Lineage-specific distribution of clonal hematopoiesis of indeterminate potential across mature hematopoietic subpopulations in cancer patients

Abstract INTRODUCTIONThe mechanism by which CHIP arises and skews hematopoietic differentiation toward specific blood cell lineages remains poorly understood. Limited studies have shown preferential expansion of CHIP clones within myeloid lineages and low lymphoid representation, yet data in cancer populations remain scarce. We aimed to characterize the baseline lineage distribution of CHIP mutations across major peripheral blood (PB) subpopulations using targeted sequencing and droplet digital PCR (ddPCR). METHODSThis descriptive and observational study included patients (pts) with a first cancer diagnosis at age ≥ 60 years and treatment-naïve. PB samples were collected at diagnosis and analyzed using a customized 13-gene NGS panel (DNMT3A, TET2, ASXL1, JAK2, PPM1D, TP53, SF3B1, GNB1, SRSF2, CHEK2, CBL, GNAS, and NRAS) to identify CHIP carriers and their variants.These variants were subsequently tracked in sorted PB granulocytes (GR), monocytes (MO), T-, B-, and NK-lymphocytes isolated by cell shorting. DNA from each fraction was analyzed by ddPCR to quantify fractional abundance (FA) across hematopoietic lineages. Cases with an extracted DNA yield <0.05 µg and/or insufficient read counts for reliable ddPCR analysis were excluded.To characterize the contribution of each subpopulation to the CHIP clone, we compared the absolute FA within each subpopulation and its normalized allelic burden, defined as the ratio of FA in the subpopulation to that in whole blood. RESULTSThe prevalence of CHIP in our cohort of 113 pts is 36% with a median number of variants of 1 (IQR 1-2). With a total of 62 variants, mutations in DNMT3A, TET2, and PPM1D were the most prevalent, accounting for 75% of all variants (36%, 26%, and 13%, respectively), followed by ASXL1 and TP53 (8% each).Variant allele frequency by NGS and FA by ddPCR showed near-perfect correlation (Spearman 0.99), indicating highly comparable measurements. FA differed significantly among PB cell types (p < 0.001), with MO showing the highest clonal burden (median FA 5.8%), followed by GR and NK cells (median FA 3.7% and 2.7%, respectively). Markedly lower levels were observed in B and, particularly, in T cells (median FA 1.4 and 0.3%, respectively; p < 0.001). Paired comparisons confirmed a myeloid predominance (both GR and MO) over B- and T, but not over NK lymphocytes, which also displayed higher FA compared to T cells (p<0.001).Gene-specific analyses uncovered striking differences in lineage involvement. DNMT3A exhibited the strongest T-cell involvement across all genes, with a median normalized clonal burden of 0.28 (IQR 0.12–0.57) compared with its wildtype counterpart (0.03, IQR 0.01–0.14) (p < 0.001). TET2 mutations displayed a broad clonal expansion with higher normalized allelic load in GR (1.36 vs. 1.13, p = 0.04), MO (1.86 vs. 1.25, p = 0.01), B (0.87 vs. 0.47, p = 0.03), and NK cells (1.66 vs. 0.67, p = 0.002), but not among T cells. Conversely, ASXL1 mutations were myeloid-restricted, showing markedly lower normalized FA in B (0.01 vs. 0.65, p = 0.01) and T cells (0.00 vs. 0.11, p = 0.05). TP53 and PPM1D mutations showed no lineage-specific differences. These lineage-specific differences were not influenced by patient age.Analysis by type of neoplasm revealed that pts with B-lymphoproliferative disorders displayed a markedly lower mutational burden in B-cell compartment compared to other lineages (median FA 0, IQR 0-2 vs. 2, IQR 1-9; p=0.02), which correlated with a higher myeloid bias compared to other tumors (median 2.42 vs. 0.97, p=0.04). In contrast, pts with solid tumours including breast, ovarian, various types of gastrointestinal neoplasms, and head and neck cancer, showed no differences in the mutational burden of subpopulations. CONCLUSIONSIn our cohort of cancer patients, CHIP preferentially expanded within the myeloid lineage, with partial involvement of NK cells, and minimal representation in B and T lymphocytes. Gene-specific analyses revealed distinct lineage tropisms: TET2 mutations were broadly distributed across most hematopoietic compartments except T-cells, while DNMT3A displayed the highest relative representation in T lymphocytes, and ASXL1 was largely restricted to myeloid lineages.These findings underscore mutation-specific patterns of hematopoietic involvement, supporting the biological heterogeneity of CHIP and its potential role in modulating clinical phenotype and disease trajectory.