BackgroundMyeloproliferative neoplasm (MPN) is a heterogeneous group of clonal disorders. The underlying mechanisms of pathogenesis, especially in subtype specification and stochastic malignant transformation, are still largely unknown. Single-cell RNA sequencing (scRNA-seq) is a novel tool that can be used to identify the transcriptomic signature of individual cells. In the current study, we aimed to employ scRNA-seq to analyze genetic profiling of individual cells at different hematopoietic hierarchy in essential thrombocythemia (ET) patients.MethodsWe enrolled 7 ET patients and one healthy adult. Individual CD34+ progenitor cells were enriched from peripheral blood (PB). Harvested viable cells were barcoded and sequenced with the Illumina HiSeq 4000. Data was visualized with 10x Genomics Loupe software. We performed t-distributed stochastic neighbor embedding (t-SNE) plotting to dissect the scRNA-seq data and cluster cells with transcriptional similarity. Cellular sub-populations were stratified by the surface markers, including hematopoietic stem cells (HSC, CD34+CD38-Lin-), common myeloid progenitors (CMP, CD34+KIT+FLT3+IL3RAlowLin-), megakaryocyte-erythroid progenitor (MEP, CD34+CD38+PTPRC-) and granulocyte-macrophage progenitor (GMP, CD34+CD38+PTPRC+). Differentially expressed genes were subjected to gene ontology analysis. Gene Set Enrichment Analysis (GSEA) and the Reactome analysis were also employed. To clarify the distinct genetic background of these patients, targeted deep sequencing of PB granulocytes was also performed.ResultsAmong the 7 ET patients, two carry JAK2 mutation [one heterozygous (h-JAK2)and one homozygous (H-JAK2)] and three carry CALR mutation (1 type I, 2 type II). The remaining two cases are triple-negative (TN). Integrative analysis showed significant activation of JAK-STAT signaling in ET patients. Compared with control, the t-SNE analysis revealed disparate expression profiling in ET patients across various hematopoietic lineages. The discrepancy grew wider as the hematopoiesis became more lineage-restricted. Significant heterogeneity existed even among different ET patients, suggesting the high diversity of the disease. In the two JAK2-mutated patients, the t-SNE analysis demonstrated divergent transcriptomic profiling which scarcely overlapped. In the H-JAK2 sample, the HSCs exhibited a distinct profile different from the rest of hematopoietic progenitors. The CMPs were more closely related to MEP, which possibly suggested skewed differentiation and resultant ET phenotype. The phenomenon, however, was not similarly seen in the h-JAK2 sample. Using GSEA, we identified a subset of miR-21-targeted genes that were down-regulated in the h-JAK2 sample. Furthermore, there was apparent aberrant signaling activity of TGF-β, widely considered a regulator of miR-21, in this ET sample as compared with the H-JAK2 sample. Therefore, it was probably not a coincidence that, two months after blood sampling, the h-JAK2 patient suffered from disease transformation to secondary myelofibrosis. Among the three CALR-mutated patients, the expression patterns and the mutational profiles were also significantly discrepant. In the patient with type I CALR mutation, the CD34+ cells exhibited aberrant activity in epigenetic regulators. Coupled with the identified somatic mutations in some epigenetic modifiers from the targeted sequencing results, it is speculated that these mutations occur in a very early hematopoietic stage and contribute to ET pathogenesis through abnormal epigenetic regulation in this patient. Lastly, principal component analysis showed that the pathognomonic molecular events initiated at different hierarchical level of hematopoiesis in the 2 TN ET patients. Reactome analysis also disclosed one patient had altered DNA repair activity, and targeted sequencing confirmed the presence of TP53 mutation. Clinically, this patient exhibited highly aggressive, treatment-refractory disease.ConclusionWe demonstrate that scRNA-seq extends our knowledge of clonal diversity and inter-individual heterogeneity in patients with ET. Combined with the results from targeted sequencing, we were able to uncover unique transcriptomic pattern in samples carrying specific somatic mutations. The obtained information could potentially leapfrog our effort in the elucidation of the pathogenesis of ET. [Display omitted] DisclosuresNo relevant conflicts of interest to declare.