Role of the immune microenvironment in the clonal evolution of chronic lymphocytic leukemia

Abstract Background: Chronic lymphocytic leukemia (CLL) is a heterogeneous B-cell malignancy characterized by clonal expansion and immune dysfunction. Although peripheral blood (PB) is frequently used for genomic and transcriptomic profiling, it may not fully capture the biological activity of the disease. The lymph node (LN) microenvironment plays a central role in promoting malignant cell survival, proliferation, and immune evasion. Genomic alterations such as TP53, NOTCH1, and SF3B1 mutations are associated with adverse outcomes, but it remains unclear how these mutations interact with the local immune contexts across compartments. We investigated how clonal architecture and immune composition differ between PB and LN, and how these features relate to evolutionary dynamics. Methods: We performed whole exome sequencing (WES) and bulk RNA sequencing on paired LN biopsies and CD19⁺-purified PB samples from 26 patients with CLL. We selected samples from patients with known CLL driver mutations. Somatic variants, copy number alterations, and clonal architecture were inferred using SuperFreq, and cancer cell fraction (CCF) was used to define clonal distribution. Variants were filtered to include only those with 3 alternate reads, 20 total reads, a variant allele frequency (VAF) 0.02, and classified as non-synonymous mutations. CCF was considered equivalent to VAF if 0.6, otherwise CCF was calculated as VAF * 2. CCF was adjusted by dividing CCF by the ancestral or founding clone CCF as a proxy for tumor cell fraction. Clones were considered compartmentalized if the absolute CCF difference between LN and PB exceeded 0.25. Exome alteration was calculated by summing the length of exonic targets falling in copy-number segments with an absolute log2 copy-ratio 0.2 and dividing by the total captured exome length. LN immune composition was estimated from bulk RNA-seq data using CIBERSORTx. Results: Of the 26 patients, 12 were treatment-naive and 18 were IGHV unmutated. In the PB, mutational frequency was as follows: 46.1% NOTCH1, 42.3% TP53, and 26.9% SF3B1. Oncoplot analysis revealed the presence of multi-hit alterations and a predominance of missense and frameshift variants. Analysis of paired PB and LN samples revealed notable heterogeneity in clonal architecture. Using SuperFreq, we identified between 2-6 clones per patient (median: 4). The median number of subclones did not differ between LN and PB (4 vs. 4), treatment naive vs. relapsed/refractory status (4 vs. 4) or between IGHV unmutated vs mutated (4.5 vs. 4). We evaluated genomic instability and found significantly higher exome alteration in patients who were IGHV unmutated (P = .03). Clonal compartmentalization (CCF difference of |.25|) was observed in 18 of 26 (69%) patients, with 5 patients showing clonal evolution in the LN (28%), 6 in the PB (33%), and 7 with distinct clones evolving in both LN and PB (39%). SF3B1 mutations were common in clones with CCF trending higher in the LN compared to PB (median ΔCCF 0.16, P = .11). Interestingly, CCF of NOTCH1 mutations was balanced across compartments (median ΔCCF = -0.02; P = .36), and CCFof TP53 mutations was numerically slightly lower in LN than PB (median ΔCCF –0.05; P = .97). LN samples with stable clonal populations had higher proportions of CD4⁺ naïve T cells compared to those with evolving clones (P = .02), but no other significant changes in T cell populations were noted. Patients with NOTCH1 or SF3B1 mutations had reduced dendritic cell abundance in LN (P = .04 and P = .06, respectively) compared to those with TP53 mutations or no known driver mutations. Conclusions: This study demonstrates distinct clonal and immune features between PB and LN compartments in CLL, a heterogeneous clonal architecture across compartments. Differences in clonal composition across LN sites within individual patients suggest that localized immune contexts play a role in clonal selection. Evolving LN clones were more common in LNs with relative decreases in CD4⁺ T cell subsets, and the presence of common driver mutations was associated with reductions in dendritic cell content. Together these observations suggest an interplay of supportive vs restrictive roles of the LN TME in CLL progression. These findings highlight the added biological insight provided by LN sampling and suggest that PB alone may not fully capture the clonal heterogeneity of CLL.