Somatic epimutations enable single-cell lineage tracing in native hematopoiesis across the murine and human lifespan

Abstract

Current approaches to lineage tracing of stem cell clones require genetic engineering or rely on sparse somatic DNA variants, which are difficult to capture at single-cell resolution. Here, we show that targeted single-cell measurements of DNA methylation at single-CpG resolution deliver joint information about cellular differentiation state and clonal identities. We develop EPI-clone, a droplet-based method for transgene-free lineage tracing, and apply it to study hematopoiesis, capturing hundreds of clonal trajectories across almost 100,000 single-cells. Using ground-truth genetic barcodes, we demonstrate that EPI-clone accurately identifies clonal lineages throughout hematopoietic differentiation. Applied to unperturbed hematopoiesis, we describe an overall decline of clonal complexity during murine ageing and the expansion of rare low-output stem cell clones. In aged human donors, we identified expanded hematopoietic clones with and without genetic lesions, and various degrees of clonal complexity. Taken together, EPI-clone enables accurate and transgene-free single-cell lineage tracing at scale.

Figure 1:. DNA methylation jointly encodes differentiation and clonal identity.

A. Scheme of the experimental design. B. Overview of the 453 CpGs covered by our based on differential methylation in bulk data and performance in an undigested control experiment. WSH=within-sample heterogeneity, DMC=differentially methylated cytosine, IMR=intermediately methylated region C. uMAP of DNA methylation data from HSPCs from the LARRY main experiment (four mice). The color highlights clone, as defined from the LARRY barcode. D. Same uMAP as in C indicating cell states, see Figure 2 for how cell state was defined. E. Scatter plot depicting, for n=453 CpGs, the average methylation rate (x axis), the statistical association with surface protein expression (y axis, see Methods) and the statistical association with LARRY clonal labels (color coded, p value from a chi-square test). The CpGs in the upper/lower central rectangle were defined as static/dynamic CpGs, respectively. F, G. Bar chart depicting the fraction of static and dynamic CpGs within annotated enhancer/heterochromatic regions and within early/late replicating domains, respectively. H. Average methylation state across all profiled cells for the dynamic/static CpGs. CpGs with more than 50% methylated cells were termed methylated, the remaining ones unmethylated.

Figure 2:. Single-cell DNA methylation profiles deliver a high-resolution map of mouse hematopoiesis.

A. uMAP of DNA methylation data for HSPCs from the LARRY main experiment, the LARRY replicate experiment and from native hematopoiesis. Batch correction was applied prior to uMAP. Colors highlight groups identified from unsupervised clustering. B. uMAP of transcriptomic data from the same experiments. C. uMAP as in A, highlighting relative methylation state of cells across all CpGs that are methylated in HSCs or MPP3/4 in bulk data. D. uMAP as in A, highlighting relative methylation states of CpGs located in vicinity of TFBS. E. Surface protein expression of Sca1, c-Kit, CD135, CD201, CD48, and CD150. The CD135/CD201 and CD48/CD150 plots only show Lin⁻Sca-1⁺c-KIT⁺ cells and the color codes corresponds to the cell type clusters defined in the DNAm uMAP F. Enrichment of CpGs specifically unmethylated in a cell-type cluster according to the vicinity to the annotated TFBS.

Figure 3:. EPI-clone reliably identifies clones only from DNA methylation data.

A. Schematic overview of EPI-clone. B. uMAP of DNA methylation computed on static CpGs only for the LARRY main experiment, highlighting clonal identity as defined by LARRY barcodes. Only cells carrying a LARRY barcode are shown. C, D. Same uMAP as in B highlighting the cells by cell state as defined in Figure 2 (C) and those that were selected as part of expanded clones based on local density in PCA space (D). E. Receiver-Operating Characteristics Curve visualizing the performance of classifying cells into expanded and non-expanded clones based on local density in PCA space. F. Heatmap depicting the association between LARRY barcode and methylation-based clonal cluster identified by EPI-Clone. The row labeled with an asterisk contains all LARRY clones smaller than 30 cells. G. Sorting scheme for the mature myeloid experiment. H, I. UMAP representation of the mature myeloid experiment. Cells cluster both by clonal identity (LARRY barcode, H) and cell state (I). J. Adjusted rand indices between the ground truth clonal label (LARRY) and the clones identified by EPI-Clone stratified by cell type

Figure 4:. Blood production in age is characterized by a decrease in clonal complexity.

A. UMAP based on the static CpGs for a native, young mouse. B. Pie charts depicting size of expanded clones identified in this study and in a study where cells were labeled using a DNA barcode in the adult. In the EPI-clone pie chart, grey refers to cells that putatively come from small clones (<75 cells); in the lower pie chart, grey refers to clonal labels that are represented in less than 1% of cells (Granulocytes). C. UMAP based on the static CpGs and the associated EPI-clone clusters for the old mouse. D. Comparison of clone sizes, measured using the percentage of cells in the clone in comparison to all cells, for the old/young mouse. Clones with less than 75 cells are shown in grey. E. Comparison of clone sizes for the 20 largest clones. F. Comparison of HSC/MPP1 output and myeloid output for the 20 clones with the highest HSC/myeloid output between the young and old mouse. G. Bubble plot visualizing the frequency (measured as the square of the HSC/MPP1 frequency) of HSC/MPP1 cells per clone for the old versus the young mouse. H. Visualizing clones in the differentiation uMAP for the old (left) and young (right) mouse. Two examples of clones are shown for both mice. I, J. Characterizing clonal behavior of all clones (rows) across the different cell types (columns) as a heatmap. The color indicates the relative frequency of the clones toward four main differentiation trajectories. Immature=HSC/MPP1, MPP2; myeloid=MPP4, myeloid progenitors; lymphoid=B cell progenitors; erythroid=MEP, erythroid progenitor. This is shown for the old (I) and young (J) mouse separately.

Figure 5:. Different degrees of oligoclonality in human ageing.

A. Experimental design. B. uMAP of DNA methylation data for donor 1. See Figure S8 for donor 2. C. Expression of surface proteins (red) and presence of chromosome Y amplicons (blue). D. EPI-Clone uMAP of static CpGs in all cells except B and T cells, see also Figure 3. From left to right: Donor 1, highlighting clone; Donor 1, highlighting loss of chromosome Y; Donor 2, highlighting clone. P-value is from a Fisher test for the enrichment of LoY in the expanded cluster. E. Fraction of total cells stemming from the different expanded clones.