IgM-MM is predominantly a pre–germinal center disorder and has a distinct genomic and transcriptomic signature from WM

Immunoglobulin M (IgM) multiple myeloma (MM) is a rare condition with an estimated incidence of <0.5%.¹ Its differentiation from more common IgM-producing plasma cell (PC) disorders, chiefly Waldenström macroglobulinemia (WM), is challenging but essential because of distinct treatment modalities and prognoses.² Recent advancements in molecular techniques have shed light on the genomic characteristics and unique genetic alterations of MM and WM^3,4; however, comprehensive profiling has not yet been performed for IgM-MM. This study investigates genomic and transcriptomic characteristics of IgM-MM through whole-genome (WGS) and transcriptome sequencing of IgM-MM cases, comparing them to non–IgM-MM, WM, and healthy donor PC.

All MM patient samples and deidentified clinical data were collected after written informed consent. We performed and compared WGS on CD138-sorted myeloma cells from 15 IgM-MM, 211 non–IgM-MM, and 55 WM patient samples, and transcriptomic data from 15 IgM-MM, 30 non–IgM-MM, 35 WM, and 3 healthy donor PCs. We analyzed WGS data with Mutect2, Facets, and Manta, and RNAseq analysis with Salmon, and R (supplemental Methods, available on the Blood Web site).

We first evaluated the WGS data for structural variants and observed that all patients with IgM-MM harbored a t(11;14)(q13;q32) translocation that fused immunoglobulin superenhancers with CCND1 (Figure 1A). Sixty percent of t(11;14) translocations in IgM-MM involved V_H-D_H-J_H regions on IGH (Figure 1B), compared with only 25% of non–IgM-t(11;14)-MM that predominantly involved immunoglobulin switch regions (Figure 1C-D). None of the WM samples featured any translocation (Figure 1E).

IgM-MM displayed copy-number variation patterns similar to those reported in non–IgM-MM (Figure 1F; supplemental Figure 1). We observed clonal changes involving del13q and del6q in 67% and 40% of patients with IgM-MM, respectively, with 1q21 and 6p gains in 40% of 15 patients. Subclonal events involving del16q were observed in 47%, with del1p and del17p in 1 patient each. Del6q and del13 were the only events that were more frequently observed in IgM-MM vs non–IgM-MM and non–IgM-t(11;14)-MM (P < .05).⁵ It is interesting to note that del6q occurred in >31% of WM and is associated with progression from IgM monoclonal gammopathy of undetermined significance (MGUS) to WM.⁶ 

Previously described MM driver mutations⁷ involving KRAS, NRAS, BRAF, CCND1, and DIS3 were also encountered in IgM-MM (Figure 1G). We have observed a high frequency of clock-like signature in IgM-MM and WM compared with non–IgM-MM and high APOBEC in IgM-MM compared with WM, suggesting an alternate mechanism in these groups (supplemental Figure 2).⁸ Matched RNA-seq variant allele frequency (VAF) was linearly correlated with genome-called VAF for other mutations, suggesting a lack of impact on transcription, except for CCND1, which displayed a disproportionate expression of the mutant allele (supplemental Table 1). Interestingly, MYD88 mutation, a hallmark of WM observed in 95% to 97% of patients with WM,⁴ was detected in 1 IgM-MM sample; however, it was a subclonal S219C mutation and not the usual L265P variant observed in WM.⁹ 

RNA-seq data from IgM-MM displayed elevated expression of transcripts relevant to MM-associated proteins, including CD38, CD138, BCMA, and SLAMF7 (Figure 2A), as well as t(11;14)-specific proteins, including CCND1, CD20, CD23, and CD79A (Figure 2B). Furthermore, CD56 and CD117 are highly expressed in MM but low or absent in IgM-MM and WM (Figure 2C). Gene set enrichment analysis showed upregulation of MYC target genes in IgM-MM compared with both non–IgM-MM and WM (Figure 2D; supplemental Figure 3; supplemental Tables 2 and 3).

Unsupervised hierarchical clustering using differentially expressed genes between MM and WM revealed a closer molecular homology of IgM-MM to non–IgM-MM compared with WM (Figure 2E). We identified 230 genes unique to IgM-MM and WM and differentially expressed from MM (Figure 2F; supplemental Figure 4). A similar analysis using only B-cell–specific transcription factors^10-12 showed separation of WM and MM and preferential clustering of IgM-MM with non–IgM-MM (Figure 2G; supplemental Figure 5). WM was noted to have closer homology to peripheral memory and germinal center (GC) B cells compared with MM, including IgM-MM, which appeared to have a more mature phenotype. It appears that IgM-MM has an intermediate phenotype compared with non–IgM-MM, contrasted with WM, which retains a fully immature signature with upregulated PAX5, MYC, and BCL-6 and downregulated PRDM1, XBP1, and IRF4. WM is theorized to originate from peripheral memory B cells, marginal zone cells, or GC B cells,^13-15 which was observed in the unsupervised clustering (Figure 2G).

IGH translocations have been observed in >40% of patients with MM, most frequently involving CCND1 on chromosome 11 (15% to 20%).¹⁶ It has been suggested that different mechanisms in B-cell development are responsible for breakpoints generated on the IGH locus. In fact, the majority (>60%) occur within the switch-region upstream of IGH constant genes, generated through class-switch recombination in mature B cells in GCs.¹⁷ A fifth of t(11;14) translocations, however, are thought to be generated through V_H-D_H-J_H recombination, occurring earlier at the pro–B-cell stage in the bone marrow (BM).¹⁷ All of the IgM-MM samples showed CCND1 translocation, the most probable MM initiating event in this subgroup, and 60% of these appear to originate from V_H-D_H-J_H recombination, suggesting their occurrence in the BM at pre-GC progenitor level. Also, although IgG/IgA-MM displayed evidence of class switch recombination (CSR) with deletions between the IGHM switch region and IGHG/IGHA switch regions, most IgM-MM samples (73%) had no such alterations. It is interesting to note that 2 IgM samples appear to have acquired the t(11;14) translocation during CSR but failed to class switch, which could explain their persistence in the IgM state, and perhaps even more intriguing, 4 samples appear to have fully classed switched and acquired the t(11:14) translocation during CSR, yet the translocated allele is inactivated through an unclear mechanism and the cells exclusively produce IgM. Although both IgM-MM and WM share the same precursor condition, namely IgM-MGUS, the presence of t(11;14) and del13 coupled with the lack of MYD88 mutation at a precursor level may serve as predictors for progression of MGUS to IgM-MM as opposed to WM.¹⁸ 

Three clinically relevant upregulated targets, BCL-2, CD20, and Bruton tyrosine kinase (BTK), were identified in IgM-MM (Figure 2H). We have confirmed that all 15 patients with IgM-MM are t(11;14), and all express CD20 at the RNA level. This is in line with reported CD20 expression in t(11;14)-MM.¹⁹ Moreover, using immunoperoxidase and in situ hybridizations as well as flow cytometric methods, we have confirmed CD20 and cyclin D1 expression in CD138⁺ PCs with monotypic cytoplasmic reactivity for λ light chain and IgM heavy chain in BM from a patient with IgM-MM. This provides the rationale for use of rituxan in patients with IgM-MM. BTK inhibitors, such as ibrutinib, are very effective in treating WM,²⁰ but have only shown modest activity in MM.²¹ BTK expression was significantly higher in IgM-MM and WM compared with non–IgM-MM, indicating a possible role for BTK inhibitors in IgM-MM. Finally, both IgM and non–IgM-t(11;14)-MM had elevated levels of BCL-2, which is a known predictor of venetoclax efficacy.²² Interestingly, the higher BCL2/BCL-XL and BCL2/MCL1 ratios observed in IgM-MM compared with non–IgM-t(11;14)-MM may suggest even better sensitivity to venetoclax in this group of patients.²³ Although certainly not unique to IgM-MM, CD38, BCMA, and SLAMF7 are well-established targets for treatment in MM, and given their high expressions in IgM-MM, similar responses are expected in this subgroup.^24,25 To date, clinical data are lacking, and further investigations are needed to fully understand the potential role of these drugs in treating IgM-MM.

In summary, we here describe a molecular and transcriptomic landscape of IgM-MM, which both identifies its origin at an early stage of PC differentiation and provides clues for a selective therapeutic targeting.

This study was supported by National Institutes of Health, National Cancer Institute grants P01 CA155258 and P50 CA100707, VA Healthcare System grant 5I01BX001584, and a Paula and Roger Riney Foundation grant.

Contribution: A.H.B. designed the research, analyzed the data, and wrote the manuscript; H.A.-L., R.S., A.A.S., Z.H., M.S., J.C., M.F., K.C.A., G.P., S.P.T., and M.M. reviewed the manuscript and discussed the results; A.H.B., H.A.-L., R.S., Z.H., S.P.T., M.K.S., N.C.M. collected the samples and generated the data; M.K.S. and N.C.M. designed the research and wrote the manuscript.

Conflict-of-interest disclosure: K.C.A. has received consulting fees from Bristol Myers Squibb, Celgene, Gilead, Janssen, Precision Biosciences, Sanofi-Aventis, Takeda, and Tolero and serves on the board of directors of and has stock options in Oncopep. N.C.M. is a consultant for BMS, Janssen, Oncopep, Amgen, Karyopharm, Legend, Abbvie, Takeda, and GSK and serves on the board of directors of and has stock options in Oncopep. The remaining authors declare no competing financial interests.

Correspondence: Mehmet Kemal Samur, Dana-Farber Cancer Institute, Harvard T.H. Chan School of Public Health, 450 Brookline Ave, Boston, MA 02215; e-mail: mehmet_samur@dfci.harvard.edu; and Nikhil C. Munshi, Dana-Farber Cancer Institute, Harvard Medical School, 450 Brookline Ave, Boston, MA 02215; e-mail: nikhil_munshi@dfci.harvard.edu.