A GCSFR/CSF3R zebrafish mutant models the persistent basal neutrophil deficiency of severe congenital neutropenia

Abstract

Granulocyte colony-stimulating factor (GCSF) and its receptor (GCSFR), also known as CSF3 and CSF3R, are required to maintain normal neutrophil numbers during basal and emergency granulopoiesis in humans, mice and zebrafish. Previous studies identified two zebrafish CSF3 ligands and a single CSF3 receptor. Transient antisense morpholino oligonucleotide knockdown of both these ligands and receptor reduces neutrophil numbers in zebrafish embryos, a technique widely used to evaluate neutrophil contributions to models of infection, inflammation and regeneration. We created an allelic series of zebrafish csf3r mutants by CRISPR/Cas9 mutagenesis targeting csf3r exon 2. Biallelic csf3r mutant embryos are viable and have normal early survival, despite a substantial reduction of their neutrophil population size, and normal macrophage abundance. Heterozygotes have a haploinsufficiency phenotype with an intermediate reduction in neutrophil numbers. csf3r mutants are viable as adults, with a 50% reduction in tissue neutrophil density and a substantial reduction in the number of myeloid cells in the kidney marrow. These csf3r mutants are a new animal model of human CSF3R-dependent congenital neutropenia. Furthermore, they will be valuable for studying the impact of neutrophil loss in the context of other zebrafish disease models by providing a genetically stable, persistent, reproducible neutrophil deficiency state throughout life.

Figure 1. CRISPR/Cas9-induced mutant zebrafish csf3r alleles.

(a) Intron/exon structure of zebrafish csf3r locus. (b) Domain structure of zebrafish Csf3r protein. Ig = immunoglobulin, FBN = fibronectin. (c) Four CRISPR/Cas9-induced csf3r nonsense mutations identified in adult F1 DNA (designated alleles 1–4 for this report) aligned to WT sequence. The corresponding predicted truncated amino acid sequences are shown: blue = native Csf3r sequence, red = predicted non-native sequence downstream of the mutation site, *premature stop.

Figure 2. Neutrophil deficiency in csf3r null F2 embryos.

(a–d) Fluorescence micrographs of 3 dpf Tg(mpx:EGFP) embryos showing that several representative csf3r mutant null allelotypes (b–d) have substantially reduced numbers of fluorescent neutrophils compared to WT (a). Panel b is a representative homozygous csf3r^3/3 null embryo, and panels c-d show representative compound heterozygous null embryos of the two allelotypes shown. (e) Quantification over 2–5 days post-fertilisation (dpf) of tail region neutrophil numbers in embryos of control (WT) matings and incrosses of 3 different pairs of csf3r null parents of the csf3r allelotypes shown. Data are mean ± SD. Within-day comparisons across genotypes are analysed by unpaired two-tailed t-tests, pooling data from the mutant compound heterozygous allelotypes shown, which are not significantly different to each other. These parental allelotype combinations were randomly selected, being those pairs that laid of those that were set up. They are not intended to be comprehensive, but they do demonstrate a consistent non-complementing neutrophil-depletion phenotype encompassing five embryonic allelotypes (1/2, 1/3, 2/2, 2/3 and 3/3). (f) Embryos carrying single csf3r null alleles have an intermediate neutrophil deficiency. A mix of heterozygous csf3r^wt/(1or2) embryos in a 1:1 ratio were generated by outcrossing a parent of genotype csf3r^1/2 to WT. Their neutrophil numbers at 3 dpf are compared with the non-contemporaneous csf3r^WT/WT and pooled csf3r^−/− mutant 3 dpf groups of panel (e). The mutant data were pooled as none of the mutant allelotype groups is significantly different to any other. p-values are from one-way ANOVA with Tukey’s multiple comparisons test. p < 0.0001. (g) csf3r null embryos have equivalent survival to WT embryos up to 5 days post-fertilisation (dpf). Kaplan-Meier plots compared by Wilcoxon rank sum test.

Figure 3. Normal macrophage numbers in csf3r null F2 embryos.

(a) Photomicrographs of representative 3 dpf WT and csf3r null embryos stained by whole mount in situ hybridisation (WISH) for the macrophage-specific marker csf1ra/cfms. (b) Quantification of torso-located macrophage numbers in csf1ra/cfms WISH embryos in shows no difference between genotypes (the torso being the region distal to the dotted red line in (a), selected for scoring for its lower density of macrophages and lower incidence of overlapping cells).

Figure 4. Neutrophil deficiency in csf3r null adults.

(a,b) Fluorescent photomicrographs of 5 month old adult Tg(mpx:EGFP) zebrafish tail fins showing reduced numbers of EGFP-positive neutrophils in csf3r null (b) vs WT (a) animals. (c) Quantification of tail fin neutrophil density. n = 4–5 animals/genotype as shown.

Figure 5. Reduced granulopoiesis in csf3r null kidney marrow.

(a) Representative examples showing gating strategy on single and viable cells, and forward/side scatter profiles (FSC ad SSC) displaying reduced number of cells in the myeloid cell gate of csf3r-mutant kidney marrow cells compared to wildtype (WT). Supplementary Fig. S3 provides this information for all samples contributing to data in this figure. (b) May Grünwald-Giemsa stained cytospin of WT myeloid gate cells. (c) Percentage of viable cells falling within myeloid gate, which is significantly reduced in csf3r^−/− mutant kidney marrows. Points represent different animals. Data are mean ± SD; p-value from Mann-Whitney test. n = 8 animals (WT) and 6 (mutant). (d) Number of viable myeloid cells/kidney, which is significantly reduced in csf3r^−/− mutant kidney marrows. Data are mean ± SD; p-value from unpaired, 2-tailed t-test. n = 3 animals/group. (e) Four-category neutrophil differential counts based on the categories illustrated (top row of panels) in WT (n = 4) and csf3r^−/− mutant (n = 3) kidney marrow cells, demonstrating no marked difference.

Figure 6. Relative preponderance of eosinophils in csf3r null kidney marrow.

(a,b) Periodic Acid-Schiff (PAS) stained cytospins of FACS-purified myeloid cells from WT (a) and csf3r^3/3 mutant (b) kidney marrow. Red arrows indicate eosinophils, recognised by their PAS-positive strongly pink-staining cytoplasm and nuclear morphology and position. (c) Higher Neutrophil/Eosinophil ratio in WT vs csf3r^3/3 mutant myeloid cells. Individual ratios determined from >150 cell differential counts. (d) Although neutrophil numbers are depleted in csf3r^−/− kidney marrow, eosinophil numbers are not, indicating that the change in neutrophil:eosinophil ratio is due to relative change in the cell population sizes.