AUTHORS
Yu Xu, Haobo Sun, PhD2, Xingyun Yao3, Chao Hu4, Jie Jin5, Jianping Ge6, Wen-Juan Yu, MD7, Hongyan Tong, MD, PhD8, Xiaofei Gao, Ph.D.9
AFFILIATIONS
1 Department of Hematology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China
2 Westlake University, Hangzhou, Zhejiang, China
3 Westlake University, Hangzhou, Zhejiang, China
4 Department of Hematology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China
5 Timing BioTech, Hangzhou, Zhejiang, China
6 Timing BioTech, Hangzhou, Zhejiang, China
7 Department of Hematology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China
8 Department of Hematology, The First Affiliated Hospital, Zhejiang University School of Medicine, Hangzhou, China
9 Westlake University, Hangzhou, Zhejiang, China
ABSTRACT
Although mature red blood cells (RBCs) are enucleated, they can retain DNA remnants generated during erythropoiesis, such as micronuclei (MN). Erythrocyte MN are well-established indicators of chromosomal instability and are widely used for in vivo genotoxicity testing. We recently developed a genome-wide sequencing approach for the DNA remnants in red blood cells (rbcDNA; Sun et al., Cell Res 2025). Our work in solid tumors further showed that tumor-derived signals can remotely perturb bone marrow erythropoiesis, generating reproducible tumor-associated rbcDNA features. Whether rbcDNA is similarly altered in hematologic malignancies remains unknown. Here, we profiled rbcDNA in diffuse large B-cell lymphoma (DLBCL), the most common aggressive B-cell lymphoma, to identify DLBCL-associated features and assess their associations with clinical risk.
We collected peripheral blood from 25 patients with newly diagnosed DLBCL before treatment and from 25 age- and sex-matched healthy controls. Patients represented disease stages I–IV, International Prognostic Index (IPI) scores of 0–4, and both germinal center B-cell–like (GCB; n=6) and non-GCB (n=19) subtypes. rbcDNA was isolated from 2 mL of peripheral blood using our previously reported workflow and analyzed by low-pass whole-genome sequencing at approximately 1.5× coverage. Normalized sequencing read counts were calculated across genomic regions, and regions that differed significantly between patients with DLBCL and healthy controls were defined as DLBCL-associated rbcDNA features. These features were evaluated by hierarchical clustering and feature prioritization and compared with rbcDNA features previously identified in colorectal, gastric, liver, lung, thyroid, and breast cancers. Prioritized features were assessed for associations with clinical variables, including IPI, lactate dehydrogenase, GCB/non-GCB subtype, and interleukin-6 and interleukin-8 levels, followed by pathway enrichment and genomic functional annotation. In parallel, erythrocyte MN were quantified in the same samples using an optimized flow and imaging cytometric assay with 50 μL of blood.
Unsupervised hierarchical clustering based on the differential rbcDNA features showed distinct grouping of patients with DLBCL and healthy controls within this discovery cohort, and feature prioritization yielded a 67-feature set. The DLBCL-associated features were distinct from those previously identified in colorectal, gastric, liver, lung, thyroid, and breast cancers. Of the clinical variables examined, IPI risk showed the strongest relationship with the prioritized rbcDNA features. Pathway analysis linked this relationship to inflammatory and cytokine responses, immune regulation, and lymphoid processes, while genomic functional annotation suggested potential involvement of BCL6-related transcriptional regulation and heat-shock or cellular-stress responses. MN-positive erythrocyte fractions were also markedly higher in patients with DLBCL than in healthy controls (median, 9.36% [interquartile range, 6.43%–18.52%] vs 0.58% [0.27%–1.67%]; P<0.0001), providing parallel cellular evidence of chromosomal instability during erythropoiesis.
To our knowledge, this is the first genome-wide analysis of rbcDNA in a hematologic malignancy. DLBCL-associated rbcDNA features differed from those identified in healthy controls and solid tumors, with the strongest clinical relationship observed for IPI risk. Their distinction from solid-tumor features may relate to the hematopoietic origin of DLBCL, although the underlying mechanisms remain to be defined. These findings support further evaluation of rbcDNA profiling for DLBCL detection and clinical risk assessment. Larger, independent prospective cohorts incorporating disease controls and serial sampling are needed to validate these findings and assess their longitudinal relevance. Beyond DLBCL, rbcDNA profiling may offer a new approach to investigating how hematologic malignancies influence erythropoiesis and whether the resulting genomic instability can inform clinical assessment.
DISCLOSURES
Yu Xu: No relevant conflicts of interest to declare.