Single-Cell DNA Sequencing Reveals an Evolutionary Pattern of CHIP in Transplant Eligible Multiple Myeloma Patients

Abstract

Clonal hematopoiesis of indeterminate potential (CHIP) refers to the phenomenon where a hematopoietic stem cell acquires fitness-increasing mutation(s), resulting in its clonal expansion. CHIP is frequently observed in multiple myeloma (MM) patients, and it is associated with a worse outcome. High-throughput amplicon-based single-cell DNA sequencing was performed on circulating CD34+ cells collected from twelve MM patients before autologous stem cell transplantation (ASCT). Moreover, in four MM patients, longitudinal samples either before or post-ASCT were collected. Single-cell sequencing and data analysis were assessed using the MissionBio Tapestri^® platform, with a targeted panel of 20 leukemia-associated genes. We detected CHIP pathogenic mutations in 6/12 patients (50%) at the time of transplant. The most frequently mutated genes were TET2, EZH2, KIT, DNMT3A, and ASXL1. In two patients, we observed co-occurring mutations involving an epigenetic modifier (i.e., DNMT3A) and/or a gene involved in splicing machinery (i.e., SF3B1) and/or a tyrosine kinase receptor (i.e., KIT) in the same clone. Longitudinal analysis of paired samples revealed a positive selection of mutant high-fitness clones over time, regardless of their affinity with a major or minor sub-clone. Copy number analysis of the panel of all genes did not show any numerical alterations present in stem cell compartment. Moreover, we observed a tendency of CHIP-positive patients to achieve a suboptimal response to therapy compared to those without. A sub-clone dynamic of high-fitness mutations over time was confirmed.

Keywords: clonal hematopoiesis of indeterminate potential (CHIP); multiple myeloma; single-cell DNA sequencing; stem cells.

Figure 1

Single-nucleotide variants (SNVs) detected by single-cell DNA sequencing of 12 transplant-eligible patients with NDMM who underwent PBSC mobilization. (A) Venn diagram of the study design. In total we analyzed 19 samples collected at the following time points: BM CD34+ cells at diagnosis (n = 4), circulating CD34+ (n = 12) or CD34− (n = 1) cells selected from leukapheresis products before ASCT, BM-MNCs during post-ASCT maintenance therapy with lenalidomide (n = 1), and BM-CD138− cells at the time of relapse (n = 1). (B) Heatmap with the oncoprint of patient samples analyzed by single-cell DNA sequencing. Each column represents a patient, and each row represents a gene with at least 1 detected coding variant (235 different variants across 18 genes). The total number of variants detected per gene is shown in the left column with a green color gradient and the number of variants detected per patient is shown in the upper row with a red color gradient. The heatmap blue color gradient indicates the number of variants per patient and gene. CD34+ BM = CD34+ cells from bone marrow at diagnosis; CD34+ PB = CD34+ cells from apheresis product; CD34− PB = CD34− cells from apheresis product; BM-MNCs post-ASCT = bone marrow mononuclear cells post-autologous stem cell transplantation; CD138− BM relapse = CD138− cells from bone marrow at relapse.

Figure 2

Type of unique single-nucleotide variants detected in MM patients. (A) Coding SNVs organized by gene. Histograms show the number of variants identified per gene in all the 19 samples analyzed. (B) Pie chart with the distribution of the 109 unique variants according to the protein translation implication (missense, synonymous, frameshift, nonsense, 3′ utr, in frame, or splicing). (C) Histograms show the number of missense mutations identified per gene in all of the 19 samples analyzed. All variants reported are mutated in at least 0.5% of the cells from a given sample.

Figure 3

CHIP clones identified in CD34+ cells selected from leukapheresis products in 12 NDMM patients. (A) Table shows the number of mutations and clones identified in CD34+ cells collected following PBSC mobilization therapy in MM CHIP+ patients. (B) Histograms show the number of mutations in CHIP+ patients organized by gene. (C) Pie chart with the distribution of the clones organized by gene. (D) Pie charts with the distribution of clones for all the CHIP+ patients.

Figure 4

Clonal architecture of CHIP clones in MM patients at the time of transplant. Bar plots show the co-occurrence of mutations in CD34+ cells collected by leukapheresis in the six CHIP+ patients. Grey boxes represent the genes with heterozygous (HET) mutations whereas wild-type (WT) genotypes are shown with white boxes.

Figure 5

Genetic architecture of CHIP#1 patient. Bar plot showing co-occurrence of mutations in (A) CD34+ apheresis-derived cells and in CD138− at relapse phase and (B) CD34× apheresis-derived cells and in CD138× at relapse phase for CHIP#1 patient. Grey box represents the genes with heterozygous (HET) mutations whereas wild-type (WT) genotypes are shown with a white box. In panels (C,D), fish plot showing acquisition of mutations in serial time points. Bar plots show the distribution of cells in the sub-clones. At relapse phase, the major clone was significantly enriched. Minor clones include cells with a different heterozygous combination of mutations of DNMT3A, GATA2, and EZH2. Significant differences detected by Fisher’s exact test are shown in the tables below the histograms. T0 (D) = Diagnosis, T1 (A) = Apheresis, T2 (R) = Relapse, BM = Bone marrow, PB = Peripheral blood. ** < 0.001, *** < 0.001 and ns = non-significant.

Figure 6

Genetic architecture of CHIP#11, CHIP#12, and CHIP#14 patients. In panel (A–C), a bar plot showcases the co-occurrence of mutations at different time points analyzed, including CD34+ BM cells at disease onset (CHIP#11, CHIP#12, and CHIP#14, respectively), CD34+ apheresis-derived cells, and BM cells post-transplant (CHIP#11) (left panel). Grey box represents the genes with heterozygous (HET) mutations whereas wild-type (WT) genotypes are shown with a white box. Fish plot showing acquisition of mutations in serial time points (right panel). Bar plots show the distribution of cells in the sub-clones. Significant differences detected by Fisher’s exact test are shown in the tables below the histograms. T0 (D) = Diagnosis, T1 (A) = Apheresis, T2 (P) = BM post-ASCT, BM = Bone marrow, PB = Peripheral blood. * < 0.05, ** < 0.001, *** < 0.001 and ns = non-significant.

Figure 7

(A) Copy number alteration (CNA) analysis in patient CHIP#9. Variant call matrix per cell and copy number profile per clone are shown. Normalized counts of sequencing reads using the amplicons in the panel were used to calculate CNAs for each amplicon locus tested. Using the wild type (WT), heterozygous (HET), and homozygous (HOM) genotypes called by Tapestri Pipeline software, clones with no mutations (WT) were assumed to also present a normal diploid copy number and were thus used as a reference to compute potential CNAs present in mutated clones. (B) Clinical outcomes of MM patients. Diagram of clinical outcome of both CHIP+ and CHIP− MM patients. Response to therapy was evaluated according to the International Myeloma Working Group criteria (IMWG). CHIP+ patients are shown in red whereas CHIP− patients are shown in blue. sCR = Stringent Complete Response, CR = Complete Response, VGPR = Very Good Partial Response and PR = Partial Response. VTD = Bortezomib–Thalidomide–Dexamethasone, VRD = Bortezomib–Lenalidomide–Dexamethasone, DARA = Daratumumab, VCD = Bortezomib–Cyclophosphamide–Dexamethasone, LENA = Lenalidomide, IXA = Ixazomib, ASCT = Autologous Stem Cell Transplantation, R = Relapse.