Hematopoietic differentiation is a stochastic process that delivers multiple distinct cell types at deterministically fixed frequencies. These opposites are reconciled through a topological representation of differentiation: even if the fate decision of each individual stem cell is stochastic, the shape of the landscape will determine the frequencies of the differentiated populations. However, how is the differentiation topology encoded? DNA methylation (DNAme) may be a key contributor to shaping the differentiation topology. DNAme is the most stably inherited epigenetic mark and therefore a strong candidate for encoding topological information. Moreover, clonal hematopoiesis in humans, a state that constitutes a significant topological disruption, often involves somatic mutations in modifiers of DNAme, including Dnmt3a, Tet2 and Idh2. To study how DNA methylation reshapes the differentiation topology we preformed single-cell RNAseq of >50,000 bone marrow progenitors from Tet2, Idh2, Dnmt3a mutated and wildtype mice. Tet2 knockout showed a decrease in erythroid-committed progenitors, and increase in monocyte-committed progenitors. Notably, this erythroid vs. monocyte fate-decision skew is caused by disruption in priming of early, uncommitted HSCs. Moreover, Dnmt3a deletion, which causes the opposite effect on methylation (hypomethylation), also results in opposite topological skews. This raises the question of what is the mechanistic link between stochastic genome-wide DNAme changes and deterministic topology skews (e.g. monocyte vs. erythroid). We hypothesize that genome-wide DNAme gain or loss may affect fate-decision through inherent biases in the TF motif CpG enrichment. Indeed, our analysis across lineage-defining transcription factor (TF) binding motifs uncovered that erythroid motifs show a marked enrichment in CpG content, compared with monocytic motifs. To further explore this hypothesis, we will examine the impact of DNAme of TF motifs on HSC priming, using two novel complementary approaches. First, ATAC-seq coupled with bisulfite sequencing will simultaneously study the sites of open chromatin critical to priming together with their methylation state. Second, to directly link DNAme and the transcriptional state of HSCs, we will apply joint single-cell bisulfite sequencing and RNAseq, to evaluate, at the single cell level, the interplay between TF binding motif DNAme and priming. To further define the impact of motif DNAme on TF binding, we will evaluate erythroid and monocytic TF binding in relation to DNAme using ChIP-bisulfite-seq. To functionally examine the impact of diffuse binding motif DNAme changes, we will apply epigenetic editing with guide RNAs targeting the binding motif itself. Finally, to examine this phenomenon in human HSC differentiation, we will apply our novel single-cell muti- omics platforms to jointly capture single-cell methylome, transcriptone and genotype. Thus, we will compare within the same individual with clonal hematopoiesis, the differentiation topology of mutant vs. wildtype HSCs, and define the role of DNAme in reshaping HSC differentiation in humans.
The formation of blood cells involves a complex process whereby stem cells morph into mature blood cells (e.g., red blood cells or white blood cells). However, what controls the choice of which mature blood cell will result from each stem cell remains poorly understood. We will apply novel single-cell sequencing technologies to create high- resolution maps of this process, and define the effect of specific changes to DNA (`methylation') in mouse models and human blood formation.
