Waldenström macroglobulinemia neoplastic cells lack immunoglobulin heavy chain locus translocations but have frequent 6q deletions

Waldenström macroglobulinemia (WM) is a B-cell lymphoproliferative disorder characterized by immunoglobulin M (IgM) paraproteinemia and the accumulation of clonal lymphoplasmacytic cells in the bone marrow (BM).1,2 With advancing disease, patients may acquire organomegaly, anemia, and hyperviscosity. A serum monoclonal IgM level that exceeds a concentration of 1.5 g/dL is characteristic of WM and is useful in differentiating WM from other neoplasms with plasmacytoid differentiation.3 The diagnosis is based on the presence of the characteristic accompanying clinical features and the pathologic findings, most commonly from the bone marrow (Table 1). Because of its morphologic and immunophenotypic features, the pathologic designation for WM has been lymphoplasmacytic lymphoma (LPL), as proposed by the revised European-American lymphoma classification.3Because LPL is not always associated with paraproteinemia, it is possible that LPL without paraproteinemia and WM (LPL with IgM paraproteinemia) are distinct pathologic entities.

It has been reported that the t(9;14)(p13;q32) translocation occurs in at least 50% of patients with LPL.4 The t(9;14)(p13;q32) results in the up-regulation of PAX-5,5,6implicating this gene as a putative oncogene in the pathogenesis of LPL.7-9 Other PAX genes are known to be involved in diverse human neoplasia.10,PAX-5encodes for the transcription factor, B-cell–specific activation protein (BSAP), which is a 50-kDa protein critical for B-cell development. BSAP up-regulation appears to cause an increase in B-cell proliferation11 and is characteristically absent in normal and malignant plasma cells.12-15 BSAP/PAX-5 is known to down-regulate the Iε promoter and the IgH 3′α enhancer, with consequent down-regulation of IgH transcription.16-19In the original series of LPL tumors with the t(9;14)(p13;q32) translocation, none of the patients had paraproteinemia, a finding consistent with the possibility that the overexpression ofPAX-5 prevents the expression of high levels of immunoglobulin that result in paraproteinemia. If this hypothesis is correct, one would expect that the t(9;14)(p13;q32) translocation would not be present in WM, a form of LPL with IgM paraproteinemia.

The underlying genetic abnormalities responsible for WM have not been identified. From a limited number of conventional cytogenetic (CC) studies, no recurrent cytogenetic abnormalities have been detected in tumor cells from patients with WM. However, there are sporadic reports of WM tumors with IgH or t(11;18)(q21;q21) translocations or 6q abnormalities, each of which occurs in other kinds of B-cell tumors. Thus, we decided to test a cohort of patients with well-defined WM to address 3 issues. First, despite the expression of IgM in WM, does legitimate or illegitimate IgH switch recombination occur, and perhaps mediate IgH translocations, as previously shown in multiple myeloma and other B-cell tumors?20 Second, using a highly sensitive and specific interphase fluorescence in situ hybridization (FISH) assay, what is the incidence of the t(9;14)(p13;q32) or other IgH translocations in WM? Third, through a combination of CC, multicolor metaphase-FISH (M-FISH), and interphase FISH assays, what is the incidence of recurrent numeric or structural karyotype abnormalities—including the t(11;18)(q21;q21) translocation or abnormalities of 6q21—in WM?

We identified patients who fulfilled the clinical and pathologic diagnosis of WM1,2 (Table 1). The study was conducted under institutional review board approval. Each patient was required to have an IgM paraproteinemia level of 1.5 g/dL or greater and clonal lymphoplasmacytic infiltration comprising at least 20% of the mononuclear cells in the bone marrow (BM), but BM involvement could be lower (10%-19%) if the patient's monoclonal protein level exceeded 3 g/dL. The cohort consisted of 74 patients who fulfilled these criteria (65% had prior chemotherapy exposure). Samples available for study included BM aspirate obtained from 33 patients at the time of routine clinical procurement (all studies with cIg-FISH), fixed cell pellets in Carnoy fixative and cultured for the purpose of clinical karyotype analysis from 17 patients, or both types of samples from 24 patients (cultured and fixed cell pellets and uncultured bone marrow aspirate).

In 19 patients, bone marrow research aspirates were further enriched by simultaneous CD138⁺ and CD19⁺ magnetic bead selection (Miltenyi Biotec, Auburn, CA). Owen et al21,22 have found that clonal cells in WM express CD19, CD138, or both. Because of the pleomorphic nature of the clonal process in WM, we thought that with the combination of CD138 and the B-cell marker CD19, we could enrich for the clonal cells of patients. When we only had stored BM samples previously cultured for metaphase analysis, it was not possible to perform magnetic bead selection or concurrent immunofluorescence staining of the cytoplasmic immunoglobulins for those patients.

This study was conducted under the approval of the Mayo Clinic institutional review board, and patients gave informed consent for the sample collection. The study was conducted in accordance with the Declaration of Helsinki for research with human subjects.

Southern blot analysis was performed according to our previously published technique.20 In brief, we used a J_H probe to detect the germ line and clonal bands to ensure the presence of clonal DNA. We subsequently performed sequential hybridization with probes flanking the 5′ and the 3′ ends of the Sμ switch region. The presence of fully concordant 5′Sμ and 3′Sμ bands indicates the absence of IgH switch recombination. Alternatively, discordant 5′Sμ and 3′Sμ bands that do not hybridize with other IgH switch probes strongly suggest illegitimate IgH switch recombination caused by a chromosomal translocation.20 A total of 2.5 to 10 μg BM DNA (these samples were not magnetic microbead-enriched) was loaded into each well after complete digestion with restriction enzyme HindIII (Figure1)

Conventional karyotypes performed for clinical evaluation were available for 37 patients. BM specimens were processed by direct and short-term culture techniques, as described previously.23At least 20 banded metaphases were analyzed for each patient, with representative karyotypes prepared from at least 2 metaphases from each clone. The karyotype was described according to the International System for Human Cytogenetics Nomenclature (ISCN, 1995).

M-FISH was performed as previously described.24,25Slides with informative karyotypes in 5 patients were processed using the Spectra Vysion (Vysis, Downer's Grove, IL) reagent. Probes for M-FISH were placed on the hybridization site, protected with a coverslip, sealed with rubber cement, and placed in the HYBrite (Vysis) for codenaturation. Slides were viewed with a Zeiss (Thornwood, NY) microscope powered by a 100-W mercury bulb. Filter sets for capturing M-FISH images and viewing metaphases were from Vysis.

We used 2 different probe strategies to detect IgH rearrangements by interphase FISH. We first used a break-apart strategy (segregation), using a cosmid probe (V_H) (labeled red) specific for the IgH variable region and a BAC clone (C_H) specific for the IgH constant (labeled green).26 Under separate experiments and to detect the t(9;14)(p13;q32) translocation, we used a fusion strategy (colocalization), using a BAC clone (clone 112gO2; Incyte Genomics, Palo Alto, CA) specific for thePAX-5 gene at 9p13 (labeled red) (Figure2), together with the V_H/C_H probes (both labeled in green).26 To validate these probes, we tested them in the cell line KIS-1, which harbors t(9;14)(p13;q32) and was kindly provided by Dr Hitoshi Ohno (Kyoto University, Japan) and in a patient sample with t(9;14)(p13;q32) that was kindly provided by Dr Herve Avet-Loiseau (Nantes, France). In both cases our probes showed 2 fusion signals of the IgH probe and the PAX-5 BAC clone, on both derivative chromosomes in each case, indicating its usefulness for the detection of this translocation.

To detect t(11;18)(q21;q21) we used a set of probes, kindly provided by Dr David James (Mayo Clinic), that span both breakpoints in all reported t(11;18)(q21;q21) translocations and that result in a double fusion, as validated by Remstein et al27 (Figure 2). To screen for possible aneuploidy, we used commercially available centromere enumeration probes (CEPs) for chromosomes 7, 9, 11, 12, 15, and 17 (Vysis). To screen for 6q21 deletions, we used the clone RPCI 91C23 (obtained from Oakland Children's Hospital, Oakland, CA) with simultaneous hybridization for CEP 6. This same cohort of patients has also been studied for deletions at 13q14 and 17p13 by interphase FISH, and the results are compared with those obtained from this study.28 

Noncommercial probes were directly labeled using standard nick translation with SpectrumRed or SpectrumGreen (Vysis). Slides and probes were co-denatured for 7 minutes at 80°C, placed in a humidified chamber, and allowed to hybridize in a 37°C oven overnight. Slides were then washed, an antifade mounting medium (Vectashield H-1000) was applied to each, and the slides were coverslipped. For slides not assessed for cytoplasmic immunoglobulin stain, DAPI (Vector Laboratories, Burlingame, CA) was added to the medium.

In all patients (n = 33) for whom cytospin slides were available, we performed cytoplasmic staining of the IgM cytoplasmic protein using an AMCA (7-amino-4-methylcoumarin-3-acetic acid)–conjugated anti-IgM antibody (Vector Laboratories), in a variation of our previously published technique.29 If we could not use the cIg-FISH technique, or if the original cytospin slides were depleted, we used slides made from the cultured and fixed cells. Because of these limitations, these slides were scored on unselected cells without regard to cellular morphology.

Using a Leica epifluorescence microscope with fluorescein isothiocyanate, Texas red, and a DAPI filter, we attempted to score at least 100 clonal cells from each patient. In the cytospin slides, only cells identified by morphologic features or cytoplasmic anti-IgM staining were scored. Fusion of probe signals (colocalization) was defined as 2 probe signals making contact (Figure2). Break-apart of probe signals (segregation) was defined by a distance of more than 3 signal widths between 2 differently labeled probe signals. A sample was said to have an abnormal pattern if the percentage of cells exceeded the mean percentage background level plus 3 SD found for normal cells. The mean percentage background was determined in normal and abnormal samples, and at least 1000 cells were scored for each set of probes.

To detect IgH rearrangements we conducted a pilot study using Southern blot analysis, but we did not find evidence of legitimate or illegitimate IgH switch recombination fragments. In all cases the presence of clonal cell DNA was confirmed by the finding of clonal rearrangements by hybridization with a J_H probe, but there was complete concordance of the 5′Sμ and the 3′Sμ bands in sequential hybridizations (n = 12) (Figure 1).

We reviewed the karyotype results of 45 patients using conventional cytogenetic analysis performed for routine clinical purposes. Six patients were excluded because chromosomal analysis was performed at the time of diagnosis of secondary myelodysplasia or leukemia. Abnormal findings on BM examination and resultant abnormal karyotypes were consistent with the diagnoses. Of the remaining 39 patients, 35 had sufficient numbers of metaphases to be evaluable (Table 2). Of these 35 patients, 22 (63%) patients were found to have only normal metaphases, possibly originating from the normal myeloid elements in the BM, including 2 patients whose only abnormality was the loss of chromosome Y. Abnormal metaphases were obtained in 13 (37%) patients, 5 of whom were not previously treated. Structural and numerical chromosomal abnormalities were encountered in the abnormal karyotypes, but usually we observed single abnormalities in an otherwise normal karyotype (Table 2). For the 13 patients with abnormal karyotype, 4 (31%) had recurrent abnormalities of chromosome 13, with del(13)(q14) in 3 and monosomy 13 in one; 2 (15%) had del(6)(q13q21)(q13q25); 1 (8%) had an addition at chromosome 6q27 (add(6)(q27)); 2 (15%) had abnormalities of chromosome 17q25 and 1 (8%) had chromosome 17 monosomy; 2 had del(5q)(13q35) (one with and one without prior therapy); and 4 had diverse abnormalities in chromosome 8 (t(Y;8)(q11;p23), i8(q10), der(8;9)(q10;q10), and −8). As a follow-up to conventional cytogenetics in 5 of 13 patients with abnormal karyotype, M-FISH studies identified karyotypic abnormalities in 2 patients but provided no additional information in 3 patients (Table 2).

Of the 31 patients studied, 30 had no evidence of IgH translocations by the V_H/C_H strategy, and in one patient 37% of cells harbored the abnormality. In addition, in none of these 31 patients could we find fusion signals indicative of t(9;14)(p13;q32) (Figure 2). When we extended our study to include 18 additional patients in whom we had CC cell suspension to study, we found that none of them had evidence of IgH translocations by the V_H/C_H break-apart strategy (n = 17) or the t(9;14)(p13;q32) translocation by the fusion strategy (n = 18). One patient had 3 or more signals from the PAX-5 BAC clone in 69% of the cells but no fusions (17% with 3 signals, and 52% with 4 signals). This same patient had evidence of trisomy for chromosome 9 in the metaphase analysis. None of the 24 patients studied had evidence of t(11;18)(q21;q21) as detected by FISH.

To screen for possible aneuploidy, tumor cells from 15 patients were studied by interphase FISH using CEP probes for chromosomes 7, 9, 11, 12, 15, and 17. With the exception of one patient, all chromosomes were normal; one patient had only one signal for chromosome 9, indicative of monosomy in 26% of the clonal cells. Although some patients showed abnormalities in the karyotype consistent with aneuploidy (Table 2), interphase FISH showed that this is the exception, and patients with karyotype abnormalities likely have aggressive variants of the disease. There is probably an inherent bias in the patient population that is subjected to scrutiny of karyotype analysis (ie, worsening disease). In addition, it is highly likely, as in multiple myeloma, that the ability to obtain informative metaphases is closely related to tumor burden and proliferative activity.30 

We also tested 24 patients (in whom we had metaphase culture slides but could not perform cIg-FISH assay) for evidence of 6q21 deletions and found that 10 (42%) had abnormalities in more than 25% of the cells. In addition 5 (21%) more patients had between 10% and 25% abnormal cells. Because in many of these latter patients the percentage of clonal involvement of the bone marrow was between 10% and 20%, it is conceivable that they could also have had deletions at this site. Therefore, between 42% and 63% of patients had deletions of 6q21.

Three patients had chromosome abnormalities detected by metaphase analysis; we also performed specific interphase FISH. In one patient −13 was confirmed by FISH, and in another patient the deletion of 17p was confirmed by FISH. In yet another patient, del(6)(q13q25) was detected by our FISH probe.

In this study we have found that clonal cells from patients with clinically defined WM do not have the t(9;14)(p13;q32) translocation, as previously found in 50% of patients with LPL but without paraproteinemia. Thus, WM, a type of LPL with IgM paraproteinemia, and LPL with no paraproteinemia differ not only in phenotype but also in the presence of the t(9;14)(p13;q32) translocation. This result is consistent with the hypothesis that the predicted biologic features of B-cell clones with PAX-5 up-regulation are incompatible with the phenotype of WM.17-19 Notably, BSAP is absent in plasma cells, and the overexpression of PAX-5/BSAP abrogates the production of the J peptide.13 This protein is an integral component of the IgM pentamers, which give rise to the hyperviscosity of patients with WM.31 In addition,PAX-5/BSAP negatively regulates the 3′α enhancer, with the likely consequent inhibition of transcription at the IgH locus,17-19,32,33 a result inconsistent with the IgM paraproteinemia observed in these patients. Finally, the inhibition of PAX-5 transcription with antisense oligonucleotide down-regulates immunoglobulin class switching,11 a result consistent with our failure to identify evidence for legitimate or illegitimate IgH isotype switch recombination in the clonal cells of IgM-producing WM.

Our finding that IgH translocations are not present in WM is consistent with the previously reported karyotypic abnormalities in these patients. In a series of 45 patients with WM reported by Louviaux et al,34 12 patients had abnormal metaphases, but no abnormalities of chromosome 14 were noted. In our review of the cytogenetic database at the Mayo Clinic, which we report here, we were unable to find any patient with 14q32 abnormalities. Unlike multiple myeloma20,35,36 and low-grade lymphomas,37,3814q32 translocations appear not to be initiation events for disease pathogenesis in WM. Scattered reports in the literature describe patients with WM and 14q32 translocations, including t(8;14)(q24;q32).39-42 Some of these reports are confounded either by the lack of consistent clinical features associated with a diagnosis of WM or by the samples originating from pleural effusions and, thus, of unknown relation to the original clone.39-42 

There are rare reports of WM patients in whom tumor cells had a t(11;18)(q21;q21) translocation,43,44 an abnormality also seen in approximately 20% of extranodal marginal zone lymphomas.45 Because we have been unable to detect this translocation in the tumor cells of 24 WM patients, it must be, at best, rare.

Despite an apparent paucity of structural and numeric karyotypic abnormalities in tumor cells from our cohort of WM patients, we did identify a high prevalence of deletions of the long arm of chromosome 6. This is consistent with what has been previously reported in selected patients with WM46,47 and other B-cell neoplasias in which abnormalities in this region are common.48-50 This is also consistent with findings from our karyotype analysis that 3 patients had 6q deletions. Further work is under way to characterize the area of minimal deletion and the search for putative genes involved in disease pathogenesis. We suspect that the inactivation of a tumor suppressor gene at this locus, as is seen in other B-cell neoplasias, is likely to be of importance for clone immortalization.

Our data suggest less genomic instability in WM than in multiple myeloma, as determined by metaphase analyses for structural and numeric abnormalities and interphase FISH for numeric abnormalities. Moreover, we have recently reported that, unlike multiple myeloma and B-cell chronic lymphocytic leukemia (B-CLL), deletions of 13q14 are rarely observed at the time of diagnosis.28 As are deletions of 17p13.1 (see below), 13q14 deletions are seen mostly in advanced stages and in clonal cells in a small fraction (approximately 15%) of patients with WM at the time of diagnosis.28 Therefore, we can conclude that WM is clearly different from myeloma in that it lacks IgH translocations, has infrequent deletions of 13q14, and has limited numeric and structural karyotypic abnormalities. WM is also different from B-CLL in the lower incidence of 13q14 deletions (seen in approximately 50% of patients with B-CLL) and the low prevalence of trisomy 12 (seen in approximately 15% of patients with B-CLL).51 WM is also different from extranodal marginal zone lymphomas in that we could not detect any patient with t(11;18)(q21;q21). The lack of IgH translocations also differentiates WM from many kinds of B lymphomas (follicular, mantle cell, diffuse large cell).52 We thus conclude that the biologic nature of WM is different from that of multiple myeloma and most lymphomas, but it appears similar to postgerminal center B-CLL (Table3).

The normal counterpart of the malignant cell in WM might be a postgerminal center IgM memory B cell that has undergone somatic hypermutation but has failed to undergo isotype class switching. Our Southern blot results, failing to identify legitimate or illegitimate IgH switch recombination rearrangements indicative of isotype class switching, and the presence of the IgM paraproteinemia are consistent with this. In addition, individual WM tumor cells show heterogeneous morphologic features consistent with variable differentiation from a B cell to a plasma cell but failure to fully differentiate into plasma cells. We speculate that the presumed genetic event(s) responsible for generating an immortalized clone of tumor cells in WM are directly or indirectly responsible for the failure of isotype class switching and the lack of differentiation to plasma cells. With the exception of frequent deletions of the long arm of chromosome 6, our knowledge of specific primary or secondary genetic events in WM remains elusive.