Many of us can recall back in primary school being asked to create a family tree, placing grand- or great grand- parents at the top and drawing branches to connect your aunts, uncles, cousins, and parents, all the way down to yourself. Not unlike this familial example, your immune cells are also related to one another, with the majority of your white blood cells originating from one multipotent progenitor cell type, hematopoietic stem cells (HSCs). Cells capable of forming long-term HSCs can be found very early during development, differentiating from hemogenic endothelium (HE) as rare precursors. Despite knowing they are there, actually distinguishing these cells from the multitude of cell types in the same region with similar protein expression profiles is quite difficult. Defining characteristic features of these cells would allow for high resolution study of HSC and HSC-precursor cells including investigating their unique environmental signals that promote their development from HE cells.
Supported by the Cancer Consortium Support Grant, at first working under the guidance of Dr. Irwin Bernstein and then independently in his own lab, Dr. Brandon Hadland wanted to investigate whether “advances in single cell transcriptomics paired with computational approaches would enable us to predict the signaling interactions regulating HSC emergence”. In their recent study published in Nature Communications, they showed that this task was no small feat, as HSCs are very rare within the aorta-gonad-mesonephros (AGM), and precursors are hard to define at early developmental stages (embryonic (E) days 9-11, as the focus of this study). The team had previously developed a unique ex vivo system that recapitulated the developing-HSC environment by co-culturing hemogenic precursors with endothelial cells derived from AGM (AGM-EC). But while most AGM cultures were able to support HSC development, they noticed that a few did not, setting the stage to ask what contributed to this observation. They used single cell (sc) RNAseq to look at broad differences in transcriptomes between these supportive and non-supportive AGM-EC and found potentially important genes that could provide some explanation, such as transcription factors and secreted signaling proteins.
To connect these transcriptomic analyses with functional characteristics, they also wanted to perform scRNAseq on hemogenic precursors as they developed into HSCs. Since these precursors were so rare, they first characterized surface markers that could reliably identify cells with bona fide capability of differentiating into HSCs. Building off previous work, and through examination of multiple different surface proteins, they identified three putative co-expressed markers - VE-Cadherin, CD61, and EPCR - that enabled enrichment of precursors by fluorescence activated cell sorting (FACS) at E9-E11. However, the individual cell potential for HSC development was quite varied within this population, further highlighting the need for better tools to specifically define precursors. When they analyzed the transcriptomic profiles of these cells using dimensionality reduction approaches, they found signatures of different cell types and differentiation states consistent with both endothelial and hematopoietic cells. Using a machine learning algorithm tool called Monocle, they were able to identify genes which varied between the endothelial and hematopoietic cells, suggesting the inferred temporal trajectory of transition between these cell types. Comparing their data with published gene sets for varying states of HSC development, they were able to successfully define a transcriptional state that appeared to identify the elusive HSC precursor cells.
Following this, a major question the authors wanted to ask using the large atlas of data they had collected was whether they could identify specific environmental signals that facilitated the development of HSCs. Dr. Hadland stated they were specifically interested in whether they could use this to “rationally design a stromal cell-independent engineered niche sufficient to generate engrafting HSC from embryonic hemogenic precursors in vitro” and noted that this would be a significant advance in the field, and applicable for translational use. To address this, they mapped a broad set of potential ligand-receptor interactions between these precursors and environmental cells and observed several notable standouts which fit with much of the previous literature on HSC formation and support. To generate their in vitro platform, they started first with serum free media and then, based on their receptor-ligand mapping, further decided to include recombinant stem cell factor (SCF), interleukin-3 (IL-3), thrombopoietin (TPO), and a ligand for VLA-4 and VLA-5. They also included anti-Notch1 and anti-Notch2, since they knew these signaling pathways were essential for HSC formation. When they tested whether this purposefully designed milieu was able to support HSC formation, they found that it could, but only for later timepoints of development (E11). However, based on their scRNAseq data, they were aware that HSC precursors existed at earlier timepoints, and so they looked for ways to improve their system to align with this. Intriguingly, they noticed that the chemokine, CXCL12, in the endothelial cells and its receptor, CXCR4, in precursors were seen at early emergence timepoints (E9-10). This was particularly interesting, as this pathway is known to have a “well-established role in HSC retention in the adult bone marrow niche”, stated Dr. Hadland, but its role in early development had not been characterized. When they included CXCL12 from their culture, this allowed for the generation of multilineage long-term HSCs from as early as day E9.
Altogether, this study provided important advancements both in understanding of HSC formation, as well as elegant experimental and computational approaches to investigate cell development more broadly. The authors point out that their “single cell transcriptomic atlas also provides a resource for further discovery of the niche signals and downstream transcriptional programs regulating HSC development”, that could be useful for the scientific community. While they recognize that there will need to be future studies to further optimize their engineered niche, they are hopeful that their initial platform will have larger implications for the potential to generate HSCs de novo from widely available pluripotent stem cells, and that these could potentially be used for cellular therapies in the future.
This research was supported by the NIH and the Cancer Center Support Grants from the Fred Hutch.
Fred Hutch/University of Washington/Seattle Children's Cancer Consortium members Drs. Irv Bernstein, Cole Trapnell and Brandon Hadland contributed to this work.
Hadland B, Varnum-Finney B, Dozono S, Dignum T, Nourigat-McKay C, Heck AM, Ishida T, Jackson DL, Itkin T, Butler JM, Rafii S, Trapnell C, Bernstein ID. 2022. Engineering a niche supporting hematopoietic stem cell development using integrated single-cell transcriptomics. Nat Commun. 13(1):1584.