Development of RUNX1-Independent Hematopoiesis in Three Zebrafish runx1 -KO Models

The transcription factor RUNX1 is required for the emergence of definitive hematopoietic stem cells (HSCs) from the hemogenic endothelium during mouse embryonic development. Consequently, Runx1 knockout mouse embryos lack all definitive blood lineages and cannot survive past embryonic day 13. Previously we found that mutating the only runx1 homologue in zebrafish also resulted in deficiency of definitive blood cells during embryogenesis. However, the zebrafish runx1 mutants, which carried an ENU-induced nonsense mutation in runx1 (runx1^W84X/W84X), were able to recover from a larval "bloodless" phase and developed to adults with multi-lineage hematopoiesis, suggesting that definitive hematopoiesis can take place without RUNX1. These findings challenged the notion that RUNX1 is indispensable for HSC emergence and suggested that HSCs can form without an intact RUNX1 protein in the zebrafish. However, this observation was based on a single mutant line and required verification in additional independent models. Now we have generated three new runx1 mutantsusing the TALEN and CRISPR-Cas9 technologies. Two of the runx1 mutant lines carry mutations in exon 4 (a deletion of 8 bp, runx1^del8/del8, and a deletion of 25 bp, runx1^del25/del25) , both of which truncatethe runt-homologous domain. The third mutation is a large deletion of exons 3 through 8 (runx1^{del(e3-8)/del(e3-8)}) , which removes most of the coding region of runx1 . All three runx1 mutant linesrecapitulated our previous observations as the runx1 mutant embryos lacked the expression of the HSC marker c-myb and failed to initiate definitive hematopoiesis during early embryonic development. However, approximately 40% of the runx1^-/- embryos developed into fertile adults with circulating blood cells of multi-lineages, further supporting the presence of RUNX1-independent mechanisms for the generation of HSCs.

Live confocal imaging revealed the presence of hematopoietic progenitor cells in the runx1^{- /-} mutants at early stages of embryonic development. Transcriptional profiling of these hematopoietic precursors at both bulk and single cell levels by RNAseq showed that the runx1 -null hematopoietic progenitors are different from wildtype cells in global gene expression. On the other hand, the transcriptional profile of the hematopoietic cells in adult kidney is similar between runx1^-/- and wildtype adult fish, except for some key myeloid and thrombocyte genes, which were downregulated in the runx1^{- /-} mutants.

Taken together, we can now provide strong evidence from multiple independent runx1 -KO zebrafish models that RUNX1-independent pathways for HSC formation and definitive hematopoiesis exist.