Next-generation sequencing of peripheral B-lineage cells pinpoints the circulating clonotypic cell pool in multiple myeloma

Improvement of long-term outcomes in multiple myeloma critically relies on a better understanding of the tumor-initiating cell in this disease. As in other malignancies, the cancer stem cell concept incriminates drug-resistant myeloma stem cells with tumor-initiating, self-renewing properties to feed the malignant plasma cell (M-PC) compartment in disease relapse and progression.^1,2  However, it remains a matter of controversy if the clonogenic population resides within the pool of terminally differentiated post-switch M-PCs or within a less differentiated (surface) immunoglobulin (Ig)-positive pre-switch B-cell compartment. The latter hypothesis has been fueled by the description of so-called clonotypic pre-switch (IgM⁺) B cells postulated to express the same patient-individual variable region Ig rearrangement as the M-PC.^3-7  This finding even provided the rationale for therapeutic targeting of this postulated CD20⁺ population with the monoclonal antibody rituximab.^8,9  However, inconsistent results of xenotransplantation experiments,^10-15  data on intraclonal evolution at the PC level,^16-22  the analysis of class switch junctions,²³  and the lack of benefit from CD20-directed targeted therapy^8,9  challenged the concept of such feeder cells in myeloma. Moreover, a recent study entirely failed to detect clonotypic rearrangements in highly purified B-cell populations devoid of contaminating M-PCs,²⁴  and our own data also pointed in this direction, because we could not phenotypically detect clonotypic B cells in the majority of patients with patient-individual Ig ligands as tracers.²⁵ 

Here we used next-generation sequencing to definitively confirm or disprove the existence of this highly controversial cell population.

Blood and bone marrow samples of 12 myeloma patients (Table 1) visiting the Freiburg and Hamburg University Medical Centers were obtained after written informed consent as approved by the institutional review boards. This study was conducted in accordance with the Declaration of Helsinki.

M-PC heavy-chain Ig rearrangements were determined as previously described from bone marrow and are shown in supplemental Table 1.²⁵  For detection of circulating clonotypic cells, peripheral blood mononuclear cells were used. Polymerase chain reaction (PCR) approaches for the qualitative detection of clonotypic rearrangements of all isotypes, of IgM isotype only, and for isotype subclass determination by Sanger sequencing are shown in supplemental Figure 1A-B,D and supplemental Table 2, available on the Blood Web site.

Ig transcripts were amplified for next-generation sequencing (NGS) from peripheral blood as described in the supplemental Detailed Methods section. The amplification strategy and primer sequences are shown in supplemental Figure 1C and supplemental Table 2. Amplicons were multiplex-sequenced on a MiSeq Illumina sequencer. Data were plotted using ggplot2 for R statistical programming by display on a matrix providing a specific position for each potential V-D-J rearrangement.

Peripheral blood cells were stained with fluorescently labeled antibodies (CD45-ECD, CD138-PC5, CD38-FITC, CD19-PC7) to discriminate B-lineage cell populations.

Peripheral blood clonotypic rearrangements were qualitatively detectable with HCDR3-specific primers in 6 of 8 patients with active disease, suggesting the presence of circulating clonotypic B-lineage cells (data not shown). None of the cases in remission after treatment was PCR-positive. All HCDR3 PCR-positive cases were also positive in an established semi-nested PCR approach using IgM-specific primers. Surprisingly, however, Sanger sequencing of PCR products revealed post-switch IgG/IgA transcripts despite the use of IgM-specific primers, indicating that the PCR may yield false-positive results in the presence of a dominant clonotypic rearrangement of alternative isotype (data not shown).

This observation prompted us to comprehensively screen Ig repertoires of circulating cells expressing pre- and post-switch isotypes for the presence of the clonotypic rearrangement by NGS. Sequencing data were displayed on a matrix providing a specific position for each potential V-D-J rearrangement (Figure 1). We found highly skewed peripheral Ig repertoires in myeloma patients compared with the polyclonal repertoires of a healthy donor (Figure 1A), reflecting the suppression of healthy polyclonal B-lineage cells by the M-PC clone in the bone marrow. None of the patients showed evidence for IgM⁺ clonotypic rearrangements, suggesting absence of pre-switch clonotypic B cells in the peripheral blood. The clonotypic rearrangement could, however, be detected within the post-switch repertoire of the M-PC clone (IgG or IgA), indicating that M-PC circulate in the peripheral blood of these patients. In the majority of cases, the clonotypic rearrangement was the most abundant rearrangement detectable within the repertoire with a median of 48% of reads of the respective isotype (supplemental Table 3). PCR-negative control cases MM001 and MM020 (Figure 1A) as well as blood samples of patients MM024 and MM031 in complete remission after treatment (Figure 1B) were negative for the clonotypic rearrangement. In patients MM031 and MM048, we found, apart from the dominant clonotypic rearrangement of the M-PC, an identical rearrangement of alternative post-switch isotype (nonclinical rearrangement). Sequencing of the exact isotype subclass revealed a dominant IgG3 clone with a small IgA1 nonclinical isotype in patient MM031 and a dominant IgA1 clone with a small IgG1 nonclinical isotype in patient MM048. Considering the architecture of the Ig gene locus, this suggested that the MM031 IgG3 clone emerged from the preexisting IgA1 clone, while the small MM048 IgG1 clone had to be considered as precursor of the dominant IgA1 clone. This finding suggested ongoing sequential switch events in the progeny of post-switch myeloma cells and therefore clonal evolution at the PC level.

To confirm our assumption that the clonotypic post-switch rearrangements may derive from circulating M-PCs, we analyzed peripheral blood cells by multiparametric flow cytometry. We found evidence of circulating M-PCs with the classical immunophenotype in all but one patient with a clonotypic rearrangement in ∼1% of leukocytes (supplemental Figure 2A). In the control cases MM001 and MM020 and posttreatment samples of patients MM024 and MM031 (supplemental Figure 2B), no circulating M-PCs could be detected. Patient MM021, who had a low burden of clonotypic rearrangements (only 2% of all IgG reads), was the only NGS-positive case with flow cytometric negativity for circulating M-PCs, most likely due to the different sensitivity levels of these assays.

The hypothesis that pre-switch clonotypic B cells act as tumor-initiating and -propagating cells in myeloma has caused considerable debate for over decades. The results presented here should terminate this controversy. Using highly sensitive and highly specific state-of-the-art technology, we disprove the existence of such cells in the blood. This is in accordance with a recently growing body of evidence by other studies.^8-23 

Two methodological aspects most likely account for the discrepancy between prior reports^3-7  and our study. First, PCR with HCDR3- and isotype-specific primers may lack specificity in the presence of abundant clonotypic transcripts of alternative isotype. Without confirmation by direct sequencing of such transcripts, this finding may have falsely suggested the existence of IgM-positive clonotypic B cells, although the transcripts derived from IgG- or IgA-positive M-PCs. Second, many of the previous studies were primarily based on B-cell subpopulations isolated by fluorescence-activated cell sorting. Because M-PCs can circulate in the peripheral blood, such B-cell populations may easily have been contaminated by circulating M-PCs, which may then have provided the clonotypic sequences falsely attributed to B cells. Consequently, when highly purified B-cell populations devoid of contaminating M-PCs were investigated, no clonotypic transcripts were found.²⁴ 

Taken together, our data show that pre-switch clonotypic B cells are inexistent in the blood of myeloma patients, while circulating M-PCs frequently occur in active disease and may contribute to spreading the disease.

This work was supported by the Wilhelm Sander Foundation (grant 2009.035.02 to M.B.).

Contribution: B.T., M.K., M.A., and D.I. performed experiments; M.B., B.T., M.K., M.T., and A.G. interpreted data; and M.B., B.T., and M.K. wrote the manuscript.

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

Correspondence: Mascha Binder, University Medical Center Hamburg-Eppendorf, Department of Oncology, Hematology, BMT with section Pneumology, Martinistrasse 52, D-20246 Hamburg, Germany; e-mail: m.binder@uke.de.