We read with interest the manuscript of Lossos et al,1 in which they disclosed the impact of BCL6mRNA expression on the clinical outcome of diffuse large B-cell lymphoma (DLBCL). The authors established a quantitative real-time polymerase chain reaction (PCR) method using TaqMan technology to measure the levels of BCL6 mRNA in clinical materials. They first analyzed 22 cases of DLBCL and found that the overall survival (OS) of patients with high BCL6 expression was significantly superior to that of patients with low BCL6 expression, and their finding was validated in an additional 39 cases. They suggested that high BCL6 mRNA expression is a new favorable prognostic factor for DLBCL.
3q27 translocation affecting BCL6 gene has been observed in 20% to 40% of DLBCL.2 The particular feature ofBCL6 translocation is that it can involve not only either one of the 3 immunoglobulin genes (Ig) but also another non-Ig partner.2,3 Our initial study suggested that DLBCL cases with non-Ig/BCL6 translocation showed a worse prognosis than those withIg/BCL6.4 In an updated comparison between 17 cases with non-Ig partners, including 2 with a deletion of a larger than 1-kb segment encompassing the BCL6 exon 1, and 26 cases with Ig partners, OS of the former group was inferior to that of the latter group (P = .0400). Thus, we propose that the non-Ig/BCL6 fusion gene is a poor prognostic indicator of DLBCL.
To test the correlation between the 2 independent studies on the prognostic significance of BCL6, we compared the levels ofBCL6 mRNA between DLBCL cases with Ig/BCL6translocation and those with non-Ig/BCL6. Total cellular RNA were prepared from cryopreserved tumor cells and subjected to real-time PCR analysis using an ABI Prism 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA). The sequences of oligonucleotide primers as well as the fluorogenically labeled probe were as described.1 The amount of BCL6 mRNA of test materials was divided by that of the endogenous reference, the glyceraldehydes-3-phosphate dehydrogenase gene (GAPDH), and the BCL6-GAPDH values were further normalized with that of Raji cells.1 As indicated in Figure1, the comparison study clearly showed that BCL6 mRNA levels of the Ig/BCL6 group (n = 6; range, 2.2-7.0; mean, 4.3) were significantly higher than those of the non-Ig/BCL6 group (n = 8; range, 0.4-1.9; mean, 1.0) (P = .0003). In contrast, the values of DLBCL cases that lacked BCL6 rearrangement were distributed from 1.2 to 10.7 (n = 9; mean, 6.5). We next performed reverse transcriptase–mediated PCR using a forward primer for BCL6exon 1 and a reverse primer for exon 4. There were no measurable amounts of normal BCL6 transcripts in cases withBCL6 translocation, though additional PCR using nested primer sets generated PCR products in a proportion of the materials tested. It is therefore likely that the observed BCL6 mRNA in both Ig/BCL6 and non-Ig/BCL6 cases represented transcription not of the normal BCL6 but of the rearrangedBCL6 allele.
DLBCL is a heterogeneous subcategory of non-Hodgkin lymphoma in terms of cell morphology, immunophenotype, and genetic abnormality. Many studies have focused on whether these parameters are associated with particular clinical features and the treatment outcome of DLBCL.5 Lossos et al1 and our own studies,4 including this report, suggest thatIg versus non-Ig partner of BCL6translocation and high versus low BCL6 mRNA expression are both prognostic indicators of DLBCL, potentially reflecting a role ofBCL6 in the pathogenesis of DLBCL. We very recently found that t(3;16)(q27;p11) translocation leads to the fusion ofBCL6 with the interleukin-21 receptor gene(IL-21R).6 Although IL-21R was actively transcribed in a DLBCL cell line (YM) carrying this particular non-Ig/BCL6 translocation, the level of BCL6mRNA, which was under the control of IL-21R promoter, was unexpectedly low (Figure 1). Thus, it is possible that BCL6expression is down-regulated in lymphoma cells with t(3;16) once the cells have obtained a malignant phenotype. In contrast, lymphoma cells with Ig/BCL6 may show a persistent expression ofBCL6 at higher levels, corresponding to a feature of germinal center B-like DLBCL.7 Further studies are needed to elucidate the mechanistic detail involved in the transcriptional control of BCL6 in lymphoma cells.