The REL proto-oncogene encodes a transcription factor in the nuclear factor κB (NF-κB) family, and the activation of the REL protein can be controlled by subcellular localization.1 The REL locus, located at chromosomal position 2p16, is amplified in many human B-cell lymphomas,2 and overexpression of REL can transform chicken lymphoid cells in vitro.3
Houldsworth et al4 recently reported on REL protein expression in a panel of diffuse large B-cell lymphomas (DLBCLs) with and without REL amplification. Using indirect immunofluorescence to assess subcellular localization of REL, these authors determined that DLBCLs with REL gene amplification did not have increased nuclear accumulation of REL protein compared with DLBCLs without REL gene amplification. This led these authors to conclude that REL protein activity is not involved in the development of DLBCLs with REL amplification, and that REL may not be the relevant oncogene in DLCBLs with amplifications of chromosomal region 2p16.
Based on the large amount of research that has been conducted on in vitro transformation of chicken lymphoid cells by v-Rel or more recently by human REL, we believe this is a faulty conclusion. First, by indirect immunofluorescence, v-Rel and human REL are largely cytoplasmic proteins in transformed chicken lymphoid cells.3,5 However, in electrophoretic mobility shift assays using nuclear extracts, there is clearly nuclear Rel DNA-binding activity in v-Rel– and REL-transformed chicken lymphoid cells,3,6 and v-Rel induces the expression of several κB site–containing target genes.7 In addition, in these transformed cells, v-Rel and REL are continually shuttling through the nucleus,3,8 and one cannot detect this movement by a static immunofluorescence image. Moreover, this nuclear shuttling appears to be important for oncogenic activity, because if the nuclear shuttling of v-Rel is decreased by the addition of a strong nuclear export signal onto v-Rel, its transforming activity is reduced.9 Thus, it is our contention that the REL immunofluorescence data of Houldsworth et al4 on DLBCLs are entirely consistent with the current model indicating that Rel proteins transform avian lymphoid cells by entering the nucleus, binding to DNA, and activating gene transcription, regardless of the primarily cytoplasmic localization in transformed cells that one observes by immunofluorescence.7 Finally, we show here that REL immunofluorescent staining in the human DLBCL cell line RC-K8 appears primarily cytoplasmic (Figure 1), even though these cells have abundant nuclear REL κB site–binding complexes.10 More importantly, the viability of RC-K8 cells appears to depend on this nuclear REL activity in that expression of the REL inhibitor IκBα kills these cells.10
Cytoplasmic REL staining in the RC-K8 human DLBCL cell line. RC-K8 cells10 were analyzed by indirect immunofluorescence using an anti-REL antiserum against the C-terminal 15 amino acids of REL, as described by Houldsworth et al.4 In the right panel, nuclear DNA was detected in the same field using DAPI (4′,6 diamidino-2-phenylindole). Original magnification, × 600.
Cytoplasmic REL staining in the RC-K8 human DLBCL cell line. RC-K8 cells10 were analyzed by indirect immunofluorescence using an anti-REL antiserum against the C-terminal 15 amino acids of REL, as described by Houldsworth et al.4 In the right panel, nuclear DNA was detected in the same field using DAPI (4′,6 diamidino-2-phenylindole). Original magnification, × 600.
In short, we believe that the recent data of Houldsworth et al4 are consistent with what is observed with Rel-directed lymphoid cell transformation in chicken cell models, wherein the absolute levels of nuclear Rel as determined by immunofluorescence are not necessarily predictive of its role in malignant transformation. As such, the recent results4 do not dismiss a role for REL as the active participant in human DLBCLs with REL gene amplifications.
Response: REL may not be the functional target of the 2p12-16 amplicon in diffuse large B-cell lymphoma
In our recent report,1 we concluded that REL may not be the functional target gene of the 2p12-16 amplification event in diffuse large B-cell lymphoma (DLBCL). This conclusion was derived from our studies in which no correlation was found between the relative levels of REL amplification, REL mRNA, and nuclear accumulation of what was presumably the active form of REL. This is distinct from classical Hodgkin disease (HD), where a recent study has shown that all specimens with 2p gain/amplification exhibited nuclear accumulation of REL.2 For specimens without 2p gain, REL nuclear staining was evident overall in fewer specimens and in a smaller percentage of cells.2 While we did not report evaluation of the levels of REL protein, our Western blotting data (J.H. and R.S.K.C., unpublished data, 2003) on a subset of the tumors also indicated a lack of correlation between REL amplification and overall REL protein levels. Based on the expected biologic features of amplified target genes and on the aggressive tumor phenotype commonly associated with gene amplification (not found for REL), we concluded that REL may not be the target gene of the 2p12-16 amplicon and suggested that a comprehensive analysis of the genes within this region may reveal a more relevant candidate(s).
Distinct from this issue, our studies did not dismiss an active role for REL in B-cell transformation per se. Immunohistochemical studies of human B-cell lymphoma have indicated clear evidence for activation of the REL/nuclear factor κB (NF-κB) pathway in classical HD and mediastinal large B-cell lymphoma, and in DLBCL, though to varying degrees.1-4 The studies quoted and performed by Gilmore et al have clearly demonstrated a role for REL in chicken B-cell transformation and highlighted the importance of the passage of REL through the nucleus for its oncogenic role. Thus, while the question remains as to whether the overall level of REL detected in the nucleus and the number of REL-positive nuclei at any given time are true measures of NF-κB activity, a role for REL in human B-cell transformation, though likely, remains to be proven in an appropriate model system.
Correspondence: R. S. K. Chaganti, Department of Medicine, Memorial Sloan-Kettering Cancer Center, 1275 York Ave, New York, NY 10021; e-mail: chagantr@mskcc.org.
Supported by the National Institutes of Health (RO1 CA 47763 [T.D.G.]). D.T.S. was supported in part by a Pre-doctoral Fellowship from the Natural Sciences & Engineering Research Council of Canada.
