Transcriptional and Spatial Heterogeneity of Mouse Megakaryocytes at Single-Cell Resolution

Megakaryocytes (MKs) have long been described solely as platelet progenitors. However, recent studies show that MKs also act as an essential component of the bone marrow niche to maintain hematopoietic stem cell function, and combat infection by engulfing and presenting antigens. However, it is not known whether these diverse programs are executed by a single cell population or distinct subsets of cells.

We performed single-cell RNA sequencing (scRNA-seq) to dissect the heterogeneity of MKs. To overcome the difficulty in obtaining the rare (0.1% in BM) and oversized (up to 65μm) highly polyploid, fragile MKs, we developed an efficient isolation strategy by combining fluorescence-activated cell sorting (FACS) sorting, manual selection of highly viable cells, and FISH verification of ploidy. We obtained 920 CD41⁺ highly-purified, bone-marrow derived, murine MKs spanning each ploidy stages (2N-32N) for scRNA-seq with modified Smart-seq2 protocol. On average, we detected over 6800 expressed genes and 250,000 transcripts in each MK. All cells expressed classical markers such as Pf4 and Itga2b (CD41).

Four cell clusters were identified using an unsupervised clustering method. Cells in Cluster 1 expressed mature MK markers CD42 and CD61, were enriched for hemostasis and platelet activation expression signatures, and consist of ≥8N cells, suggesting these MKs may represent platelet generating MKs. Cells in cluster 2 had lower CD42 and CD61 expression, were low ploidy (≤8N), and had higher expression of inflammation-related genes, including Ctss and Itgam ("inflammatory response-associated MKs"). Cells in Cluster 3 were enriched for DNA replication and DNA strand elongation (GO terms) and were in all ploidy stages ("MKs in polyploidization stage"). Cluster 4, most of which were high-ploidy, expressed high levels of CD42, CD61, TGF-β, and IGF1: factors regulating HSC behavior ("HSC niche cells"). Furthermore, we identified cell population-specific surface markers and transcription factors (TFs) for each of the 4 clusters. Then, immunostaining using antibodies against corresponding markers were performed to confirm the presence of respective MK subpopulations in the bone marrow. Our analyses suggest that defining MK stages by ploidy and traditional markers CD42 and CD61 alone may result in a genetically and developmentally heterogenous population of MK. Rather, MKs at various stages may be more specifically identified by these gene signatures.

MKs with different functions are known to have specific spatial distribution in the bone marrow (BM) niche. To test whether MKs with distinct expression signatures are uniquely localized within BM, we performed immunofluorescence staining on BM sections using antibodies against cell population-specific marker genes. Remarkably, we observed that Cluster 1 cells directly contacted blood vessels while most of Cluster 4 cells resided within one cell diameter of HSCs. The unique spatial distribution of cluster 1 and 4 population are in consistency with their respective transcriptomic signatures, and support that platelet generation and HSC maintenance are carried out by two distinct MK subpopulations. We further investigated the potential intrinsic relationship of these four Clusters during megakaryopoiesis. Developmental time courses were reconstructed using Monocle analysis, demonstrating that polyploidization (Cluster 3) occurs at the early stage of MK development with subsequent differentiation toward three orientations. While MKs with low ploidy appear to have two distinct cell fates (immuno-modulation or polyploidization), MKs with high ploidy (≥8N) differentiate towards populations associated with platelet production or stem cell regulation.

In summary, our study provides the first in vivo transcriptomic profile of megakaryopoiesis and a potential map of megakaryocyte heterogeneity at the single-cell resolution. MKs may be classified into different functional subpopulations irrespective to their developmental stage and degree of ploidy. These observations suggest that megakaryopoiesis does not occur merely in a stepwise process, but is dynamic and adaptive to locations in the BM and biological needs.