generation chemotherapy3,4,27,28,29. Eight patients died before the completion of therapy, with incomplete evaluations for disease stage, or of infectious complications during or shortly after treatment. Their cases were not considered capable of evaluation with respect to remission, but they were included in the analysis of overall survival and freedom from disease progression. One patient was judged to be in complete remission, which was confirmed by autopsy after death from infectious complications three weeks after the completion of protocol treatment. All deaths, regardless of cause, were considered to be end points in the analysis of overall survival.
Median survival was determined by the method of Kaplan and Meier30. Correlations between gene rearrangements and clinical features were analyzed with Fisher's exact test12. Means were compared by two-sample t-tests. Univariate comparisons of survival and the duration of freedom from disease progression were made with the log-rank test; the resulting estimates are expressed with their confidence intervals. Multivariate analysis was performed with the Cox regression model31. Stepwise multiple logistic regression was used in the multivariate analysis of factors prognostic of a complete response. For all statistical analyses, a P value of less than 0.05 based on a two-sided test was considered to indicate statistical significance.
Results
Of 102 cases of DLLC studied at the time of diagnosis, 23 involved rearrangement of a bcl-6 gene, 21 involved a t(14;18) karyotype or rearrangement of bcl-2, and 58 showed no evidence of either bcl-6 or bcl-2 rearrangement. Representative results of hybridization analysis for rearrangement of bcl-6 are shown in Figure 1. The clinical characteristics of the patients are summarized in Table 1 according to the presence or absence of bcl-6 and bcl-2 rearrangement. The histologic subtypes and clinical features of the patients with bcl-6 rearrangement are shown in Table 2.
|
|
|
The patients with rearranged bcl-6 had a mean age of 64.1 years at presentation and a high frequency of extranodal involvement by lymphoma; 19 of the 23 patients had stage IE, IIE, IIIE, or IV disease, as compared with 48 of the 79 with germ-line bcl-6 (unrearranged bcl-6) (P = 0.07). Extranodal sites included muscle or subcutaneous tissue (six cases), the stomach (five cases), the lung or pleura (five cases), or the skin, breast, bowel, thyroid, pancreas, or kidney, as assessed by biopsy or as indicated by radiographic abnormalities that improved after chemotherapy. Of the seven patients with stage IE or IIE disease, five had primary extranodal lymphomas and two had extranodal extensions from a primary nodal site. Two patients had primary splenic lymphoma, and two had only peripheral adenopathy. The proportion of patients with stage IV disease among those with bcl-6 rearrangement did not differ from the proportion among those without a rearranged bcl-6 gene. By contrast, bone marrow involvement was observed in 15 of 75 patients with germ-line bcl-6, as compared with only 1 of the 23 with rearranged bcl-6 (P = 0.1).
All but 1 of the 23 patients with bcl-6 rearrangement at the time of diagnosis received chemotherapy including anthracyclines. This patient remained free of disease eight years after resection of a primary splenic large-cell lymphoma. After a median follow-up of two years, 21 of the 23 patients were alive; the actuarial survival was 91 percent (95 percent confidence interval, 80 to 100 percent). Two patients died during or immediately after treatment; autopsy of one patient revealed no evidence of lymphoma. This patient and 19 others were judged to have had a complete remission after treatment. Two patients relapsed with recurrent disease in the lung, and two patients had persistent subcutaneous masses. One of the patients with a relapse (Case 295) subsequently underwent autologous bone marrow transplantation and has remained free of disease for 78 months.
With respect to known prognostic variables, the proportion of patients with serum lactate dehydrogenase levels above 500 U per liter was similar in the groups with and without bcl-6 rearrangement (3 of 23 vs. 13 of 79, P = 0.99). Five of 23 patients with bcl-6 rearrangement had bulky disease, as compared with 35 of 79 without bcl-6 rearrangement (P = 0.1). The proportion with "limited stage" disease (stage I, IE, II, or IIE) was comparable in the groups with and without bcl-6 rearrangement (Table 1).
Multivariate Analysis of Clinical Outcome
The median duration of freedom from disease progression had not been reached at the time of analysis in the group with rearranged bcl-6, whereas it was 70 months in the group with germ-line bcl-6, regardless of bcl-2 status (P = 0.009) (Figure 2A). The projected proportion with freedom from disease progression at 36 months was 82 percent in the group with rearranged bcl-6 (95 percent confidence interval, 66 to 98 percent) and 49 percent in the group with germ-line bcl-6 (95 percent confidence interval, 37 to 60 percent). Multivariate analysis revealed that four variables -- bcl-6 rearrangement, stage IV disease, bulky disease, and high lactate dehydrogenase levels (log transformed) -- were the most powerful predictors of freedom from disease progression (P = 0.007, P = 0.01, P = 0.03, and P = 0.05, respectively) (Table 3). Multivariate analysis also demonstrated that bulky disease, high lactate dehydrogenase levels, bcl-6 rearrangement, and stage IV disease were the most useful predictors of overall survival (P = 0.01, P = 0.02, P = 0.02, and P = 0.05, respectively).
|
|
The patients with rearranged bcl-2 demonstrated a trend toward poorer survival, as compared with the patients with germ-line bcl-2, regardless of the configuration of the bcl-6 gene (P = 0.12). When the results for bcl-6 and bcl-2 were considered together (Figure 2B), the projected actuarial survival at 36 months was 91 percent in the group with rearranged bcl-6 (95 percent confidence interval, 80 to 100 percent), as compared with 59 percent in the group with germ-line bcl-6 and germ-line bcl-2 (95 percent confidence interval, 44 to 74 percent) and 46 percent in the group with rearranged bcl-2 (95 percent confidence interval, 21 to 72 percent). The log-rank test demonstrated a significant difference in survival between the group with bcl-6 rearrangement and the two groups without this feature (P = 0.02) (Figure 2B); the main reason for this difference was the better survival of the group with rearranged bcl-6. The projected proportion with freedom from disease progression at 36 months was 82 percent in the group with rearranged bcl-6 (95 percent confidence interval, 66 to 98 percent), 56 percent in the group with germ-line bcl-6 and germ-line bcl-2 (95 percent confidence interval, 43 to 70 percent), and 31 percent in the group with rearranged bcl-2 (95 percent confidence interval, 8 to 53 percent). The median length of follow-up of survivors was two years. Rearranged bcl-2 did not emerge as an independent prognostic marker in the multivariate analysis of survival or freedom from disease progression.
There was also no significant effect of the generation of chemotherapy (first, second, or third generation; see the Methods section) on survival or freedom from disease progression (P = 0.95 and P = 0.21, respectively). There was a trend toward a higher rate of complete responses in the group with bcl-6 rearrangement (P = 0.1) (Table 1), although logistic regression revealed that only the clinical variables of high lactate dehydrogenase levels, stage IV disease, and bulky disease were independent predictors of response.
Relation between bcl-6, bcl-2, and 8q24 Rearrangements
Of the 79 lymphomas without bcl-6 rearrangement, 21 had the t(14;18)(q32;q21) karyotype or rearrangement of bcl-2 according to molecular analysis. In these cases the patients were older at diagnosis but similar with respect to lactate dehydrogenase levels and distribution of histologic types to the patients in whom neither bcl-2 nor bcl-6 was rearranged (Table 1).
Nine lymphomas had the translocation t(8;14) (q24;q32), which is characteristic of involvement of the myc gene. Two of these lymphomas also had a rearranged bcl-6 gene. The t(8;14) translocation had no effect on survival, whether or not the bcl-6 rearrangement was present, and the two lymphomas with t(8;14) translocation and bcl-6 rearrangement did not show evidence of histologic transformation or other unusual histologic features.
Cytogenetic Features, Including the Relation between 3q27 Aberrations and bcl-6 Rearrangement
Of the 65 lymphomas with karyotypic abnormalities, 14 had translocations involving 3q27 and 1 had a deletion affecting band 3q27. Only 11 of these 15 tumors had rearrangement of bcl-6. Five lymphomas with apparently normal chromosomes 3 demonstrated bcl-6 rearrangements on DNA analysis.
Discussion
The DLLC group of lymphomas has not been associated with a characteristic genetic abnormality8. The t(14;18)(q32;q21) karyotype or its molecular equivalent, bcl-2 rearrangement, which occurs in the vast majority of follicular lymphomas, has been observed in 20 to 30 percent of cases of DLLC8. In such cases, the t(14;18) abnormality may reflect a follicular origin of the tumor. The recognition of translocations involving 3q27 and the sites of immunoglobulin genes, 14q32, 22q11, and 2p12 in predominantly diffuse nonHodgkin's lymphoma led to the molecular cloning of bcl-6 and its localization to 3q2713,14,15,16,17,18. Although not unique to diffuse large-cell lymphomas, translocations affecting 3q27 were observed in only 7 of more than 200 cases of follicular non-Hodgkin's lymphoma with abnormal karyotypes that were reported in a catalogue of chromosomal abnormalities in cancer32. Of 28 cases of follicular non-Hodgkin's lymphoma analyzed in a previous study, none involved rearrangement of bcl-618. Rearranged bcl-6 is thus the most common genetic lesion characteristic of DLLC.
Unlike the 18q21 translocations in non-Hodgkin's lymphoma, which to date have involved only immunoglobulin-gene loci as reciprocal partners, the 3q27 translocations found in this study showed a marked promiscuity of partners in rearrangement. In addition to the sites of the immunoglobulin genes, we and others have documented reciprocal translocations involving the 3q27-29 region with at least 12 other loci; a total of 79 cases of DLLC with 3q27 translocations have been reported13,14,32,33.
The frequent rearrangement of bcl-6 in DLLC characterized by extranodal involvement represents one of the few genetic markers for this subgroup of lymphomas8. By contrast, rearrangements of bcl-1, bcl-2, or bcl-3 occur infrequently among extranodal lymphomas,34,35,36 although 5 of 12 extranodal lymphomas of gastric origin in one series had a myc (8q24) rearrangement36. We could not confirm the association between 8q24 rearrangement and gastric lymphoma, although we saw the t(8;14) translocation in five cases of extranodal lymphoma, one of which also had bcl-6 rearrangement. The proportion of patients with stage IE, IIE, IIIE, or IV disease was higher among cases with rearranged bcl-6 than among those with germline bcl-6; in the latter group, stage IV DLLC was commonly due to bone marrow involvement. Whether this association of rearranged bcl-6 with extranodal lymphoma reflects an effect of the primary deregulation of bcl-6 or "secondary" genetic events associated with tumor progression8,21 is unclear. The observation of the t(3;22), t(2;3), or t(3;14) translocation as a solitary cytogenetic abnormality in some tumors13,14 is consistent with a primary pathogenetic role for this type of translocation.
The favorable prognosis in stage IE or IIE extranodal DLLC with bcl-6 rearrangement is consistent with earlier reports of a good prognosis associated with localized extranodal large-cell non-Hodgkin's lymphoma treated with surgery or radiotherapy37. Extranodal involvement in advanced disease, noted in the majority of cases with bcl-6 rearrangement, has generally been considered a poor prognostic factor in DLLC, although its negative consequences were most evident when this factor was combined with other adverse indicators such as bulky disease, high lactate dehydrogenase levels, or low performance status5,22,38. In contrast, bone marrow involvement, observed in 22 percent of cases of DLLC and considered an extranodal site according to the international prognostic index,5 was rare in the group with rearranged bcl-6. A recent study39 showed a slight but not statistically significant survival advantage in patients with DLLC and laz-3(bcl-6) rearrangement, although the ascertainment of cases of extranodal disease was low and overall survival of both groups was lower than predicted with the international index. The favorable outcome of treatment in the group with bcl-6 rearrangement in our study must be weighed against the observation of relapse or residual disease in three of the patients still alive. An additional patient with relapse remains in remission six years after salvage autologous bone marrow transplantation.
Some clinical markers of favorable prognosis were found in patients with DLLC and rearranged bcl-6. With respect to serum lactate dehydrogenase levels and the proportion of patients with disease in stage I through stage IIIE, the groups with and without bcl-6 rearrangement were comparable. However, the proportion of patients with bulky disease or bone marrow involvement was lower in the group with bcl-6 rearrangement. Multivariate analysis suggested that bcl-6 gene rearrangement added independent prognostic power when analyzed together with clinically derived variables of the international prognostic index5. This observation is exemplified in Case 352, in which there were both bcl-6 rearrangement and clinical features consistent with a high level of risk as defined by the international index (elevated lactate dehydrogenase levels, extensive extranodal involvement, low performance status, and stage IV disease) but a durable remission.
The identification of favorable and unfavorable prognostic markers offers the potential to stratify groups of patients for therapeutic approaches to DLLC4,5,6,7,22,38. The probability of treatment failure remains as high as 25 to 40 percent in the subgroups of DLLC with the most favorable prognoses based on current models, highlighting the need for genetic or other prognostic markers5. In addition to its potential diagnostic and prognostic applications, the identification of bcl-6 breakpoint regions may allow development of polymerase-chain-reaction-based indicators of minimal residual disease40. The availability of bcl-6 rearrangement as a molecular marker of large-cell lymphoma constitutes a potentially important clinical tool in the care of patients with this disease.
Supported by grants from the National Institutes of Health (CA-08748, CA-34775, and CA-44029), the Matheson Foundation, and the Lymphoma Foundation; by Clinical Oncology Career Development Awards (to Drs. Offit and Ladanyi) from the American Cancer Society, and by funds (to Dr. Lo Coco) from the Associazione Italiana contra le Leucemie.
We are indebted to Drs. James P. O'Brien and David Straus for their comments and suggestions; to Drs. Ellin Berman, Joseph Bertino, Charles Farber, Timothy Gee, Subi Gulati, Gwen Nichols, and Andrew Zelenetz and physicians referring patients to Memorial Hospital for providing clinical data on presentation and follow-up; and to Reiner Siebert and Claudia Little for assistance in assembling clinical data and for data entry and analysis.
Source Information
From the Cell Biology and Genetics Program (K.O., N.Z.P., R.S.K.C.), the Departments of Medicine (K.O., C.P.), Pathology (D.C.L., D.A.F., S.J.), and Epidemiology and Biostatistics (D.L.), and the Cytogenetics Service (K.O., M.L., S.J., R.S.K.C.), Memorial Sloan-Kettering Cancer Center; and the Division of Oncology, Department of Pathology, College of Physicians and Surgeons, Columbia University (F.L.C., B.H.Y., F.L., R.D.-F.); and Columbia University (A.R.) -- all in New York; and Sezione di Ematologia, Dipartimento di Biopatologia Umana, Universita "La Sapienza" (F.L.C.), and Centro Studi e Ricerche della Sanita dell'Esercito (F.L.) -- both in Rome.
Address reprint requests to Dr. Offit at the Memorial Sloan-Kettering Cancer Center, Box 192, 1275 York Ave., New York, NY 10021.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
The Advantage of Rearranging bcl-6
Mooi W. J., Kluin P. M., Offit K., Chaganti R.S.K.
Extract |
Full Text
N Engl J Med 1994;
331:1459-1460, Nov 24, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
