Cxcl8 Expands Hematopoietic Stem and Progenitor Cells and Alters Their Interaction with the Endothelial Niche

We currently have little understanding of the mechanisms by which hematopoietic stem and progenitor cells (HSPCs) gain a selective advantage in patients with clonal hematopoiesis and other myeloid neoplasms. The chemokine CXCL8 is elevated in a subset of patients with myeloid neoplasms. Our previous work in zebrafish has discovered a novel role for cxcl8 and its receptor, cxcr1, in supporting colonization of HSPCs within the sinusoidal endothelial cell niche of the embryonic zebrafish known as the caudal hematopoietic tissue (CHT). We hypothesized that mosaic overexpression of cxcl8 in a population of HSPCs during development would alter HSPC-niche interactions, selectively favor HSPCs expressing cxcl8 and lead to their expansion in adults. To test this hypothesis, we microinjected DNA constructs encoding cxcl8-2A-GFP or GFP alone under the control of the HSPC-specific Runx1+23 enhancer into zebrafish embryos at the single-cell stage. Time lapse fluorescence video microscopy and single-cell tracking was performed on HSPCs within the CHT. Overexpression of cxcl8 nearly doubled the amount of time HSPCs resided within the CHT when compared to expression of GFP alone as a control (cxcl8: 4.94 ± 0.86 h vs GFP: 2.54 ± 0.18 h, p=0.01, N=142 tracked cells). Substitution of WT cxcl8 with a mutant cxcl8 construct lacking the ELRCXC motif required for receptor binding reduced these effects (WT cxcl8: 6.6 ± 0.48 h vs ELRCXC-cxcl8: 5.3 ± 0.33 h, p=0.02, N=355 tracked cells). To observe HSPC-niche interactions, kdrl:mCherry endothelial cell reporter zebrafish were microinjected with Runx1+23:cxcl8-2A-GFP or Runx1+23:GFP DNA constructs. The percent of time individual HSPCs spent closely interacting with a single group of CHT endothelial cells (endothelial cell cuddling) was quantified over the period from 52 to 72 hours post-fertilization. Overexpression of cxcl8 by HSPCs increased HSPC-endothelial cell cuddling time by 30% (cxcl8: 87% vs GFP: 57%, p=0.001). To directly test competition between wild type and cxcl8 overexpressing HSPCs, zebrafish embryos were microinjected with a 1:1 molar ratio of Runx1+23:cxcl8-2A-mCherry and Runx1+23:clover DNA. Single cxcl8-2A-mCherry⁺ and clover⁺ competitor cells were tracked by time-lapse fluorescence confocal microscopy. HSPCs expressing cxcl8 resided longer within the CHT than competitor HSPCs when quantified over the period from 72 to 96 hours post-fertilization (cxcl8: 4.0 ± 0.20 h vs competitor: 2.5 ± 0.25 h, p=2.0 x 10^-6, n=426 tracked cells). Single cell RNA-sequencing (scRNA-seq) of zebrafish embryos with mosaic expression of cxcl8 in HSPCs showed upregulation of cxcl12a in endothelial cells compared to endothelial cells from control embryos (p=5.19 x 10^-3), suggesting a possible mechanism to explain the increased CHT residency time. Zebrafish with mosaic expression of Runx1+23:cxcl8 were raised to adulthood and the kidney marrow cells were analyzed by flow cytometry. Compared to clutchmate controls, Runx1+23:cxcl8 mosaic transgenics had a higher hematopoietic progenitor/precursor to lymphocyte ratio, suggesting a mild differentiation block and possible lineage skewing (cxcl8: 2.0 ± 0.15 vs control: 1.6 ± 0.10, p=0.048, N=25 animals). Taken together, these data support a model in which pre-malignant HSPC clones aberrantly express cxcl8 and acquire a selective advantage over normal clones through enhanced interactions with the endothelial cell niche.