Acquisition of CLL-like epigenetic and transcriptomic states in low-count MBL Free

Despite substantial advances in targeted therapies, chronic lymphocytic leukemia (CLL) remains incurable, underscoring the need for strategies to predict or even prevent its development. Low-count (LC-) and high-count (HC-) monoclonal B cell lymphocytosis (MBL) are precursor states of CLL, but the molecular events driving their progression remain poorly understood due to limited characterization of their epigenetic, transcriptional, and proteomic landscapes. To address this gap, we performed a multi-modal single cell (sc) analysis on 34 peripheral blood samples from individuals with LC-MBL (n=6; P1-P6), HC-MBL (n=7; P7-P13), and CLL (n=11; P14-P24), including 10 matched HC-MBL/CLL pairs (median follow-up: 1,926 days, range 916-2,859). We integrated transcriptomic (RNA-seq), epigenetic (ATAC-seq), and surface marker profiling (ASAP-seq) with B cell receptor (BCR) sequencing, mitochondrial DNA (mtDNA) lineage tracing, and copy number variation (CNV) analysis to dissect clonal relationships and cell states.

We first mapped transcriptional and epigenetic states across physiologic B cells, LC-, HC-MBL and CLL using 301,853 scRNA-seq and 168,533 scATAC-seq profiles, which we annotated as MBL/CLL, B, T, NK cells or monocytes. While B cells retained conserved epigenetic profiles, MBL/CLL cells formed unique individual-specific gene expression and chromatin accessibility clusters, even at the stage of LC-MBL. These changes were consistent with CLL-like surface immunophenotypes in LC-MBL, such as overexpression of CD5 and CD200 or downregulation of CD49d, Integrin β7, CD73, and CD35. Single cell BCR analysis demonstrated that LC-MBL retained only small residual polyclonal B cell populations (median: 25% polyclonal; range: 3-95%), which was consistent with the low percentage of CD19⁺ CD5^- non-MBL cells on flow cytometry (median: 21%; range: 0.3-91%). In sum, LC-MBL is characterized by heterogeneous gene regulatory and expression states with early clonal dominance.

We next sought to identify CLL-associated core gene regulatory programs. On differential chromatin accessibility analysis, we found decreased accessibility of transcription factor (TF) motifs implicated in B cell development (i.e., EBF-1, BACH1/2, POU2F2), cell differentiation (AP-1), or chromatin remodeling (SMARCC1), but also increased TF activity in B cell anergy (NFAT family) or stem cell pluripotency (TCF4) shared across the cohort. These distinct profiles were already detectable in LC-MBL and remained stable throughout progression from HC-MBL to CLL. Analysis of transcriptional profiles mirrored these results with increased expression of pro-survival pathways (P2RX7, BCL2, PTPN22, LEF1) and loss of B cell identity (EBF1) or apoptosis regulation (FOS). Together, CLL-like epigenetic and transcriptional states are acquired as early as LC-MBL.

Finally, we wondered what clonal dynamics drive the trajectory from physiologic B cells to MBL/CLL. Comparison of mtDNA mutations between physiologic and monoclonal B cells showed that 245 of 253 (97%) variants were shared. Amongst these, 7 were selectively enriched (i.e., 2435G>A and 3488T>C in P1; 7161G>A in P2; 13886T>C in P5) and 6 depleted (i.e., 8396A>G in P1; 9247G>A and 13453C>T in P2) in LC-MBL. Integration of mtDNA mutations with BCR sequences (nanoranger) and CNVs (numbat) validated these results. In one HC-MBL, targeted genotyping of 3,093 scATAC-seq profiles for SF3B1^Y623C (GoT-ChA) revealed that physiologic B cells (1% SF3B1^Y623C) were enriched for 16186C>T, while MBL (31% SF3B1^Y623C) harbored 1412G>A and 4925C>T. Finally, we used mtDNA mutations to resolve subclonal architectures within MBL/CLL populations, identifying a median of 33 subclones per case (range 4–90). The majority of subclones (median 6 per case [73%], range 3–87) persisted throughout HC-MBL/CLL progression, consistent with clonal proliferation rather than selection. Together, shared mtDNA mutations between physiologic and monoclonal B cells analysis suggest a common ancestry, while their persistence demonstrates clonal stability through the HC-MBL/CLL transition.

In summary, the HC-MBL/CLL transition is characterized by subclonal, epigenetic, and transcriptomic stability, indicating that key transformative events occur earlier, prior to the development of LC-MBL. As a heterogeneous premalignant state, LC-MBL warrants further investigation to define the earliest molecular steps in CLL initiation and identify potential targets for prevention.