Distinct molecular subtypes in Waldenström macroglobulinemia Free

In this issue of Blood, Gagler et al¹ report a single-cell multi-omic study of MYD88-mutated (MYD88^Mut) Waldenström macroglobulinemia (WM), identifying 2 distinct subtypes of disease: memory B-cell–like (MBC-like) and plasma cell–like (PC-like). The MBC-like subtype showed a blockade in differentiation at the memory B-cell stage, whereas the PC-like subtype showed partial differentiation toward a plasma cell. Among their key findings are that CXCR4 mutations (CXCR4^Mut), along with MAP3K14 and ARID1A mutations, were prevalent in the MBC-like subtype, whereas chromosome 6q deletions (del6q) occurred in the PC-like subtype. Distinct mutations and transcriptional and epigenomic features were also observed in both MYD88^Mut WM subsets.

Although MYD88^Mut occurs in 95% to 97% of patients with WM, evidence for the existence of at least 2 distinct subtypes within this population has been building over the past 10 years. This distinction was first suggested by differences in signaling dynamics and clinical presentation based on CXCR4 mutation status.^2,3,CXCR4^Mut, which occur in 30% to 40% of patients with MYD88^Mut WM, show relative exclusivity with del6q, which are found in 40% to 50% of patients with WM.^4,5 The del6q region includes many genes with important regulatory functions for B-cell receptor, apoptosis, BCL2, and NF-κB signaling.⁵ Roos-Weil and colleagues⁶ also reported the existence of 2 distinct WM clusters in patients with MYD88^Mut WM based on CpG methylation analysis. One subset subtyped with memory B cells, and the other with plasma cells. CXCR4^Mut were also enriched in the memory B-cell subgroup, whereas del6q occurred almost exclusively in the PC-like subgroup, consistent with Gagler et al. More recently, Sklavenitis-Pistofidis and colleagues⁷ used single-cell RNA sequencing to show 2 subtypes of MYD88^Mut WM. Their subtypes were associated with either CXCR4^Mut or high levels of DUSP22 and CD9 expression. Using a multi-omic analysis of patients with wild-type and MYD88^Mut WM, we also observed 2 methylation clusters: 1 with a subgroup matching healthy donor memory B cells, and 1 clustered with immunoglobulin M multiple myeloma cells and distinct from healthy donor plasma cells.⁸ The 2 clusters were also associated with either an increased CXCR4^Mut rate or the high expression of DUSP22, respectively.

Although Gagler et al have advanced the observations just described, they also added important context to the dysregulated B-cell differentiation machinery that underlies the 2 subtypes of WM. They postulated that the differentiation block that characterizes WM is driven by SBPI1 and SPIB in the MBC-like cluster, and by FOXO1 along with IRF4 in the PC-like WM cluster. The authors also observed that both subtypes expressed a combination of germinal center and activated B-cell programming at the level of motif enrichment and gene expression. Furthermore, they speculated that a germinal center B cell may be the cell of origin, with the ultimate subtype determined by the mutations gained as the clone evolves out of the germinal center. This argument is somewhat undercut by the fact that all major mutational events were seen in both subtypes, albeit at differing rates. Similar findings were also described by Roos-Weil and colleagues in their methylation studies.⁶ 

Taken together, the aforementioned studies support the same conclusion that there are at least 2 subtypes of MYD88^Mut WM, with important differences in their underlying genomic, immunophenotypic, and clinical presentations.^6-8 These differences are likely to impact therapy. Indeed, the expression of CXCR4^Mut, which is enriched in MBC-like WM, has been associated with inferior outcomes with ibrutinib.⁹ However, changes in signaling that accompany the evolution from early WM into either MBC-like or PC-like subtypes are also likely to impact outcomes of targeted therapies.^8,10 Therefore, both the evolutionary state (E score) and subtype of WM need to be investigated in prospective clinical trials. Important to the advancement of E score and subtype-specific investigation for WM will also be the development of clinically adaptable molecular or pathological tools for evaluation.

In conclusion, the era of molecular subtyping in WM appears at hand and it offers the promise for development of more individualized prognostic and predictive testing as well as treatment approaches for WM.

Conflict-of-interest disclosure: S.T. and Z.R.H. have been named as inventors of MYD88 and CXCR4 testing for WM and have assigned all interests to their institution.