BCL2 expression in chronic lymphocytic leukemia: lack of association with the BCL2 −938A>C promoter single nucleotide polymorphism

One of the hallmarks of chronic lymphocytic leukemia (CLL) is high-level expression of the antiapoptotic protein BCL2. This has led to the development of therapeutic approaches targeting BCL2, either via antisense approaches to down-regulate BCL2 RNA¹  or through the development of BH3 mimetics.^2,3  The latter appear to be remarkably effective against CLL in vitro with a nanomolar EC₅₀.³ 

What drives BCL2 expression in CLL is not clear. Unlike follicular lymphoma, most patients lack BCL2 chromosomal translocation.⁴  Different mechanisms have been proposed, including promoter hypomethylation,⁵  loss of microRNA (miRNA) expression (specifically, loss of miRNA-15a and miRNA-16-1 that are frequently deleted in CLL⁶ ), and expression of nucleolin.⁷  The 5′ untranslated region of BCL2 mRNA also contains an untranslated open-reading frame (uORF), which may repress translation.⁸  The relative importance of transcriptional and posttranscriptional mechanisms in the control of BCL2 protein levels in B-cell malignancies lacking t(14;18)(q32;q21) remains to be determined and may have therapeutic implications.

BCL2 transcription is controlled by 2 major promoters, P1 and P2.⁹  P1 is located 1386 to 1423 bp upstream of the translation initiation site and is a G-C–rich, TATA-less promoter with multiple transcription initiation sites.⁹  In contrast, P2, which lies 1300 bp proximally, contains a CCAAT box, an octamer motif, and a TATA element. Negative regulatory elements between P1 and P2 decrease transcription.¹⁰  In normal B cells, the P1 promoter is used predominantly, but in B-cell lymphoma cell lines with t(14;18)(q21;q23), P2 becomes active due to changes in transcription factor abundance.¹¹ 

Although most CLLs express levels of BCL2 protein comparable to those seen in follicular lymphoma with t(14;18)(q32;q21.3), several studies have shown that some express lower levels.^5,12,–14  The causes of this variation are unknown. It has recently been suggested that a single nucleotide polymorphism (SNP) within the inhibitory region of the BCL2 promoter may not only control levels of expression of BCL2 protein, but may also be associated with a number of clinical parameters.¹⁵  We sought to confirm these observations in a population-based series of patients with CLL attending a single center. We failed to confirm the association of the BCL2 promoter SNP with either expression of BCL2 protein or any other clinical parameter. Patients with CLL with low-level expression of BCL2 were heterogeneous, both biologically and clinically.

Blood samples were taken from 276 patients with CLL following informed, written consent and with University Hospitals Leicester Ethical Committee approval. Genomic DNA and total cellular RNA were isolated using Qiagen columns (Hilden, Germany) from leukemic mononuclear cells prepared by Ficoll gradient centrifugation. The SC-1 B-cell lymphoma cell line¹⁶  carrying t(8;14;18)(q24;q32;q21) chromosomal translocation was obtained from DSMZ (www.dsmz.de, Braunschweig, Germany). An Epstein-Barr virus (EBV)–transformed normal B-cell line HRC-57 was obtained from Dr Jude Fitzgibbon (Cancer Research UK [CRUK], London, United Kingdom). Amplification and sequence analysis of BCL2 promoter regions was performed by DNA polymerase chain reaction (PCR) as described.¹⁵  Amplification and analysis of IGHV rearrangements was performed according to BIOMED-2 protocols.¹⁷  Western blot was carried out as previously described.¹⁸  Blots were quantified by densitometry using ImageQuant software (GE Healthcare, Buckinghamshire, United Kingdom). BCL2 expression was normalized first against β-actin and then compared with the BCL2 expression of the SC-1 cell line as described in Document S1 (available on the Blood website; see the Supplemental Materials link at the top of the online article). Quantitative reverse transcription (QRT)–PCR for BCL2 and TBP on an ABI 7000 lightcycler (Applied Biosystems, Cheshire, United Kingdom) was performed as described in Document S1. Fluorescence in situ hybridization (FISH) was performed for detection of deletions in 13q14, 11q22 and 17p13, trisomy 12 and IGH breakpoints, and, in cases with IGH breakpoints, IGH-BCL2 fusions using methods described in the Document S1 and in Martin-Subero et al.¹⁹  Genotype-dependent differences in protein expression were assessed using 1-way analysis of variance (ANOVA). Association of the BCL2 SNP with sex, stage diagnosis, IGHV, treatment, and survival were analyzed using chi-square. Time to progression was calculated from date of diagnosis to first treatment in months and also analyzed using ANOVA and plotted using Kaplan-Meier. Statistical significance was assumed at a P value less than .05. All statistical tests were carried out using the GraphPad Prism Version 4.0 software (GraphPad Software, San Diego, CA).

We assessed the frequency of the BCL2 promoter SNP −938A>C in 276 patients with CLL derived from our database. This database reflects a consecutive, population-based series of patients attending the hematology clinic in Leicester (A.M., R.W., S.G., R.S., D.B.J.K., and M.J.S.D., manuscript in preparation). Of these 276 patients, 97 were A/A genotype, 127 were A/C genotype, and 52 were C/C genotype. These data are very similar to those reported previously.¹⁵  We correlated the genotype with a number of parameters (Table 1) and with BCL2 protein expression (Figure 1A-D). We found no significant correlation of SNP genotype with either clinical parameters or laboratory parameters such as IGHV mutational status and interphase FISH abnormality. A total of 98 patients have been treated to date. In contrast to the previous study, there was no association with time to first treatment, as shown in Figure 1E. There were no differences in the SNP genotype between treated and nontreated patients (Table 1).

To assess possible linkage of BCL2 protein levels with the −938C>A SNP, we performed quantitative Western blotting on 100 patients using the cell lines SC-1 and HRC-57 as high- and low-level comparators, respectively. Both treatment-naive (64 patients) and previously treated patients (36) were studied. First, to ensure that BCL2 protein levels remained relatively constant over time, 10 patients for whom multiple samples were available were studied; in all 10 patients, there were no significant changes in BCL2 protein levels with time, despite all patients having received therapy (Figure S1). All 100 patients expressed BCL2 protein, but to variable levels. Overall, there was no association of BCL2 protein levels with the promoter SNP in either treatment-naive or treated subgroups (Figure 1A-D; data not shown).

BCL2 protein level was a continuous variable in CLL (Figure 1D). The 19 patients showing the lowest levels of BCL2 protein expression, comparable with those seen in HRC-57, were studied further (range, 0.16-0.48; mean levels of expression in HRC-57, 0.34). Patients expressing lower amounts of BCL2 were distributed equally between the different SNP −938 genotypes with no preferential association of the C/C genotype. There was no preferential expression of MCL1 in low BCL2-expressing CLL (data not shown). Within this subgroup there was both clinical and biological heterogeneity. A total of 5 of 15 patients examined by QRT-PCR showed high-level RNA expression comparable with that seen in the SC-1 cell line. A total of 4 patients expressing low amounts of BCL2 pursued an aggressive clinical course (Table S1). Low levels of BCL2 expression were not due to mutations involving the BCL2 epitope recognized by the antibody used in this study (Document S1), and did not reflect high level of normal residual T cells. All patients with either heterozygous or homozygous deletion of 13q14 showed high-level expression of BCL2 (Figure S2). BCL2 chromosomal translocations are seen rarely in CLL.⁴  A total of 5 typical CLLs with t(14;18)(q32;q21) were studied. All showed high-level BCL2 protein expression (Figure S3).

The possible prognostic significance of levels of the antiapoptotic proteins BCL2 and MCL1 or BCL2/BAX ratios in CLL remains controversial. A recent study showed no prognostic significance of any of these proteins in a large cohort of patients.²⁰  We have failed to confirm that the −938C>A polymorphism associates with BCL2 protein levels and disease progression in CLL. Patients with CLL expressing lower levels of BCL2 protein were biologically and clinically heterogeneous. Overall, our data indicate that BCL2 protein levels reflect a complex interplay of transcriptional and posttranscriptional controls. Whether BCL2 levels determine outcome to BCL2-targeted therapies remains to be determined.

This work was supported by the UK Medical Research Council, the Hope Foundation, and the Habilitierten Fund of the Medical Faculty Kiel.

Contribution: A.M., O.T., R.W., S.G., and R.S. performed experiments, analyzed data, and wrote the paper. D.B.J.K. and M.J.S.D. designed research and contributed patient material. M.J.S.D. analyzed data and wrote the paper.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Martin J. S. Dyer, MRC Toxicology Unit/Leicester University, Hodgkin Bldg, Rm 402, Lancaster Road, Leicester, United Kingdom LE1 9HN; e-mail: mjsd1@le.ac.uk.