Clonal Hematopoiesis in Diffuse Large B-Cell Lymphoma: Clinical and Molecular Correlations

Introduction: Clonal Hematopoiesis (CH) is an age-related phenomenon in which somatic mutations accumulate in hematopoietic stem cells (HSCs). Despite CH occurring in >30% of patients (pts) with diffuse large B-cell lymphoma (DLBCL), its impact remains controversial. CH may promote DLBCL in two ways: it can seed mutations in the B-cell progenitors of germinal center (GC) B-cells, contributing to malignant transformation. Alternatively, the clonal and pro-inflammatory tumor microenvironment derived from CH may be relevant in lymphoma promotion.

In the current study, we aim to: i) determine the prevalence of CH in a cohort of newly diagnosed DLBCL pts; ii) evaluate the clinical impact of CH on the risk of hematological toxicity of R-CHOP and on progression-free survival (PFS); iii) establish the correlation between CH and DLBCL genetic lesions; iv) evaluate at a single-cell level the frequency of co-occurrence of DLBCL driver mutations and CH mutations; v) evaluate if CH is enriched in cells of the lymphoma microenvironment compared to blood.

Methods: Pts with DLBCL, who participated in either the IOSI-EMA003 study or the SAKK 38/19 trial (NCT04604067) underwent Cancer Personalized Profiling by Deep Sequencing (CAPP-seq). For the detection of DLBCL mutations, somatic copy number abnormalities, and BCL6 translocations, a validated customized lymphoid panel (LyV4.0) was used and applied to plasma cell-free DNA (cfDNA) and paired granulocytes DNA (to filter out polymorphisms). NMF clustering was used for DLBCL subtyping. To detect CH mutations, a validated customized myeloid panel (EOncoLAB v1.0) was used and applied to genomic DNA (gDNA) from granulocytes. Moreover, multiomic single-cell DNA sequencing and protein expression were explored in six pts with paired peripheral blood or bone marrow (BM) and disaggregated lymph nodes (LN). A customized Tapestri (MissionBio) panel was designed to target CH mutations of interest and to barcode the dominant DLBCL clone by covering trunk oncogene mutations, enabling simultaneous genotyping and phenotyping of different cell types (B cells, various T cell subtypes, and myeloid cells).

Results: A total of 307 newly diagnosed DLBCL pts were analyzed by CAPP-seq. Clinical data are already available for 174 pts from the IOSI-EMA003 study. Overall, 44.3% of pts exhibited CH with a VAF >1%, surpassing the prevalence in the general population but aligning with findings in solid cancer series. DNMT3A was the most affected gene (22%), followed by TET2 (13%), with all other genes mutated in 3% or fewer pts. CH was associated with older age (p=0.006) and male sex (p=0.035). Univariate and multivariate analyses, adjusted for the IPI, showed no association between CH (HR=1.23, p=0.42), DNMT3A (HR=1.24, p=0.42), or TET2 (HR=0.56, p=0.12) mutations and PFS. Five-year PFS was 61.7% for wild-type (wt) CH and 50.6% for mutated CH (p=0.1), 59.6% for wt DNMT3A and 44.9% for mutated DNMT3A (p=0.07), and 54.7% for wt TET2 and 66.4% for mutated TET2 (p=0.28).

Among CH mutations, TET2 mutations were significantly enriched in EZH2-mutated DLBCLs (p=0.01).

Tapestri multiomic sequencing was performed on paired BM and LN from two DLBCL pts. One case (pt_1) harbored mutations in TET2 and CD79b/EZH2, while the other (pt_2), used as a control, had an ETV6 mutation but no CH mutations. CD79b/EZH2-mutated cells in pt_1 clustered entirely within the CD19+/CD10+ lymphoma population according to protein phenotyping, while TET2-mutated cells were ubiquitous across various cell compartments. These findings were consistent in both BM and LN samples. Moreover, CD79b/EZH2 mutations were mutually exclusive with respect to TET2 mutations, indicating that they were not seeded in the B cell that subsequently transformed to DLBCL. When considering solely cells of the microenvironment, TET2-mutated clones were enriched in the LN (6.4%) compared to the BM (2.6%) (p<0.001). Tapestri multiomic analyses on four additional pts are ongoing and will be presented during the congress.

Conclusion: CH mutations, particularly DNMT3A and TET2 mutations, do not impact the outcome of pts with DLBCL. TET2 mutations are specifically associated with EZH2-mutated DLBCLs. Preliminary single-cell results demonstrate the enrichment of TET2-mutated cells in the nodal microenvironment of EZB DLBCL.