Zebrafish Kelch-like family member 4 is required for vasculogenesis and hematopoiesis

Abstract

Molecular mechanisms regulating vascular development and hematopoiesis are still incompletely understood. The KLHL (Kelch-like) family of proteins function as adapters to target proteins for ubiquitination. However, their role in vascular development has not been previously analyzed. Here we have characterized a novel regulator of vascular development, kelch-like family member 4 (klhl4) in zebrafish . We show that zebrafish klhl4 is expressed in early vascular endothelial and hematopoietic progenitors, while its expression is restricted to vascular endothelial cells during later developmental stages. To determine the functional role of klhl4 , we generated loss-of-function zebrafish mutants using CRISPR/Cas9 genome editing. klhl4 mutant embryos were viable, yet they exhibited delayed sprouting of intersegmental vessels (ISVs), which correlated with reduced expression of vascular endothelial and erythroid specific molecular markers. Time-lapse imaging showed that vascular endothelial and hematopoietic progenitor cells exhibit delayed migration towards the midline and undergo increased apoptosis and reduced proliferation in klhl4 mutants. Expression of npas4l and etv2/etsrp, two master regulators of endothelial and hematopoietic development, was reduced in klhl4 mutants, suggesting that some vascular defects could be caused by the reduction of npas4l and etv2 expression. However, npas4l or etv2 overexpression failed to rescue ISV sprouting defects in klhl4 mutants, suggesting that klhl4 may promote vasculogenesis by additional mechanisms. In summary, our findings demonstrate a novel role for zebrafish klhl4 in regulating vascular endothelial and hematopoietic development during embryogenesis. Because the Klhl4 protein sequence is highly conserved between different vertebrates, it is likely that it may play a similar role in other organisms.

INTRODUCTION

The earliest blood vessels during embryogenesis form by the process of vasculogenesis, i.e. differentiation of vascular endothelial progenitor cells (also known as angioblasts) from undifferentiated mesoderm (Risau and Flamme, 1995). While it is difficult to study vasculogenesis in mammalian embryos due to embryo inaccessibility, zebrafish has emerged as an advantageous model system for developmental studies. In zebrafish embryos, similar to many other vertebrates, the earliest vascular endothelial progenitors originate bilaterally from the lateral plate mesoderm during late gastrulation and early somitogenesis stages. Subsequently they migrate to the midline, where they coalesce into the major axial vessels, the dorsal aorta (DA) and the posterior cardinal vein (Childs et al., 2002; Eriksson and Lofberg, 2000; Fouquet et al., 1997; Jin et al., 2005). This is followed by sprouting angiogenesis, when the first intersegmental vessels (ISVs) emerge from the DA starting from 21-22 hours-post-fertilization (hpf) and migrate towards the dorsal myoseptum (Childs et al., 2002; Lawson and Weinstein, 2002). In zebrafish, similar to other vertebrates, vasculogenesis is also tightly linked to hematopoiesis, and vascular endothelial and hematopoietic cells are thought to share a common progenitor, the hemangioblast (Vogeli et al., 2006). While much progress has been made in recent years to elucidate genetic and molecular pathways involved in hematopoietic and vascular development, the mechanisms that regulate the development of embryonic vasculature and hematopoiesis are still incompletely understood.

An ETS transcription factor etv2 / etsrp and bHLH-PAS transcription factor npas4l / cloche have been demonstrated to function as master regulators of embryonic vasculogenesis and hematopoiesis in zebrafish (Reischauer et al., 2016; Stainier et al., 1995; Sumanas et al., 2008; Sumanas and Lin, 2006). npas4l is first expressed in the presumptive hemangioblast and vascular endothelial progenitors at the tailbud – 2-somite stages, and is required for both vascular and hematopoietic development, since npas4l/cloche loss-of-function mutants show the loss of both lineages (Reischauer et al., 2016; Stainier et al., 1995). etv2 functions downstream of cloche, and its expression is enriched in early vascular endothelial progenitors, where it directly induces expression of multiple endothelial specific genes (De Val et al., 2008; Reischauer et al., 2016; Sumanas and Lin, 2006). Inhibition of etv2 function results in loss of vascular endothelial differentiation, and reduced differentiation of hematopoietic lineages (Chestnut et al., 2020; Pham et al., 2007; Sumanas et al., 2008; Sumanas and Lin, 2006).

Both npas4l and etv2 exhibit highly dynamic expression during vasculogenesis (Reischauer et al., 2016; Sumanas and Lin, 2006). ETV2 expression is known to be regulated at both the transcriptional and translational levels (Moore et al., 2013). However, the upstream regulators for the expression of both factors are still not well understood, and the factors involved in their post-transcriptional regulation are largely unknown.

Protein ubiquitination is a conserved post-translational modification, which targets many cellular proteins for proteasome-mediated degradation. In addition, ubiquitination regulates many cellular processes through non-proteolytic mechanisms (Rape, 2018). Kelch-like (KLHL) protein family is one of the key mediators of protein ubiquitination (Dhanoa et al., 2013; Zhou et al., 2024). KLHL proteins interact with Cullin Ring E3 ligase, which mediates ubiquitination of many cellular proteins. KLHL proteins comprise a large family of 42 proteins, which share 4-6 repetitive Kelch motifs. Kelch motifs are involved in substrate recruitment, thus targeting them for ubiquitination (Dhanoa et al., 2013; Zhou et al., 2024). It is currently unknown if any KLHL proteins are involved in vascular development.

We have previously used transcriptomic analysis to identify genes that are misregulated in cloche mutant zebrafish embryos (Wong et al., 2009). klhl4 was found to be downregulated in this analysis. Its sequence is evolutionarily conserved, and human KLHL4 has been previously associated with X-linked cleft palate (Braybrook et al., 2001). Studies in mammalian cell culture have shown that its expression is increased by p53 or DNA damage and it inhibits cell proliferation by activating p21 (Choi et al., 2020). KLHL4 has been implicated in regulating ectodermal patterning and neurulation by interacting with small GTPases GIT-PIX-PAK and affecting ubiquitination of PAK1 (Asmar et al., 2023). Previous work has demonstrated expression of zebrafish klhl4 in vascular endothelial cells (Gomez et al., 2012). However, its role in vascular development or function has not been previously investigated in any model system.

Here we investigate the role of klhl4 in vascular development in zebrafish embryos. We show that klhl4 mutants display defective vasculogenesis and hematopoiesis and show increased apoptosis of endothelial and/or hematopoietic cells. These defects were also associated with inhibited sprouting of intersegmental vessels and reduced expression of master regulators of vasculogenesis and hematopoiesis npas4l and etv2. These results demonstrate a novel requirement for klhl4 in vasculogenesis and hematopoiesis.

MATERIALS AND METHODS

Zebrafish lines and crosses.

Generation of klhl4^ci50 line. CRISPR/Cas9 editing was used to create the klhl4^ci50 allele. Guide RNAs were designed using ChopChop (http://chopchop.cbu.uib.no) (Labun et al., 2019). Two single-guide RNAs (sequences: AAATGTTCGTTAGCAGTACT and GGTAAAGTTTATCCATCAGA) were synthesized by in vitro transcription as previously described (Shah et al., 2016). gRNAs were incubated with NLS-Cas9-NLS protein (New England Biolabs), and the mixture was injected into one-cell stage fli1:GFP transgenic embryos. The injected animals were outcrossed and PCR was performed on pooled embryos to identify carriers of klhl4 mutations. The primers used, klhl4-F and klhl4-R: TGGGACATTTAGGCTGTATGTTT and ACCGTTTTTCATGGCTATGTCT are outside of the promoter deletion region. A single carrier that produced embryos with 473 bp deletion was identified and used to establish klhl4^ci50 line. The genomic sequence flanking the 473 bp deletion is listed below: GGCTTCTATATAATTACACACAAACCATCT.......473 bp deletion........ GGGATGAGAAAGTTAAGTCAGGCAGTATTC.

Other lines used in the study were: wild-type AB, Tg (fli1a:GFP)^y1 (Lawson and Weinstein, 2002), TgBAC(etsrp:GFP) ^ci1 (Proulx et al., 2010), Tg(kdrl:mCherry)^ci5 (Proulx et al., 2010), etv2^y11 mutants (Pham et al., 2007).

For most experiments (unless noted otherwise) klhl4−/− mutant embryos were obtained by the incross of klhl4−/− parents in fli1:GFP background, and wild-type (wt) controls were obtained by the incross of wt fli1:GFP adults. In some experiments, klhl4+/− heterozygous controls were used, which were obtained by crossing klhl4−/− sibling parents in fli1:GFP background to wt fli1:GFP adults.

Whole-mount in situ hybridization & Hybridization Chain Reaction and image analysis.

Whole mount in situ hybridization (ISH) was performed as previously described (Jowett, 1999). DIG-labeled antisense RNA probes for klhl4 (Wong et al., 2009), npas4l (Metikala et al., 2022; Reischauer et al., 2016), etv2 (Sumanas et al., 2005), cdh5 (Sumanas et al., 2005), kdrl (Thompson et al., 1998), gata1 (Detrich et al., 1995), and hbae3 (Brownlie et al., 2003) were synthesized as previously described. Following ISH, embryos were whole mounted in 0.6% low melting point agarose and imaged under light microscopy at 10X or 20X with Nikon Eclipse compound microscope. A series of z-slices were acquired using NIS Elements software to produce extended focus projected images.

Whole mount fluorescent in situ hybridization was performed using the hybridization chain reaction (HCR) as previously described (Choi et al., 2018). Fluorescently labeled RNA probes for klhl4, etv2, scl/tal1 and gata1 were purchased from Molecular Instruments, Inc. Following HCR, embryos were whole mounted in 0.6% low melting point agarose and imaged at 10X using a Nikon Eclipse confocal microscope. A series of z-slices were acquired using NIS Elements software to produce maximum intensity projection images. Fluorescence intensity of mRNA expression was analyzed using integrated density in Fiji / ImageJ. Specifically, fluorescence intensity of bilateral cells expressing etv2 mRNA (Fig. 6) was measured by selecting ten fluorescent cells using the multipoint tool, and integrated density was recorded. Subsequently, ten points on the yolk of the embryo were selected and integrated density recorded to account for background fluorescence. To obtain overall fluorescence intensity, background integrated density was subtracted from the bilateral cell integrated density.

Figure 6. Reduced expression of endothelial and hemangioblast markers in klhl4 mutants.

(A,B) In situ hybridization analysis of npas4l expression in the trunk region of klhl4 mutant and wild-type (wt) control embryos at the 10-somite stage. Note that npas4l expression is present in two bilateral stripes of medial (arrowheads) and lateral (arrows) cells in wt embryos (A). klhl4 mutants (B) show reduced npas4l expression, and the medial domain of npas4l expression is largely not apparent in many embryos. Flat-mounted preparations of the trunk region are shown, anterior is to the left. (C) Quantification of the number of embryos, which show npas4l expression in both lateral and medial cells or lateral cells only in the trunk region. (D,E) In situ hybridization analysis of etv2 expression in the trunk region of klhl4 mutant and wt control embryos at the 15-somite stage. Note etv2 expression in the lateral (arrows) and midline (arrowheads) vascular progenitors in wt embryos (D). Etv2 expression is greatly reduced in the midline cells in klhl4 mutants, which largely correspond to arterial progenitors (E). Flat-mounted preparations of the trunk region are shown, anterior is to the left. (F) Quantification of the number of embryos that show reduced etv2 expression at the midline. (G-I) Hybridization chain reaction (HCR) analysis of etv2 mRNA expression in klhl4 mutant and wt control embryos at the 15-somite stage. Dorsal view of the trunk region, anterior is to the top. Note etv2 expression in the lateral (arrows) and midline (arrowheads) progenitors in wt embryos (G). klhl4 mutants show reduced intensity of staining in the lateral cells and absent midline staining (H). Boxed area was quantified for the staining intensity. (I) Quantification of etv2 mRNA fluorescence intensity in klhl4 mutant and wild-type control embryos. Cells located bilaterally and at the midline were quantified separately. (J-M) Immunofluorescent analysis of Etv2 protein in wt and klhl4 mutant embryos at the 15-somite stage. Note the reduction in fluorescence intensity in klhl4 mutants. Boxed area was quantified for the fluorescence intensity and the number of cells. (L,M) Quantification of Etv2 fluorescence intensity (L) and the number of Etv2-expressing cells (M) in wt and klhl4 mutant embryos. Fisher’s exact test was used for the statistical analysis in (C,F,M), while two-tailed t-test was used in (I,L). Graphs show mean±s.d.

Fluorescence intensity of medial cells expressing etv2 mRNA (Fig. 6) was measured by drawing a rectangular box (68x251 pixels) spanning across fluorescent cells at the midline in the dorsal region of the embryo, and integrated density was recorded. An equal sized rectangle was drawn on the yolk of the embryo to account for background fluorescence, and integrated density was recorded. Overall fluorescence intensity was obtained by subtracting background integrated density from medial cell integrated density.

Real-time quantitative PCR (RT-qPCR).

Pools of 15-20 embryos were frozen on dry ice at the required developmental stage. RNA was isolated using RNAqueous – 4PCR kit (Thermofisher). cDNA synthesis was performed using the SuperScript VILO cDNA Synthesis Kit (Thermofisher). RT-qPCR was carried out using SYBR Green Master Mix (Thermofisher) in Azure Cielo Real-time Thermocycler (Azure Biosystems). Fold change was calculated using relative standard curve methods and normalized to housekeeping EF1a expression. Primer sequences used: EF1a-F (TCACCCTGGGAGTGAAACAGC), EF1a-R (ACTTGCAGGCGATGTGAGCAG), etv2-F (GAGCTGTTGCACAAAGGTCA), etv2-R (CAGAGAGGGACGAGGTTCTG), klhl4-qPCR-F (ACTTGTCCTGAGTGCAGTTTCA), klhl4-qPCR-R (ACTTGTGAGAGCTGGAGGAGAC).

Quantification of Intersegmental Vessel Extension.

ISV extension was quantified at various early developmental timepoints in wild-type and klhl4−/− embryos. Embryos were imaged under confocal microscopy at 10X magnification using a Nikon Eclipse microscope, and z-stack images were captured using NIS Elements software. Using maximum intensity or 3D projection images, extension of each ISV was quantified in embryos with the help of Imaris software 10.0 (Andor Inc). Typically, 18 pairs of anterior ISVs were counted in each embryo where feasible. The area of intersegmental vessel quantification included the entire trunk and tail region until about the mid-tail venous plexus. ISVs were individually scored as: 1) fully extended (past horizontal myoseptum), 2) partially extended (at or below horizontal myoseptum), or 3) very short/absent (no or very short extension apparent). The percentage of vessels falling into each category for each embryo was recorded. A two tailed t-test was used to compare each category of vessels between mutant and wild-type embryos.

Quantification of cell number in ISVs.

Embryos from wild-type fli1:GFP and klhl4−/−; fli1:GFP adults were collected and incubated at approximately 28.5°C until they reached approximately 48-52 hpf at which point they were fixed in 4% paraformaldehyde (PFA) in 1x phosphate buffered saline (PBS) overnight. Embryos were then transferred to 1xPBS and stored at 4°C until they were mounted laterally in low melting point agarose and imaged in the mid-trunk region at 10x magnification using the Nikon AX confocal microscope system with tiling. Images were denoised using the NIS Elements denoising tool and the total number of cells in each embryo were quantified in 6 ISVs of the midtrunk region, beginning at the end of the yolk extension and working anteriorly, in only the set of ISVs closest to the objective. When the end of the yolk extension was not visible counting started on the most posterior ISV. LUTs were adjusted for each embryo to best confine the fluorescence to nuclei and make sure all nuclei were visible and any nuclei within the bounds of the ISV were counted to represent a single cell.

Time-lapse imaging.

Embryos were mounted in 0.6% low melting point agarose with 0.2 mg/ml tricaine anesthetic on a glass slide and imaged under confocal microscopy for 20 hours beginning at the 15 somite stage. Z-stacks were captured using NIS elements software. Maximum intensity projections were used to analyze vascular progenitor cell migration in klhl4−/− and klhl4+/− embryos.

Apoptosis assay.

Embryos were fixed at the desired time point in 4% paraformaldehyde (PFA, in 1x phosphate buffered saline, PBS) at room temperature for two hours. Following fixation, embryos were washed for 4x15 minutes in 1X PBS with 0.2% Tween-20 (PBST) at room temperature, then blocked in 10% lamb serum for two hours at room temperature. Detection of Caspase3 was carried out by incubating samples overnight at 4°C with purified rabbit anti-active Caspase3 antibody prepared in 1X PBST with 10% lamb serum (1:200, Becton Dickinson catalog #559565). Embryos were washed for 6x15 minutes with 1X PBST at room temperature. This was followed by overnight incubation at 4°C in goat anti-rabbit Alexa 594 prepared in 1X PBST with 10% lamb serum (1:1000, Thermofisher catalog# A-11037). Following antibody treatments, embryos were washed for 4x15 minutes in 1X PBST. Embryos were mounted laterally in 0.6% low melting point agarose for imaging at 10X using a Nikon Eclipse confocal microscope. Area of apoptosis was measured using integrated density tool in Fiji / Image J. In brief, a rectangular box (396x105 pixels) was drawn slightly above the yolk extension over the area of apoptotic cells in the embryo. Cells expressing Caspase3 were outlined and selected by adjusting the color threshold, and integrated density was recorded.

Cell proliferation assay.

Approximately 21-22 somite stage wild-type and klhl4−/− embryos in fli1:GFP background were fixed in 4% PFA at 4°C overnight. Fixed embryos were permeabilized by incubating with 50% tetrahydrofuran (Sigma Millipore, 1:1 dilution in water) overnight, followed by 3 hr incubation in PBST(1% Tritonx100). Samples were washed in PBS and blocked in blocking solution (PBS+5%BSA+2% lamb serum). Followed by overnight incubation with Mouse anti-phospho-histoneH3 antibody (ThermoFisher, Cat. No. MA5-15220; 1:200 dilution) and secondary anti-mouse Alexafluor647 antibody (ThermoFisher, Cat. No. A32787; 1:400 dilution). Embryos were washed, stained with Hoechst 33342 and imaged using confocal microscopy (Nicon AX). For quantification, fli1:GFP-positive cells in the trunk region (region of interest, ROI) that spans approximately 400 μm in length starting next to the tail were selected. GFP/ phospho-HistoneH3 double positive cells were manually counted using Nikon NIS software. Cluster of GFP cells at the tip of the caudal region were excluded from the quantification.

Fluorescence Activated Cell Sorting (FACS).

Between 7 and 20 klhl4−/−; fli1:GFP mutant embryos were obtained by the incross of mutant parents, while 20 wt fli1:GFP+ and 20 wt fli1:GFP– embryos were obtained by the incrossing of wt fli1:GFP parents, which were screened and separated based on GFP expression. At approximately 48 hpf the embryos were manually dechorionated and transferred to 1.5mL microcentrifuge tubes for dissociation into single cell suspension by cold protease (Potter and Potter, 2019) and any excess water was removed. All steps were performed on ice. 1mL of deyolking buffer (55mM NaCl, 1.8mM KCl, 1.25mM NaHCO₃ in 1XPBS) was added to each of the three samples and pipetted up and down until full dissolution of the yolk was achieved and the samples were centrifuged at 300G for 1 minute. The supernatant was aspirated, and the samples resuspended in 500μL 0.5x Danieau Buffer before centrifugation at 300G, 4°C for 1 minute and subsequent aspiration of supernatant. Resuspension in 0.5x Danieau Buffer, centrifugation, and supernatant aspiration were repeated once before 1mL of Bacillus lichenformis enzyme mix (10mg/mL B. lichenformensis, 125U/mL DNAse I, 0.5mM EDTA in BSA Fraction V (7.5%); prepared immediately prior to addition) was added to each sample and immediately triturated 15 times using a 1mL pipette with subsequent trituration 15 times approximately every 2 minutes for approximately 30-60 minutes, until dissociation was confirmed by visualization of a 1-2μL aliquot of each sample with minimal clumps of cells remaining. Samples were then centrifuged at 1200G, 4°C for 5 minutes. Supernatant was aspirated and pellets were resuspended in 1mL 7.5% BSA Fraction V before being passed through a 20 micron strainer into a clean 1.5mL tube and the strainer was rinsed with 500mL of 7.5% BSA Fraction V. Cells were pelleted at 1200G, 4°C for 5 mins, resuspended in 1mL 7.5% BSA Fraction V and pelleted again at 1200G, 4°C for 5 mins. Cells were then resuspended in 300μL of eBioscience^™ Flow Cytometry Staining Buffer and transferred to 5mL round bottom tubes for analysis of total fluorescent cell counts using the Cytek^® Aurora spectral flow cytometer. After quality control was performed, the wild type fli1:GFP+ and the GFP- cell samples were used as reference controls for the experiment, then subsequently the percentage of fluorescent cells in each of the three samples was acquired.

ETV2 immunostaining.

Embryos were fixed at the desired time point in 2% PFA diluted in 0.1% PBST overnight at 4°C. Following fixation, embryos were washed for 4x5 minutes with 0.1% PBST at room temperature followed by 1x30 minutes with 0.5% PBS-TritonX-100 at room temperature. Embryos were blocked for two hours at room temperature in blocking solution (0.1% TritonX-100, 10% normal goat serum, 1% BSA, 0.1% PBST) and then incubated overnight at 4°C in ETV2 rabbit polyclonal serum (1:500, Kerafast, Cat No 1654) in blocking solution. Embryos were washed for 6x40 minutes at room temperature in 0.1% PBST and incubated overnight with goat anti-rabbit Alexa 594 (1:1,000 ,Thermofisher, Cat No A-11037) in blocking solution. Immunostained embryos were mounted in 0.6% low melting point agarose and imaged using a Nikon Eclipse confocal microscope. Z-slices were acquired using NIS elements software, and maximum intensity projection images were used for analysis. The multipoint tool on ImageJ was used to select and quantify the number of bilateral cells expressing ETV2. Additionally, fluorescence intensity of bilateral cells expressing ETV2 was recorded by selecting ten fluorescent cells using the multipoint tool, and integrated density was recorded. Subsequently, ten points on the yolk of the embryo were selected and integrated density recorded to account for background fluorescence. To obtain overall fluorescence intensity, background integrated density was subtracted from bilateral cell integrated density.

npas4l mRNA overexpression.

mRNA encoding npas4l was synthesized in vitro using the mMESSAGE mMACHINE SP6 Transcription Kit (Invitrogen) from a pCS2-npas4l plasmid construct, obtained from Addgene Inc. (Reischauer et al., 2016). Approximately 50 pg of npas4l mRNA was microinjected into the yolk of zebrafish embryos at the one-cell stage. Embryos were cultured under standard conditions until the shield stage and subsequently fixed in 4% paraformaldehyde for ISH analysis.

Single-cell RNA-seq analysis of klhl4 expression.

The data from DanioCell (Sur et al., 2023) is freely available here (https://daniocell.nichd.nih.gov/). The Daniocell2023_SeuratV4.rds, cluster_annotations.csv, and the daniocell_load.R script were downloaded from the DanioCell website and used to annotate cell types for the clusters. Utilizing R packages: Seurat (Hao et al., 2021), ggplot2 (Wickham, 2016), dplyr (Wickham H, 2023), reshape2 (Wickham, 2007), and tidyverse (Wickham H, 2019) the expression of klhl4 and hemangioblast marker npas4l was investigated using custom written R code (Hao et al., 2024; R Core Team, 2025; Rstudio Team, 2020). The UMAPs and feature plots were generated with this data annotated by DanioCell as the hematopoietic cluster, then sub-setting was performed to select cells from early developmental time points <=14hpf.

RESULTS

It has been previously shown that zebrafish klhl4 expression is localized within the intermediate cell mass (ICM) and vasculature (Armant et al., 2013; Gomez et al., 2012). However, detailed analysis of its expression has not been reported. Therefore, we performed in situ hybridization (ISH) and hybridization chain reaction (HCR) analysis of klhl4 expression at the stages of 15-somite – 48 hpf. At the 15-somite stage, klhl4 was observed in two bilateral stripes within the anterior (ALPM) and the posterior lateral plate mesoderm (PLPM), that likely correspond to vascular endothelial progenitors (Fig. 1A). klhl4 expression at the 18-somite stage was apparent in both arterial and venous progenitors of the trunk vasculature, and also in the cranial vessels (Fig. 1B). Its expression at 26 hpf stage was localized to all vascular endothelial cells and the endocardium (Fig. 1C,D). klhl4 expression was greatly reduced in etv2^y11 mutants (Pham et al., 2007), which show inhibited differentiation of vascular endothelial cells (Fig. 1E,F).

Figure 1. In situ hybridization (ISH) analysis of klhl4 expression.

(A-C) Chromogenic ISH analysis of klhl4 expression at 15-somite through 26 hpf stages. Note klhl4 expression in presumptive vascular endothelial progenitors in the anterior (arrows, A) and posterior lateral plate mesoderm (arrowheads, A) at the 15-somite stage, and vascular endothelial cells at 18-somite and 26 hpf stages including both the forming dorsal aorta (DA, black arrowheads, B,C) and the posterior cardinal vein (PCV, white arrowheads, B,C).

(D) Ventral view of the head region of a flat-mounted deyolked embryo at 26 hpf. Note the expression in the endocardium (arrow).

(E,F) klhl4 expression is lost or greatly reduced in etv2−/− mutant embryos at 24 hpf. DA (black arrowheads) and PCV (white arrowheads) are indicated.

(G-I) Hybridization chain reaction (HCR) analysis for klhl4 (green) and etv2 (red) co-expression at 24 hpf in the trunk and tail region. Note that expression of both markers largely overlaps in vascular endothelial cells. DA (green arrowheads), PCV (white arrowheads) and caudal vein (white arrow) are indicated.

(J-L) HCR analysis for klhl4 (green), gata1 (red) and kdrl:GFP (magenta) co-expression at 24 hpf in the trunk and tail region. Note that expression of klhl4 and gata1 has no or very little overlap. Inset in the upper right corner (J,K) shows magnified view of the region outlined in a box with a dotted line. kdrl:GFP and klhl4 are co-expressed in most endothelial cells (green arrowheads show the DA) except for the caudal vein (white arrow), which is positive for klhl4 but negative for kdrl:GFP. White arrowhead indicates PCV.

(M-P) HCR analysis for klhl4 (magenta), etv2 (red) and scl/tal1 (green) expression at the 10-somite stage. The mid-trunk region is shown, anterior is to the left. Note that etv2 and klhl4 expression co-localize in vascular endothelial progenitor cells (white arrows). scl expression labels both vascular endothelial and hematopoietic progenitors. Note klhl4 and scl co-expression in putative hematopoietic progenitors, which are negative for etv2 expression (white arrowheads). A selected region (outlined with a dotted line) shows a klhl4+scl+ positive putative hematopoietic cell at a higher magnification in the right top corner.

We used HCR analysis to analyze co-localization of klhl4 with other vascular endothelial and hematopoietic markers. klhl4 overlapped extensively with etv2 expression at 24 hpf stage, which confirms its expression in vascular endothelial cells (Fig. 1G–I). In contrast, it did not overlap with erythroid marker gata1 expression, suggesting that klhl4 has little or no expression in hematopoietic cells at this stage (Fig. 1J,K). Interestingly, klhl4 and endothelial kdrl:GFP expression overlapped in the trunk except for the most posterior region of the caudal vein, which was positive for klhl4 but negative for kdrl:GFP (Fig. 1L). This region contains vascular endothelial progenitors that are derived from the endoderm and are initially negative for kdrl:GFP expression (Nakajima et al., 2023), suggesting that klhl4 expression is present in early vascular progenitors and may precede kdrl expression.

To determine if klhl4 is expressed in angioblasts or hematopoietic progenitors prior to the 15-somite stage, we used HCR analysis to probe expression of klhl4, endothelial progenitor marker etv2 and endothelial and hematopoietic progenitor marker scl/tal1 (Liao et al., 1998). No klhl4 expression was detected at the 5-somite stage (data not shown). At the 10-somite stage, klhl4 expression was apparent in the lateral plate mesoderm of the embryonic trunk (Fig. 1M). Its expression co-localized with etv2 expresion in vascular endothelial progenitors (Fig. 1N). In addition, some klhl4-positive cells were negative for etv2 expression, and positive for scl expression (Fig. 1O,P). This suggested that early klhl4 expression is present in both hematopoietic and vascular endothelial progenitors, while it is absent from hematopoietic cells at later stages. To confirm this, we reanalyzed the previously published single-cell RNA-seq (scRNA-seq) dataset at the 10-somite (14 hpf) stage (Sur et al., 2023). As expected, klhl4 was expressed in vascular endothelial cells (Suppl. Fig. 1A,B). In addition, it showed expression in the putative hemangioblast cells, the common precursor of vascular endothelial and hematopoietic cells, where it partially overlapped with npas4l expression (Suppl. Fig. 1C), a known marker for hemangioblast and early vascular endothelial progenitor cells (Reischauer et al., 2016). Thus, klhl4 appears to be present in the early hematopoietic and vascular endothelial progenitors, while its expression is restricted to vascular endothelial cells at later stages.

To study the functional role of klhl4, we generated a loss of function mutant allele using CRISPR / Cas9 mutagenesis. Two single-guide (sgRNAs) were designed to delete a portion of klhl4 proximal promoter (Fig. 2A). Subsequently, klhl4^ci50 mutant line that contains a 473 bp promoter deletion was obtained. klhl4 mutants showed strong reduction in klhl4 mRNA expression as confirmed by in situ hybridization and qPCR, with only approximately 7% of klhl4 transcript remaining (Fig. 2B–D). Nevertheless, klhl4−/− embryos did not show any apparent morphological defects (Fig. 2E,F) and were viable as adults.

Figure 2. Generation and characterization of klhl4^ci50 promoter deletion mutants.

(A) A schematic diagram showing the genomic structure of klhl4 gene and an approximate position of two sgRNAs used to delete 473bp of the proximal promoter region of klhl4. The approximate position of the two genotyping primers, klhl4-F and klhl-R, is shown. Black boxes and lines indicate exons and introns, respectively. (B,C) In situ hybridization analysis for klhl4 expression in klhl4 homozygous mutants (C) and wild-type (wt) controls (B) at 24 hpf. Note the strong reduction in vascular endothelial klhl4 expression (arrows) in the klhl4−/− embryo. (D) qPCR analysis for klhl4 expression in wild-type control and klhl4−/− embryos at 28 hpf. Error bars show ±SEM. (E,F) Bright-field imaging of wild-type control and klhl4−/− embryos at 48 hpf. No obvious morphological defects are apparent in klhl4 mutants.

To analyze for any potential defects in vascular development, klhl4 mutant embryos in vascular endothelial fli1:GFP background were imaged using confocal microscopy at 20-somite (19 hpf) to 48 hpf stages. klhl4 mutants showed reduced number of fully or partially extended ISVs, and increased number of ISVs that were very short or absent at 20-somite to 24 hpf stages, indicative of delayed ISV formation (Fig. 3A–I). At 48 hpf stage, approximately 13% ISVs were partially extended and 2% ISVs were very short or absent in klhl4−/− embryos, while all ISVs had fully extended in wild-type (wt) fli1:GFP control embryos (Fig. 3J–L). Despite these ISV extension defects, the average number of vascular endothelial cells in each ISV was not significantly affected in klhl4 mutant embryos (Fig. 3M–O). To test if the overall number of fli1:GFP cells was altered in klhl4 mutants, we performed fluorescence activated cell sorting (FACS) at 48 hpf stage. However, no significant difference in the percentage of GFP-positive cells was observed between klhl4 mutant and wt control embryos (Fig. 3P). In order to confirm that the observed defects in ISV sprouting were not due to variations in the genetic background between mutant and control embryos, we obtained sibling embryos from the cross of heterozygous klhl4+/− ; kdrl:mCherry parents, imaged using confocal microscopy at 24 hpf, and subsequently genotyped. In this analysis, klhl4 mutant embryos also showed increased incidence of reduced or delayed ISV formation, compared to wild-type sibling embryos (Suppl. Fig. 2).

Figure 3. klhl4 mutants show reduced intersegmental vessel (ISV) extension and increased apoptosis.

(A-I) Confocal imaging of the trunk region of wild-type (wt) fli1:GFP (A,D,G) and klhl4−/−; fli1:GFP embryos (B,E,H) at 20-somite (19 hpf, A-C), 22 hpf (D-F) and 24 hpf (G-I) stages. Note that sprouting of intersegmental vessels (ISV, arrows) is reduced in klhl4−/− mutants. Also note the appearance of round GFP+ cells (white arrowheads), located laterally on the surface of the yolk adjacent to vasculature, which are present in klhl4−/− embryos only. (C,F,I) Graphs show quantification of ISV extension in the trunk region of wt and klhl4−/− embryos. (J-L) ISV extension analysis in the trunk region of klhl4−/− and wt embryos in fli1:GFP background at 48 hpf. Note the shortened ISVs in klhl4 mutants (K, arrowheads). (M-O) Analysis of ISV cell number in the trunk region of klhl4−/− and wt embryos in fli1:GFP background at approximately 52 hpf. Endothelial cell nuclei (arrows) were counted in the selected trunk region. No significant difference in ISV cell number between klhl4−/− and wt embryos was observed (O). Arrowheads in (N) show truncated ISVs in klhl4 mutants. (P) FACS analysis of fli1:GFP cells in wt and klhl4 mutant embryos in fli1:GFP background at 48 hpf. Percentage of GFP-positive cells among all viable cells was analyzed. No significant difference was observed between wt and klhl4 mutant embryos. (Q-S) Immunostaining for Caspase 3 to assay apoptosis in klhl4−/− mutant and klhl4+/− control embryos at the 20-somite stage. Fluorescence area was quantified in (S). (T-V) Magnified region of klhl4−/− embryo displayed in (R), showing an overlap with fli1:GFP expression. Note that many round fli1:GFP cells next to the yolk (arrowheads) are positive for Caspase 3 expression. (W-Y) Immunostaining for cell proliferation marker phospho-histone H3 in wild-type fli1:GFP and klhl4−/− embryos at the 21-22-somite stage. The number of proliferating cells in the trunk region that co-express fli1:GFP was quantified in (Y). Two-tailed t-test was used for statistical analysis in all graphs; error bars show mean ± s.d.

In addition to ISV defects, klhl4 mutant embryos exhibited multiple round fli1:GFP fluorescent cells that were located bilaterally along the yolk extension in the trunk and tail region, and were not connected to the rest of the vasculature (Fig. 3B,E,H, arrowheads). We hypothesized that these cells were apoptotic or dying cells. We then performed anti-Caspase3 immunostaining, which detects apoptotic cells. An enlarged area of apoptosis in the ICM region was observed in klhl4 −/− embryos compared with klhl4 +/− controls, obtained by mating sibling klhl4−/−; fli1:GFP adults to wild-type fli1:GFP (Fig. 3Q–S). Multiple round bilateral fli1:GFP cells stained positive for caspase 3 (Fig. 3T–V), confirming that many fli1:GFP positive cells undergo apoptosis in klhl4 mutant embryos. In contrast, we did not see bilateral apoptotic cells in kdrl:mCherry embryos (Suppl. Fig. 2), suggesting that these apoptotic cells were negative for kdrl expression. While kdrl expression labels vascular endothelial cells, fli1:GFP expression is known to label vascular endothelial progenitors at an earlier stage than kdrl, and is also observed in hematopoietic cells (Lawson and Weinstein, 2002), suggesting that increased apoptosis may be observed in either endothelial or blood progenitors, or both.

To test if endothelial or hematopoietic cell proliferation was affected in klhl4 mutants, we performed immunostaining against phospho-histone H3 in klhl4 mutant and wild-type control embryos at the 21-22-somite stage. The number of fli1:GFP cells in the trunk region, positive for pH3 staining, was significantly reduced in klhl4 mutant embryos compared to wt controls (Fig. 3W–Y). These results argue that in addition to increased apoptosis, vascular endothelial and / or hematopoietic fli1:GFP cells also show reduced cell proliferation.

We then performed ISH staining to analyze for expression of molecular markers, associated with vascular or hematopoietic development. Expression of vascular endothelial markers cdh5 and kdrl was reduced at 24 hpf stage in klhl4 mutants (Fig. 4A–F). In addition, reduced or absent staining in ISVs was apparent (arrowheads, E). ISV sprouting was defective in klhl4 mutant embryos, similar to the sprouting defects observed using fluorescent reporter lines. Interestingly, expression of erythroid specific markers gata1 and hbae3 was also reduced in klhl4 mutants (Fig. 4G–L), arguing that klhl4 deficiency affects both vascular development and hematopoiesis.

Figure 4. In situ hybridization analysis of genetic markers of vascular endothelial and red blood cell differentiation.

(A-F) Expression of vascular endothelial markers cdh5 (A-C) and kdrl (D-F) in the trunk and tail region at 24 hpf in klhl4−/− mutants (B,E) and wild-type (wt) controls (A,D). Note reduced cdh5 expression (arrows, B) and absent intersegmental vessels (arrowheads, E) in klhl4 mutants. Percentage of embryos showing reduced marker expression is plotted in (C,F). p<0.0001, Fisher’s exact test. (G-L) Expression of red blood cell markers gata1 (G-I) and hbae3 (J-L) at 22 hpf in klhl4−/− mutants (H,K) and wild-type controls (G,J). Percentage of embryos showing reduced marker expression is plotted in (I,L). p<0.0001, Fisher’s exact test. The numbers in the lower right corner show the number of embryos exhibiting the described phenotype out of the total number of embryos analyzed.

It is possible that defects in ISV sprouting and hematopoiesis observed at 24 hpf are caused by early defects in endothelial and hematopoietic cell specification, migration or survival. Vascular endothelial and hematopoietic cells first emerge bilaterally along the trunk and tail region, and then migrate towards the midline where they coalesce into the axial vessels, the DA and the PCV, and the hematopoietic blood island / ICM (Childs et al., 2002; Detrich et al., 1995; Eriksson and Lofberg, 2000; Fouquet et al., 1997; Jin et al., 2005). To test if there were any defects in the migration of early hematopoietic and endothelial progenitors, we performed time-lapse imaging of klhl4−/− and control klhl4+/− embryos in fli1:GFP background starting at approximately 15-somite stage. klhl4 mutant embryos showed delayed migration of vascular endothelial and hematopoietic cells (Fig. 5, Suppl. Movies 1,2). Multiple bilaterally located fli1:GFP progenitors failed to migrate towards the midline and underwent apoptosis in klhl4 mutants (Fig. 5, yellow arrows).

Figure 5. Time-lapse imaging of endothelial and hematopoietic cell migration.

klhl4−/− and klhl4+/− embryos were imaged starting at approximately 15-somite stage. Dorsal view, trunk and tail regions are shown, anterior is up. Note that in klhl4+/− embryos many angioblasts are apparent at the midline (white arrows), while the most posterior angioblasts, as well as fli1:GFP-positive blood cells are still in the process of migrating (arrowheads). klhl4−/− mutants show greatly delayed migration of angioblasts (arrowheads). In addition, multiple round apoptotic cells appear bilaterally in klhl4−/− embryos (yellow arrows), while few such cells are present in klhl4+/− embryos.

npas4l /cloche is one of the earliest regulators of vascular endothelial and hematopoietic differentiation (Reischauer et al., 2016; Stainier et al., 1995). Its expression in the trunk region of wt embryos was observed in two distinct groups of endothelial progenitors at the 10-somite stage, medial and lateral cells, which presumably correspond to arterial and venous progenitors, respectively (Kohli et al., 2013; Perens and Yelon, 2024; Reischauer et al., 2016) (Fig. 6A). npas4l expression was greatly reduced or absent largely from the medial cells in klhl4 mutants, while its lateral expression was not significantly affected (Fig. 6A–C). Similarly, expression of etv2 / etsrp, a marker of vascular endothelial progenitors (Kohli et al., 2013; Sumanas and Lin, 2006), was greatly reduced or absent in the medial / arterial progenitors in klhl4 mutants at 15-somite stage, while its expression in the lateral / venous progenitors was only mildly affected (Fig. 6D–F). Because the conventional ISH is not quantitative, we performed fluorescent HCR analysis to quantify etv2 expression level in arterial and venous progenitors. etv2 expression was reduced in both medial and lateral cells, while the level of reduction was greater in the medial cells, supporting results obtained using ISH (Fig. 6G–I). Immunofluorescence for Etv2 protein also showed a similar reduction in both Etv2 protein expression level, and in the number of Etv2+ cells at the 15-somite stage (Fig. 6J–M). These results show that expression of npas4l and etv2 are both reduced in klhl4 mutants, and this reduction is greater in the arterial progenitors.

Previous studies have demonstrated that Etv2 plays an important role in the survival of vascular endothelial cells, and etv2 mutants show excessive apoptosis of endothelial cells, as well as reduced expression of vascular endothelial markers and defective ISV sprouting (Craig et al., 2015; Pham et al., 2007). This suggested that vascular defects in klhl4 mutants may be caused by the reduction in etv2 expression. We therefore attempted to rescue vascular defects by overexpressing etv2. For overexpression, we used TgBAC(etv2:GFP) line that we had previously generated (Proulx et al., 2010). This line shows very strong expression of GFP reporter, likely due to insertion of multiple concatemerized copies. During the BAC modification strategy used to make this line, Etv2 coding sequence was left intact, and therefore TgBAC(etv2:GFP) line shows increased etv2 expression, which was confirmed by qPCR analysis (Suppl. Fig. 3G). Differently from other strategies, such as etv2 mRNA or DNA construct injection, etv2 overexpression using BAC transgenesis is more specific, as it is restricted to the cells that normally express etv2. Activity of TgBAC(etv2:GFP) was sufficient to nearly completely rescue intersegmental sprouting and blood circulation defects in etv2^y11 mutant embryos (Suppl. Fig. 3A–D). Nevertheless, sibling klhl4 mutant embryos with or without etv2:GFP transgene showed similar ISV sprouting defects at 24 hpf stage (Suppl. Fig. 3E,F,H). This argues that increased etv2 expression level is not sufficient to rescue ISV sprouting defects in klhl4 mutant embryos.

We also attempted to rescue ISV sprouting defects in klhl4 mutants by injecting npas4l mRNA. In previous experiments, npas4l mRNA overexpression partially rescued vascular defects in cloche / npas4l mutant embryos (Reischauer et al., 2016). npas4l mRNA overexpression effectively induced etv2 mRNA expression in wild-type control embryos, confirming its activity (Suppl. Fig. 4A,B). However, npas4l mRNA injection failed to rescue ISV sprouting defects in klhl4 mutant embryos (Suppl. Fig. 4C–E), suggesting that additional factors are responsible for the failure of ISV sprouting.

DISCUSSION

In the current study we investigated the function of klhl4 in zebrafish vascular and hematopoietic development. We show that klhl4 expression is initiated in the putative hemangioblast cells, while it is later restricted to vascular endothelial cells. klhl4 zebrafish mutants display reduced intersegmental vessel sprouting, diminished vasculogenesis, increased apoptosis, reduced proliferation of endothelial and/or hematopoietic cells, and reduced erythroid differentiation. This correlates with the diminished expression of etv2 and npas4l, two key regulators of vasculogenesis and hematopoiesis.

klhl4 mutants showed significant reduction in erythroid differentiation, and reduced expression of erythroid genes gata1 and hbae3. However, ISH and HCR analysis demonstrated that klhl4 expression is restricted to vascular endothelial cells and their progenitors and is not observed in blood cells. One likely explanation is that klhl4 functions in hemangioblast cells, a common progenitor of blood and endothelial cells (Vogeli et al., 2006). Therefore, its loss of function would affect both vascular endothelial and hematopoietic cells. Indeed, our scRNA-seq and HCR analysis suggests that klhl4 is expressed in the putative hemangioblast cells and both vascular endothelial and early hematopoietic progenitors. After the two cell lineages separate, klhl4 expression is maintained in vascular endothelial cells and excluded from blood cells (Fig. 7).

Figure 7. A proposed model for Klhl4 function in hematopoietic and vascular development.

npas4l / cloche initiates klhl4 expression in hemangioblast cells. In a feedback loop klhl4 is required, directly or indirectly, to maintain npas4l expression. After hematopoietic and vascular endothelial lineages separate, klhl4 expression is only maintained in vascular endothelial cells downstream of Etv2 and is excluded from hematopoietic cells. In vascular endothelial cells Klhl4 then functions directly or indirectly to maintain etv2 mRNA and protein expression. The figure was created using Biorender.

It is intriguing that etv2 expression is downregulated in klhl4 mutants. A partial loss of etv2 expression results in reduced ISV sprouting (Pham et al., 2007; Sumanas and Lin, 2006). Furthermore, etv2 mutants show increased endothelial cell apoptosis, and defective migration of endothelial cell progenitors to the midline, which was also observed in klhl4 mutants. This suggested a possibility that vasculogenic and angiogenic defects in klhl4 mutants were at least in part due to reduced etv2 expression. However, increased etv2 expression failed to rescue ISV sprouting defects in klhl4 mutants. It is possible that the approach used for rescue, which utilized TgBAC (etv2:GFP), did not sufficiently recapitulate the timing or the level of endogenous etv2 expression. However, the same TgBAC(etv2:GFP) rescued vasculogenesis and ISV sprouting defects in etv2^y11 mutants, arguing against this scenario. A more likely explanation is that klhl4 may affect expression or activity of multiple genes involved in vasculogenesis and hematopoiesis, and etv2 expression alone is not sufficient to compensate for klhl4 absence. In fact, etv2 mutants do not show the same hematopoietic defects that are observed in klhl4 mutants, further supporting this explanation. In contrast, hematopoietic and vascular defects are observed in npas4l mutants, suggesting that klhl4 affects hematopoiesis and vasculogenesis by regulating npas4l expression or function. However, npas4l overexpression also failed to rescue ISV sprouting defects in klhl4 mutants. It is possible that mRNA overexpression did not sufficiently recapitulate the endogenous pattern of npas4l expression. Alternatively, ISV defects may be caused by klhl4 effect on multiple genes involved in vascular patterning, in addition to npas4l and etv2.

Many Klhl family members are known to target their interacting partners for ubiquitination by interacting with Cullin3-RING Ubiquitin Ligases (Dhanoa et al., 2013; Zhou et al., 2024). Ubiquitination is the most common mechanism to degrade many cellular proteins. However, substrate ubiquitination can also result in non-proteolytic outcomes, including protein activation, trafficking or changes in protein interaction (Dhanoa et al., 2013). In fact, Klhl4 has been shown to promote ubiquitination of a small GTPase PAK1, which resulted in its activation during ectodermal patterning and neurulation (Asmar et al., 2023). The mechanism of Klhl4 function during vascular and hematopoietic development is not clear and will require further investigation. It is likely that Klhl4 either directly or indirectly regulates etv2 or npas4l transcription, or alternatively, stability of etv2 or npas4l transcripts.

In summary, this study provides the first demonstration of Klhl4 requirement during vascular and hematopoietic development. Because Klhl4 protein sequence is highly conserved between different vertebrates, it is likely that mammalian Klhl4 homologs may also play similar functional roles. Further studies of Klhl4 function will advance our understanding of mechanisms involved in vasculogenesis and hematopoiesis during normal development and disease.