Association of Ig/BCL6 translocations with germinal center B lymphocytes in human lymphoid tissues: implications for malignant transformation

The proto-oncogene B-cell lymphoma 6 (BCL6), a critical transcriptional repressor for germinal center (GC) reaction,¹  is a frequently targeted partner for chromosomal translocations (CTs) associated with various lymphomas such as diffuse large B-cell lymphoma (DLBCL) and follicular lymphoma (FL).^1-4  Reciprocal CTs between the immunoglobulin (Ig) gene locus on 14q32 and the BCL6 gene on 3q27 (t(3;14)(q27;q32)) are found in a significant portion of DLBCL and FL,¹  and they strongly implicate Ig/BCL6 CT in the development of these lymphomas. Ig/BCL6 CT primarily occurs between the Ig switch regions and the noncoding exon 1–intron 1 region of the BCL6 gene,^2-6  presumably mediated by the similar mechanism for Ig class switch recombination (Ig CSR).⁷  As a result, expression of the BCL6 from the translocated locus comes under the transcriptional control of Ig region promoter/enhancer, a genetic alteration known as promoter substitution.⁶  The replacement of the natural BCL6 promoter, which is subjected to tight control during GC reaction,⁸  by the Ig promoters/enhancers results in dysregulation of BCL6 expression.⁶  As BCL6 expression is critical for B-cell proliferation, differentiation and survival in GC, the deregulated BCL6 expression is believed to contribute to B-cell malignant transformation for a subgroup of DLBCL and FL.^6-9  The long-held speculation that deregulated BCL6 expression driven by Ig promoters/enhancers plays an important role in B-cell malignant transformation during lymphomagenesis gains support from recent data showing in mice models that Ig promoter/enhancer-controlled BCL6 expression promotes lymphomagenesis.^10,11 

Although Ig/BCL6 CT has been implicated in the lymphomagenesis as related to DLBCL and FL, it is not clear what the exact role of the CT is during the malignant transformation, and whether the Ig/BCL6 CT per se is sufficient to lead to the malignant transformation in the context of human B cells. These important questions regarding the biology of the Ig/BCL6-translocated cells prior to their malignant transformation have not been previously addressed because of the lack of an assay capable of detecting and measuring such rare Ig/BCL6-translocated cells in their precancerous stages. We reasoned that one way to gain insight into the role of Ig/BCL6 translocation in B-cell malignant transformation and lymphomagenesis is to examine the Ig/BCL6 translocations prior to malignant transformation. In this paper, we report the results of using a novel system to detect and measure rare Igμ/BCL6 CT events in mixed cell populations and to characterize the CT cells and quantify their frequencies in lymphoid tissues of otherwise healthy humans.

Human tonsils and spleens were collected following tonsillectomy or splenectomy performed at the UCLA Center for Health Sciences and Kaiser Permanente Medical Center at Panorama City, CA, after obtaining appropriate institutional review board approval. Human peripheral mononuclear cells (PBMCs) from healthy donors were obtained from the Core Virology Laboratory of the Jonsson Cancer Center at UCLA. Diffuse large B-cell lymphoma tissues were previously collected from patients with this diagnosis and stored. Informed consent was obtained in accordance with the Declaration of Helsinki. The human B-cell line Ramos 2G6 was purchased from ATCC (Manassas, VA).^12,13  BL-2 and DG75 cells were provided by Dr Q. Pan-Hammarstrom (Karolinska Institute, Sweden),¹⁴  and CL-01 was provided by Drs P. Casali (UC Irvine, CA) and A. Cerutti (Cornell University, NY).¹⁵  All the cells and cell lines were cultured or maintained in complete RPMI 1640, as described previously.¹³ 

Tonsillar B cells were purified from tonsil mononuclear cells as described previously.¹⁶  Specific cell populations (B cells and subpopulations, T cells, dendritic cells) were isolated by positive magnetic bead selection (Dynal beads; Invitrogen, Carlsbad, CA) with an appropriate antibody as follows. Antibodies to human CD19 (HIB-19), IgD (1A6-2), CD38 (HIT2), CD77 (5B5), CD3 (HIT3a), and CD209 (DCN46) (BD Pharmingen, San Diego, CA) were incubated with B cells (for B-cell subpopulations) or total tonsillar cells (for T cells and dendritic cells) for 30 minutes at 4°C with agitation, followed by incubation with goat anti–mouse IgG– or anti-IgM (for CD77)–coated Dynal beads at a ratio of 1 projected target cell per 3 beads for 1 hour at 4°C with agitation. The magnetic bead attached cells were selected 3 times with a magnet stand and subsequently subjected to RNA or DNA preparation. In some experiments, the CD38⁺, IgD⁺IgM⁺, CD10⁺, or CD77⁺ B cells were isolated by fluorescence-activated cell sorting (FACS) that was labeled with anti-CD38–APC, anti-IgM–PE, anti-IgD–FITC, anti-CD10–FITC (HI10a), or anti-CD77–FITC.

Total RNA was extracted by Trizol reagent (Invitrogen) as described.¹³  Total RNA (1 μg) or RNA from a given number of cells as indicated was reverse transcribed to cDNA by murine Moloney leukemia virus (MMLV) transcriptase priming with oligo-dT₁₅, as previously described.¹³  Detection of the chimeric Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts was performed with a nested PCR protocol with regular Taq polymerase in the presence of 2.5 mM MgCl₂ (Promega, Madison, WI).¹³  The first-round PCR was conducted at 95°C for 2 minutes to inactivate the MMLV, followed by 95°C for 15 seconds, 65°C for 30 seconds, and 72°C for 30 seconds for 40 cycles. For the second-round PCR, 2 μL of the first-round PCR products was served as DNA templates, and the PCR was conducted at 95°C for 15 seconds, 68°C for 30 seconds, and 72°C for 30 seconds for 40 cycles. The 5′ primer Iμ3.1 (5′-CGACTGGCTGCTCAGGCCCCAGC-3′) and 3′ primer BCL6-3C4.1 (5′-ACTGGCATGGCGGGTGAACT-3′) were used for first-round, and the 5′ primer Iμ4 (5′-TAATGGACTTGGAGGAATGATT-3′) and 3′ primer BCL6-3C1 (5′-GATACAGCTGTCAGCCGGCGAGGCCA-3′) for second-round PCR for the amplification of Iμ-BCL6E2 transcripts. For BCL6E1-Cμ1 transcripts, the 5′ primer BCL6E1a (5′-AATTGAGCTCTGTTGATTCTTAG-3′) and 3′ primer CHμ1.2 (5′-CCTCTCAGGACTGATGGGAAGC-3′) were used for the first-round, and 5′ primer BCL6E1 (5′-GTTCTTAGAAGTGGTGATGCAAG-3′) and 3′ primer CHμ1.0 (5′-TGGAGGAGAAAGTGATGGAGTCG-3′) for the second-round PCR. To monitor the cDNA loading, GAPDH was simultaneously amplified with the first-round PCR with a forward primer (5′-TGAAGGTCGGAGTCAACGGATTTGGT-3′) and a reverse primer (5′CCATGTAGTTGAGGTCAATGAA-3′) that amplify a 120-bp product. For detection of Igμ germ-line transcripts (Iμ-Cμ transcripts), the primers Iμ3.1 and Cμ1.1 (5′-TGCTCTGATGTCAGAGTTGTTC-3′) were used to amplify a 255-bp product. For BCL6 transcripts, the primers BCL6E1 and BCL6-3C4.1 were used to amplify a 147-bp product. The PCR product was electrophoresed with 2% TBE agarose gel. To prevent possible cross contamination, cell manipulation, RNA preparation, and cDNA synthesis were performed in a different location from that of PCR product analysis. An aliquot of PCR master mix without template DNA was always included as negative control for both first- and second-round PCR.

For quantification of chimeric Iμ-BCL6E2 transcript–expressing cells, total tonsillar cells were 2-fold serially diluted in duplication starting at 4 × 10⁶ cells. The RNA prepared from each cell dilution was subjected to RT reaction and PCR with the method described in the preceding paragraph. The end dilution with at least one positive signal from the duplicates was scored as positive and arbitrarily defined as one translocation event. The translocation frequency was normalized to event(s) in every 1 × 10⁶ total tonsillar cells.

Genomic DNA was prepared from DG75 cell line, DLBCL cells, tonsil and spleen B cells, and isolated cell subpopulations using DNA extraction kits (Promega), as previously described.¹³  DNA (400 ng/reaction) was amplified using an expanded PCR kit (Roche Applied Science, Indianapolis, IN). The primer pair Sμ3B (5′-ACTCAGATGGCTAAACTGAGCCTAAGCT-3′) and BCL6-5B (5′-CTGCACCGCCTCCTCTATATCTC-3′) was used for first-round PCR under the following conditions: 95°C for 5 minutes, 60°C for 5 minutes, and 72°C for 10 minutes for one cycle followed by 95°C for 30 seconds, 65°C for 1 minute, and 72°C for 5 minutes for 40 cycles. For the nested PCR, 2 μL of the first-round PCR product was the source of the DNA templates, and the primer pair Sμ3C (5′-AGCCTGAGCTAACAGGCTGAACT-3′) and BCL6-4.9B (5′-GTGGGTTGTGAGTGTCATTACCCTGT-3′) was used under the following conditions: 95°C for 30 seconds, 68°C for 1 minute, and 72°C for 5 minutes for 40 cycles. PCR bands were purified from 0.8% agarose gels and cloned into the Topo-zero blunt cloning vector (Invitrogen). The plasmids were sequenced with M13 forward and M13 reverse sequencing primers.

The CT between the Igμ gene on chromosome 14 and the BCL6 gene on chromosome 3 in cells from a subset of DLBCL patients generates a fusion gene that expresses chimeric Iμ-BCL6 (Figure 1)⁶  or VDJ-BCL6 transcripts.¹⁷  Since any Igμ/BCL6 translocation occurring within the germ line or switched Igμ switch (Sμ) region (∼ 5 kb for germ-line Igμ, and various sizes for Sμ-Sx switched locus, where Sx is a switch region of the downstream Ig isotype)¹⁸  and the exon 1–intron 1 region of the BCL6 gene (∼ 10 kb)^5,6  should uniformly produce the chimeric Iμ-BCL6 transcripts derived from the splicing of the Iμ exon to the exon 2 of the BCL6 gene,⁶  we reasoned that such chimeric transcripts could serve as a marker for the rare Igμ/BCL6 translocation events that we predict occur in heterogeneous cell populations prior to their malignant transformation.

To test this possibility, we established a nested RT-PCR assay to detect the expression of the chimeric Iμ-BCL6 transcripts. As shown in Figure 1 (top right panel), the assay was able to detect such Iμ-BCL6 transcripts from both human tonsillar and splenic B cells. Cloning and sequence analysis demonstrated that the major RT-PCR products (145 bp) from the human tonsil and spleen cells were derived from the accurate splicing of the Iμ exon of the Igμ gene to the exon 2 of the BCL6 gene as has been reported with Igμ/BCL6-translocated DLBCL⁶  (hereafter referred as to Iμ-BCL6E2 transcripts) (Supplemental Figure S1A, available on the Blood website; click on the Supplemental Figures link at the top of the online article). A minor RT-PCR product (106 bp) that represents alternative splicing of the Iμ to the exon 3 of the BCL6 gene (hereafter referred as to Iμ-BCL6E3 transcripts) was also observed from most tonsil- and spleen-cell preparations (data not shown; see Figure 2 for examples). These data show that the Iμ-BCL6 transcripts, the product of one side of reciprocal Ig/BCL6 translocations in chromosome der(3) that are expected to lead to the deregulated expression of BCL6, can be detected in normal human tonsillar and splenic B cells.

Igμ/BCL6 CT, as defined in lymphomas, is a balanced translocation between the Igμ gene locus on 14q32 and the BCL6 gene on 3q27. Such a reciprocal CT also creates a fusion gene comprised of the 5′ region of the BCL6 gene, including the BCL6 promoter and BCL6 exon 1 region, juxtaposed to the 3′ region of the Igμ gene in d(14) chromosome, including the 3′ Sμ region and the Igμ heavy chain constant region (Cμ) gene (Figure 1). As a result, the Cμ gene comes under the transcriptional control of the BCL6 promoter, and this fusion gene would theoretically produce chimeric RNA BCL6E1-Cμ transcripts representing splicing from the exon 1 of the BCL6 gene (BCL6E1) to Cμ gene. Such transcripts have been detected from DLBCL with Ig/BCL6 CT.¹⁹  As existence of the reciprocal BCL6E1-Cμ transcripts in our cells would provide independent evidence to corroborate the genomic origin of the Igμ/BCL6 translocation–derived chimeric Iμ-BCL6 transcripts, we established an RT-PCR assay to detect BCL6E1-Cμ transcripts in human tonsillar B cells. Indeed, BCL6E1-Cμ transcripts (190 bp) were detected from the same human tonsil B-cell preparations that were positive for the Iμ-BCL6E2 transcripts (Figure 1 lower right panel). Sequencing revealed that in the BCL6E1-Cμ transcripts, the BCL6 exon 1 sequence is accurately spliced to the splicing acceptor site of the exon 1 of the Cμ gene (Cμ1) (Figure S1B). This result demonstrates that the chimeric BCL6E1-Cμ1 transcripts, the products of the other side of reciprocal Ig/BCL6 translocations in chromosome d(14), are present in human tonsillar B cells.

The reciprocal nature of Ig/BCL6 translocations theoretically should lead the CT cells to express both chimeric Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts in the same cells. To determine whether these 2 types of chimeric transcripts are simultaneously expressed, we examined their expression in the same cell populations. As shown in Figure 2A, the BCL6E1-Cμ1 transcripts were detected in the same CD38⁺ B-cell subpopulations that also were expressing Iμ-BCL6E2 from different tonsils (n = 5). In contrast, those cell populations that did not express Iμ-BCL6E2 transcripts, including IgD⁺ B cells, and CD77⁺ and CD38⁺ B cells from tonsil no. 32 (T32), also did not express BCL6E1-Cμ1 transcripts (Figure 2A). These results indicate that both types of chimeric transcripts are correlatively expressed in the same cell subpopulations. Comparison of the assay sensitivities between the Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts revealed that the assay sensitivity for the Iμ-BCL6E2 transcripts was about 9-fold higher than that for the BCL6E1-Cμ1 transcripts (Figure 2B). Therefore, we chose the assay for Iμ-BCL6E2 transcripts as the primary detection method in further studies.

Chimeric Iμ-BCL6E2 or BCL6E1-Cμ1 transcripts can serve as a specific marker for Igμ/BCL6 translocations only if they are transcribed from the translocated chromosomes and do not result from other mechanisms potentially capable of generating such transcripts (for example, trans-splicing between the pre-mRNAs of the germ-line Igμ transcripts [Iμ-Cμ] and BCL6 transcripts). To test whether the chimeric transcripts can be generated from the trans-splicing between the Iμ-Cμ and BCL6 pre-mRNA, we searched for the Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts from several human B-cell lines that spontaneously produce both Iμ-Cμ and BCL6 transcripts. As shown in Figure 3, human Burkitt lymphoma cell lines Ramos, DG75, CL-01, and BL-2 all produce both Iμ-Cμ and BCL6 transcripts, but no Iμ-BCL6 or BCL6E1-Cμ transcripts were detected in nonstimulated or IL-4 plus anti-CD40–stimulated cultures even when using an assay capable of detecting Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts from human tonsil cells that served as a positive control (Figure 3). These results indicate that both Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts are unlikely to be derived by the mechanism of trans-splicing between the Iμ-Cμ and BCL6 transcripts.

To clearly demonstrate that Ig/BCL6 translocations occur in human primary lymphoid tissues, we sought to isolate and identify the recombination junctions of the Igμ/BCL6 translocations with a genomic DNA PCR approach. To optimize this PCR assay for amplification of Igμ/BCL6 translocations from primary human tonsillar B cells, we used previously diagnosed DLBCL cells with a known Igμ/BCL6 translocation as a DNA template for the amplification (Figure 4A). A series of these DLBCL cells diluted in the background of 1 × 10⁵ DG75 cells (equivalent of 400 ng DNA) were subjected to nested PCR amplification. This PCR strategy was reliably able to amplify Igμ/BCL6-translocated products (approximate size of 2.5 kb) from 10 copies of DLBCL cells but not reliably from fewer than 3 copies (Figure 4A). Fifty-eight attempts to amplify Igμ/BCL6-translocated products from the unfractionated tonsillar cells failed. However, from a total of 48 amplifications from CD38⁺-enriched B cells of 3 different donors (tonsil 14, tonsil 16, and spleen 3), we successfully amplified and isolated 3 clones with Igμ/BCL6 translocations (Figure 4C). Sequence analysis revealed that all 3 clones joined their Sμ sequences directly to the intron 1 region sequence of the BCL6 gene (Figure 4D; Figure S2). The minimum nucleotide overlaps in the recombination junctions (1 nt in T14, 3 nt in T16, and 2 nt in S3) suggest that the translocations were mediated through the nonhomologous end joint process responsible for Ig CSR. There was a 213-bp internal deletion of the Sμ region between 587 and 818 of the GenBank accession number X56795 in T14. Two mutations in the vicinity of the recombination breakpoints were found in 1 of the 3 clones (T16). These results, which show the presence of the recombination breakpoints, further demonstrate with an additional approach that Igμ/BCL6 translocation occurs in normal human primary B cells.

To further analyze the Igμ/BCL6 translocation events in normal human lymphoid tissues and in circulation, we investigated the distribution of the Iμ-BCL6E2 transcripts in human lymphoid tissues and PBMCs. We detected Iμ-BCL6E2 transcripts in 39 (96%) of 41 randomly collected human tonsils (Figure 5A; Table 1) and 3 (100%) of 3 human spleens (Figure 5B) by using a total of 1 × 10⁶ tonsillar or splenic cells. T32 tonsil was the only sample that has been confirmed to be negative for both types of chimeric transcripts from both total tonsil cells and CD38⁺ B cells through multiple tests (Figure 2A; Table 1). We could not confirm the negative Iμ-BCL6 expression results for T9 (Figure 5A) with CD38⁺ B cells, as it was a previously frozen sample. These results show that Igμ/BCL6-translocated cells are present in lymphoid tissues in the vast majority, if not all, of otherwise healthy individuals. However, peripheral-blood mononuclear cells (PBMCs) from 14 donors were negative when 1 to 2 × 10⁶ cells were analyzed for Iμ-BCL6 transcripts (Figure 5C) (“Discussion”).

Analysis of magnetic bead selected cells was undertaken to define the distribution of Iμ-BCL6E2 transcripts with B-cell subpopulations. Iμ-BCL6E2 transcripts were consistently detected in CD19⁺, CD38⁺, and IgD^– B-cell populations, as well as in total tonsillar cells, but not in CD3⁺ or CD209⁺ tonsillar-cell populations (Figure 5D; Table 1; n = 7). The association of Iμ-BCL6E2 with CD77⁺ marker was less consistent, as 5 of 7 enriched CD77⁺ B cells were positive (Figure 2A; and data not shown). Using FACS sorting to extend these studies, Iμ-BCL6E2 transcripts were detected in B cells, CD38⁺, CD77⁺, CD10⁺, and IgM^–IgD^– cells, but not in IgM⁺IgD⁺ or CD38^– B-cell populations (Figure 5E, n = 3). Overall, these results show that the Iμ-BCL6E2 transcripts primarily exist in CD38⁺ and/or CD10⁺ B-cell subpopulations, and indicate that Iμ-BCL6 transcripts are associated with GC-stage B cells, but not naive B cells, T cells, or dendritic cells.

We next sought to define the sensitivity of our assay for the quantitative measurement of chimeric transcripts as a biomarker for Ig/BCL6 translocation in mixed cell populations. As shown in Figure 6A, the assay could reliably detect chimeric Iμ-BCL6E2 transcript in the RNA equivalent of 1 to 3 translocated DLBCL cells (0.25-0.75 pg RNA). This was determined by serially diluting Iμ-BCL6E2–positive DLBCL-cell RNA into the RNA equivalent of 1 × 10⁶ DG75 cells (4 μg RNA), a Burkitt lymphoma cell line that harbors an Ig/c-myc but not an Ig/BCL6 translocation.¹⁴  To estimate the frequency of Igμ/BCL6-translocated cells, we performed a semiquantitative analysis of limiting dilution to determine the frequencies of the Iμ-BCL6E2 transcript–expressed cells in human tonsillar cells. The frequency of the Igμ/BCL6-translocated cells was estimated by using the highest dilution of the input cells able to give a positive Iμ-BCL6E2 signal as “one Ig/BCL6 CT event,” as shown in Figure 6B, where the frequency was 1 event for every 10⁶ tonsillar cells. Hence, one CT “event” is at least one (or more than one) CT cell in the given cell population. Since the transcriptional rate for Iμ-BCL6 transcripts may be different between DLBCL and primary B cells carrying an Igμ/BCL6 CT, the assay sensitivity obtained from the reference DLBCL cells may not exactly apply to the primary B cells. The frequency was further normalized to the number of Iμ-BCL6E2 expression events in every 1 × 10⁶ cells. Thirteen tonsillar cells analyzed with this approach showed that the frequencies of the Igμ/BCL6-translocated cells were in the range of 0.5 to 16 CT events in every 1 × 10⁶ tonsil cells (Table 1).

Analysis of Ig/BCL6 CT cells in humans prior to their full malignant transformation has not been possible previously due to the lack of an appropriate system capable of detecting and/or measuring Ig/BCL6-translocated cells in mixed cell populations. Previously developed Ig/BCL6 CT detection systems used for tumor tissue analysis are simply not sensitive enough to effectively detect and/or quantify the rare precancerous cells with this CT.²⁰  To better understand the biologic nature of the Ig/BCL6 CT cells prior to their malignant transformation, we established a sensitive RT-PCR assay for detection of the chimeric Iμ-BCL6 transcript as a marker reflecting Igμ/BCL6 CT. Because trans-splicing mechanism is not likely responsible for the production of the observed chimeric transcripts, the chimeric transcripts are therefore the reliable markers for the Igμ/BCL6 CT. Indeed, several lines of evidence—including the coexpression of both chimeric Iμ-BCL6E2 and BCL6E1-Cμ1 transcripts reflecting the reciprocal Igμ/BCL6 CTs from the same B-cell subpopulation, the special association of the chimeric transcripts with GC markers, and the isolation of the Ig/BCL6 DNA recombinational products from GC B cells—further support the notion that the chimeric transcripts can serve as the markers for Ig/BCL6 CT in mixed cell populations.

It is generally believed that a CT is a critical minimum factor in the development of many types of leukemias/lymphomas, as many such CTs are strongly associated with specific leukemias/lymphomas.²¹  However, several types of CTs per se are known to be insufficient to result in full lymphomagenesis (for example, Ig/BCL2 [t(14;18)(q32;q21)],²²  Ig/c-myc [t(8;14)(q24;q32)],²³  and NMP/ALK [t(2;5)(p23;q35)])²⁴  in humans and various types of CT in mice.²¹  These latter data support a “2-hit” hypothesis,²⁵  in which another factor(s), including genetic or epigenetic alterations in addition to the CT per se, is required for malignant transformation in these specific CT-associated lymphomas. However, one cannot simply extrapolate this outcome to other lymphoma-associated CTs and, especially, in those involved in Ig/BCL6 CT that are frequently associated with DLBCL and FL. With an assay capable of detecting as few as 1 to 3 Igμ/BCL6 CT cells in 10⁶ cells, we were able to identify and to quantify the frequencies of these CT cells in human tonsils, spleen, and blood. Essentially, all tonsil and spleen samples demonstrated the presence of the chimeric transcripts indicative of the Igμ/BCL6 CT, while such cells were not found in PBMCs. Therefore, we showed that Ig/BCL6 CT cells exist in human lymphoid tissues of otherwise healthy individuals. Our data also show that the frequent occurrence of an Ig/BCL6 CT in human lymphoid tissues is not sufficient to result in malignant transformation. This is of particular interest in that the deregulated BCL6 expression driven by an Iμ-promoter is not sufficient to lead to lymphomagenesis in the context of human B cells, while it is able to promote DLBCL-like pathology in a transgenic mouse model.¹¹ 

The association of Ig/BCL6 CT cells with GC phenotypes suggests that the CT occurs at the GC stage of B-cell development (indicating that these CT cells are the products of the GC reaction). This is consistent with the GC origin of DLBCL.⁷  Indeed, the machinery for Ig SHM and CSR, 2 major genetic alteration events for antibody diversification that occur during GC reactions,^26,27  has been implicated in the generation of BCL6 mutations²⁸  and in Ig/BCL6 translocations (this study), as they are characteristics of the activation-induced cytidine deaminase (AID)–dependent Ig SHM and Ig CSR.^6,28,29  Therefore, both BCL6 SHM and Ig/BCL6 CT are likely the byproducts reminiscent of the GC reactions mediated by the machinery for Ig SHM and CSR.

The fate of the cells that carry the observed Ig/BCL6 CT remains to be fully explored and will be of great interest in terms of the potency of the precancerous cells for the subsequent malignant transformation. Those CT cells that do not express B-cell receptors (BCRs) are not expected to survive the GC selection, since CT-associated lymphomas with GC origin usually express BCR and the CT cells are predominantly involved in the nonproductively rearranged Ig alleles.^7,30  Even though the number of Ig/BCL6 CT cells that will survive through GC selection is not clear at this time, a portion of these CT cells is expected to express BCR, and therefore is likely able to survive the GC selection, as there is no typical reason for why the Ig/BCL6 CT should selectively occur in the nonproductive Ig alleles. Potentially Ig/BCL6 CT cells surviving the GC selection and entering the peripheral pool would become memory B cells and, therefore, they probably would have a higher chance of undergoing malignant transformation due to their exposure to antigen stimulation, or genetic or epigenetic alterations. Alternatively, the Ig/BCL6 CT cells could also gain an additional hit for malignant transformation inside the GC, without undergoing GC selection and/or exiting GC reaction, because AID-mediated mutational/recombinational activities for further genetic alterations theoretically could also occur in the GC stages of CT cells. Malignant transformation also could occur when the Ig/BCL6 CT event hit the cells already harboring other genetic alterations, such as mutations in other proto-oncogenes, as such mutations are frequently associated with DLBCL.³¹ 

Once the Ig/BCL6 CT cells survive the GC selection, they should exit GC and become memory B cells; therefore it is possible that such CT-bearing memory B cells would appear in the circulation. We could not detect Ig/BCL6 CT cells in PBMCs, but whether this is because they are truly absent or simply too few in percentage of total PBMCs remains to be addressed. The total amount of PBMCs used (1-2 × 10⁶ cells) likely contains too few post–GC-stage B cells for detection as only about 4% of PBMCs are memory B cells. To identify whether Ig/BCL6 CT cells are present in PBMCs, a specific B-cell population–enriched fraction (such as CD27⁺ memory B cells) needs to be tested.

The correlated expression of both types of chimeric transcripts in the same cell subpopulations not only provides support for the notion that the chimeric transcripts are the sensitive markers for this type of CT, but also suggests that the Igμ/BCL6 CT in primary B cells is reciprocal, as it is present in lymphomas, and indicates that these CT cells are very likely the precursors of those lymphomas harboring Ig/BCL6 CT. However, the conclusion that the Igμ/BCL6 CT in primary B cells is reciprocal can be confirmed at the single-cell level only with appropriate technology such as fluorescent in situ hybridization (FISH), which can formally demonstrate reciprocal CT. Unfortunately, due to the sensitivity limitation of the FISH assay,³²  an attempt to use FISH to demonstrate the reciprocal nature of the CT in primary B cells failed (data not shown).

With this sensitive assay, we will be able not only to detect the rare Igμ/BCL6 CT in lymphoid tissue but also to measure the frequency of this CT as the potential precursor for DLBCL and FL in different samples. This can be done while examining the effects of variables such as age, gender, infection (HIV), chemical exposure, or other potential risk factors. It is also possible that a quantitative assessment of Igμ/BCL6 CT via this chimeric transcript approach will serve as a broad indicator for general genomic instability associated with GC reactions in subjects. Detection and measurement in blood B cells will be a big step forward in this regard. In addition, enrichment of the CT cells will allow study of their growth characteristics in vitro and their further molecular characterization. Finally, our assay will allow the detection of residual CT tumor cells after therapy in appropriate lymphoma patients.

We are grateful to Dr A. Saxon for his continuous support and encouragement, his stimulating discussions, and his critical reading of this paper. We also thank Minna Jyrala and Felix Zhang for their excellent technical assistance; C. Belz for her professional editing of the paper; Drs Q. Pan-hammarstrom, P. Casali, and A. Cerutti for cell lines; and Drs M. Teitell and R. Wall for critical reading of the paper.

The authors have no conflicting commercial interest.