Mutation analysis of the 5′ noncoding regulatory region of the BCL-6 gene in non-Hodgkin lymphoma: evidence for recurrent mutations and intraclonal heterogeneity

Biopsy specimens from 11 NHL patients with follicular center lymphoma (FCL)8 and DLBCL3 that were stored as frozen viable single-cell suspensions were used in this study. Tumor samples were routinely immunophenotyped by flow cytometry for expression of Ig heavy and light chains and B- and T-cell markers. For B-cell enrichment, tumor cells were incubated with anti-CD 19 microbeads (Miltenyi Biotec, Auburn, CA) for 15 minutes at 4°C according to the manufacturer's protocol. After incubation, the cells were passed through a 70-μm nylon cell strainer (Becton Dickinson, San Jose, CA) and were enriched by magnetic cell separation using MidiMacs system (Miltenyi Biotec). Cells from every specimen were stained with anti-CD 20 and anti-CD 5 or anti-CD 3 antibodies (Becton Dickinson) before the enrichment, after enrichment, and in the flow-through part of the specimen. Tumor B cells were enriched to more than 95% in all the specimens. Normal T lymphocytes from 2 tumor specimens were enriched by CD 3 microbeads (Miltenyi Biotec). Blood mononuclear cells from 10 healthy persons were isolated by Ficoll-Isopaque density centrifugation Amersham Pharmacia Biotech, Piscataway, NJ.

High molecular-weight DNA was extracted from 5.0 × 10⁶ cells with a commercially available kit as described by the manufacturer (QIAamp Tissue Kit; Qiagen, Valencia, CA). The first intron region of the BCL-6 gene that has previously been shown to undergo extensive mutation in B cells was amplified by polymerase chain reaction (PCR) using 0.5 to 1 μg of genomic DNA and the following primers: 5′-CCGCTGCTCATGATCATTATTT and 5′-TAGACACGATACTTCATCTCAT.

The reaction was primed from bp 1702, approximately 650 base pairs (bp) 3′ of the TATA box, and extended 790 bp downstream. The PCR was performed in a final volume of 50 μL containing a final concentration of 0.5 μmol of each primer, 20 mmol/L Tris-HCl (pH 8.4), 50 mmol/L KCl, 1.5 mmol/L MgCl₂, 200 μmol/L of each dNTP, and 2.5 units of Taq polymerase (GibcoBRL). The PCR conditions were as follows: 94°C for 5 minutes, 57°C for 30 seconds, 75°C for 1 minute, 1 cycle; 94°C for 30 seconds, 57°C for 30 seconds, 75°C for 1 minute, 30 cycles; and 75°C for 6 minutes, 1 cycle. For each PCR, a control with no added template was used to check for contamination. PCR products were analyzed by 2% agarose gel electrophoresis and staining with ethidium bromide. Bands of appropriate size were excised from the gels and purified by adsorption to a silica matrix (QIAquick columns, Qiagen).

The purified PCR amplicons were cloned into a TA-PCR cloning vector (Invitrogen, Carlsbad, CA). After the transformation of competentEscherichia coli (1 Shot INV αF′, Invitrogen, Carlsbad, CA) and plating on selective agar (50 μg/mL kanamycin, 40 μg/mL  × gal), 10 to 12 white colonies were picked per sample and used in a second round of PCR. At least 5 amplicons were selected for sequencing. DNA sequencing of PCR amplicons was performed on a 373 automatic DNA sequencer (Applied Biosystems, Foster City, CA) using ABI Prism Big Dye Terminator Kit (Perkin Elmer, Foster City, CA) as recommended by the manufacturer. The same primers used for the PCR were used for forward and backward sequencing.

Sequence analysis was performed using the programs MacVector and Assembly Lign (Oxford Molecular Group, Campbell, CA). Sequences were aligned with the BCL-6 hypermutation region germ line sequence sequenced from the blood mononuclear cells of the healthy people by the same methods. The Taq polymerase error frequency in our laboratory is 0.09%, which amounts to 0.71 mutations per BCL-6 clone.

The following definitions were used in this manuscript:

Unconfirmed mutation: a substitution mutation observed in only 1 of the BCL-6 gene clones from the same tumor specimen.

Confirmed mutation: a mutation observed more than once in the BCL-6 gene clones from the same tumor specimen.

Recurrent mutation: a mutation observed in the BCL-6 gene clones from 2 or more tumor specimens from different patients.

Homozygous mutation: mutation status in tumor specimens in which only 1 identically mutated BCL-6 gene sequence was observed in all the tested BCL-6 gene clones.

Heterozygous mutation: mutation status in which 2 or more different BCL-6 gene sequences were observed in the genes cloned from a given tumor.

The number of single unconfirmed substitution mutations in each tumor specimen is included in Table 1, but these mutations were disregarded in our determination of the tumor mutation status, because at least some of them could have resulted from Taq polymerase error, as described later.

The first nucleotide of the amplified BCL-6 gene region, corresponding to the first nucleotide of the sense primer was arbitrarily defined as position +1.

BCL-6 gene cloning and sequencing was performed in 10 healthy donors and 11 patients with untreated FCL8 and DLBCL.3 

An identical sequence of the BCL-6 first intron region was obtained from multiple molecular clones derived from the blood mononuclear cells of 10 healthy donors (Figure 1). This sequence differs at 7 positions from the previously published BCL-6 first intron sequence.12 

Single but unconfirmed substitution changes from the established germ line BCL-6 gene sequence were observed in all the tested tumor specimens. Their number ranged from 1 to 15 per specimen, with an incidence of 0.2 to 1.5 bp/BCL-6 gene clone. In 7 samples (samples 1,5,6,8,9,10, and 11 in Table 1), the incidence of the single base-pair substitutions was similar to or lower than the 0.71 bp/ BCL-6 gene clone Taq error rate. Thus, the unique base-pair substitutions observed in these samples most probably originated from Taq polymerase error. In 4 samples (samples 2,3,4, and 7 in Table 1), the incidence of the single substitutions was higher than the Taq polymerase error rate, suggesting that at least some of these substitutions were real mutations. Because it is impossible to discriminate between real single substitution mutations and Taq polymerase errors, we decided to be conservative in our analytical approach and to exclude all the single unconfirmed substitution mutations from our subsequent analyses. In contrast, deletions and insertions that were observed only once were included in the subsequent analyses, because they cannot be attributed to Taq polymerase errors.

Deletions, insertions, and repeated confirmed nucleotide substitution changes from the established germ line BCL-6 gene sequence were detected in all FCL specimens and in 1 of 3 DLBCL specimens (Table1). The number of these changes varied from 1 to 10 per specimen, with an incidence ranging from 1.3 × 10⁻³ to 1.3 × 10⁻²/bp. A total of 21 distinct confirmed substitution changes from the established germ line BCL-6 gene sequence, 3 deletions, and 1 insertion were disclosed in the tested tumor samples. Eleven of these substitution changes were transitions, whereas 10 were transversions. One deletion mutation (Δ T 520) and 3 substitutions (123 [C→T], 186 [T→C], 397 [G→C]) were recurrently observed in more than 1 tumor specimen from different individuals, but in none of the BCL-6 gene sequences from healthy normal individuals studied in our laboratory. One report has noted that deletion Δ T 520 and substitution 397 [G→C] may be polymorphisms.15 None of these recurrent BCL-6 gene changes were observed in the G nucleotide of the RGYW motif (where R = purine, Y = pirimidine, and W = A or T) that is preferentially mutated in Ig genes and represents a mutational hot spot.17 Only 1 of these recurrent substitutions occurred in another well-established mutational hot spot in Ig genes — GC or TA dinucleotide motifs (position 123).18 

To examine the possibility that inherited polymorphism may have accounted for these recurrent changes, the BCL-6 gene sequence was determined in normal T cells from samples 1 and 8 (Table 1). These 2 tumors together demonstrated all the 4 recurrent changes that we had observed. The BCL-6 gene sequences derived from T cells of sample 8 were identical to the established germ line sequence, thus demonstrating that all the observed changes in this sample, including the recurrent 186 (T→C) change, represent real mutations. In the T cells from sample 1, 2 distinct BCL-6 gene sequences that varied at position 397 (G or C) were found. Therefore, the G→C change at position 397 represents polymorphism, as was previously reported.15 All the remaining changes, including recurrent 123 (C→T) and (Δ T 520) were absent from the normal cells and, thus, represent somatic mutations in the tumor.

The recurrent occurrence of identical mutations at the same nucleotide may suggest the presence of mutational hot spots. To further evaluate the recurrent nature and the potentially skewed distribution of BCL-6 gene mutations, we combined our data with all the published mutations in NHL and Hodgkin lymphoma for which nucleotide change and exact mutation positions have been previously reported (Figure2).15,19-22 A total of 249 nucleotide substitutions, deletions, and insertions were found, including the mutations presented here. From this compilation, it is evident that certain nucleotide positions are frequently effected by mutations. The summary of the nucleotide changes at these positions is presented in Table 2. Some of the recurrent mutations observed in our study (positions 123 and 186) were also observed in previous studies,15 21 but previously they had not been noted to be recurrent.

All the 8 mutated lymphoma samples were heterozygous (Table 1) with retention of germ line sequence in 6. In 5 of these samples, an additional 2 or more distinct BCL-6 gene sequences were observed. In these samples, the retained germ line BCL-6 gene sequence may have originated from the unmutated BCL-6 gene within some of the tumor cell subclones or from normal cells infiltrating the tumor. The unmutated DLBCL cases were homozygous, whereas the unmutated FCL sample was heterozygous for 397(G or C) polymorphism.

In 6 NHL samples (5 FCL and 1 DLBCL), the nucleotide sequences showed evidence of intraclonal heterogeneity (Table 1) consistent with a stepwise ongoing mutational process affecting the BCL-6 gene in the tumor cells. Representative examples of these mutations and the intraclonal heterogeneity of the tumor clones are demonstrated in Figures 3 and 4.

By examining the sequences from a given case for the presence of the polymorphism or the pattern of mutations, we could assign the 2 alleles of the tumor. Mutations occurred in both alleles in 6 of the 8 mutated lymphomas. Moreover, in 3 tumor samples, identical mutations affecting both alleles were observed (Figure 3).

After recent reports on somatic mutation in the 5′ noncoding regulatory region of the BCL-6 gene, we examined this region for mutation in lymphomas by evaluation of multiple molecular clones of the BCL-6 gene from each lymphoma sample. In previous reports, BCL-6 mutation analysis was performed by the technique of single strand conformational polymorphism (SSCP). Although the SSCP method is rapid, it is less sensitive for the detection of mutations than extensive actual sequencing of molecular clones, and it may bias the analysis or underestimate the extent of the mutational process.23 By applying a sequencing approach, we demonstrate (a) a germ line sequence different from that previously published12; (b) a polymorphism of G-C at position 397; (c) ongoing BCL-6 gene mutation in 6 of the 11 tested NHL samples; and (d) the presence of recurrent mutations observed repeatedly in lymphoma samples from different patients.

An ongoing mutation process at the Ig locus is a hallmark of germinal center derived lymphomas, such as FL,24,25 Burkitt lymphoma,26 and a subset of DLBCL.27,28 In the current study, we have shown that the untranslated regulatory region of the BCL-6 gene – the single non-Ig gene currently known to undergo somatic hypermutation,29 also shows intraclonal heterogeneity in some FCL and DLBCL cases. The extent of the intraclonal heterogeneity is conservatively estimated here because we excluded single unconfirmed substitution mutations from the analysis. The functional significance of this intraclonal divergence of the BCL-6 regulatory region is currently unknown. Previously, it had been shown that some mutations can significantly deregulate BCL-6 expression.15 Therefore, it is possible that, like in the Ig genes, in which mutations are selected by the antigenic stimulation permitting preferential lymphocyte survival,24 BCL-6 gene mutations could be selected by growth advantage afforded by the lymphomatous state. Ongoing BCL-gene mutations may eventually lead to a subclone with the BCL-6 gene as deregulated as if it were translocated to another chromosome and provide a critical event in lymphoma progression or in the conversion of FL to a higher grade malignancy.

The observed incidence of the BCL-gene mutations and the excess of transitions over transversions are in agreement with previous reports.14-16 However, we have detected the presence of recurrent mutations and identical biallelic mutations in some of the lymphoma samples. Identical biallelic mutations in the BCL-6 gene were recently reported in 2 cases of primary effusion lymphoma.21 One potential problem with comparing gene sequences from different individuals is that of inherited polymorphism. In the case of the human BCL-6 hypermutational region, there is only 1 published DNA sequence.12 The normal sequence that we determined has a total of 7 differences from the published sequence, none of which occurred at the positions of recurrent substitution mutations observed in our tumor samples. To examine the possibility of polymorphism, we sequenced the BCL-6 gene from T cells derived from 2 tumor samples that exhibited the 4 recurrent changes. Only the 397(G→C) change was observed in these cells, thus representing a polymorphic change, whereas 186(T→C), 123(C→T), and (Δ T 520) are nonpolymorphic recurrent mutations observed in our NHL cohort. Furthermore, review of the literature15 19-22confirmed the recurrent occurrence of the mutations at positions 123 and 186 and demonstrated presence of recurrent mutations at additional positions (Figure 2 and Table 2).

The significance of these recurrent mutations is unknown. It is possible that they occur at mutational hot spots similar to the hot spots present in Ig genes.17 However, identical types of mutation at the same position in more than 1 lymphoma sample argues not just for hot spots but for hot specific events. That is, there must be a propensity to change not just at certain position but to make certain specific mutations. These recurrent mutations could result from mutational events that for mechanistic reasons are more probable, or they could be selected for by their effect on the regulation of BCL-6 gene, leading to expansion and survival of tumor clones containing these mutations. The nucleotide motifs predisposing to these recurrent mutations in the BCL-6 gene are still unknown, but most probably they are different from the mutational hot spot motifs of the Ig gene, such as RGWY and dinucleotides GC and TA.17 18 To evaluate the potential influence of these mutations on BCL-6 regulation, we searched for possible transcription factors binding sites in and around the sites of recurrent mutations observed by us (TFSEARCH on the Internet). Mutations 186(T→C) and 123(C→T) occurred in the vicinity of potential binding sites of GATA 2 and PBX-1 transcription factors, respectively. Whether these transcription factors have any role in BCL-6 gene regulation is unknown.

In conclusion, the current study shows intraclonal heterogeneity of the BCL-6 gene regulatory region in NHL samples due to the process of ongoing somatic mutation. Furthermore, our data suggest the presence of mutational hot spots in the BCL-6 gene. Further studies are needed to determine the functional significance of recurrent mutations and the potential role of the BCL-6 gene regulatory region mutations in tumor progression and transformation.