dock8 deficiency attenuates microglia colonization in early zebrafish larvae

Microglia are tissue-resident macrophages that carry out immune functions in the brain. The deficiency or dysfunction of microglia has been implicated in many neurodegenerative disorders. DOCK8, a member of the DOCK family, functions as a guanine nucleotide exchange factor and plays key roles in immune regulation and neurological diseases. The functions of DOCK8 in microglia development are not fully understood. Here, we generated zebrafish dock8 mutants by CRISPR/Cas9 genome editing and showed that dock8 mutations attenuate microglia colonization in the zebrafish midbrain at early larvae stages. In vivo time-lapse imaging revealed that the motility of macrophages was reduced in the dock8 mutant. We further found that cdc42/cdc42l, which encode the small GTPase activated by Dock8, also regulate microglia colonization in zebrafish. Collectively, our study suggests that the Dock8-Cdc42 pathway is required for microglia colonization in zebrafish larvae.

The term microglia was first used in 1919 and morphological and functional characterization of these cells from development to injury was described by the Spanish neuroscientist Pio del Rio-Hortega [12]. In mice, microglial precursors, namely erythromyeloid precursors (EMPs), are derived from the yolk sac mesoderm (E7.0-E8.5) [13][14][15][16]. EMPs generate various types of tissue-resident macrophages, including those moving to the brain to differentiate into microglia (E9.5) [14,17,18]. Zebrafish microglia colonization in the brain is remarkably similar to that observed in the mice. The microglia are derived from two sources in zebrafish. Embryonic microglial precursors initiate from the rostral blood island (RBI), the equivalent of mouse YS for myelopoiesis, at 11 h postfertilization (hpf) [19]. These cells move to the brain at 60 hpf [20]. Adult microglial precursors arise from the ventral wall of the dorsal aorta (VDA) and began to populate the brain from 15 days postfertilization onward (dpf) [19]. Embryonic microglia of zebrafish are transient and are eventually replaced by adult microglia [21]. At embryonic stages, zebrafish microglial precursors enter the optic tectum of the midbrain by two paths, including the lateral periphery between the eyes and brain and the ventral periphery of the brain in a circulation-independent manner [22]. The Il34-Csf1r pathway could regulate microglial precursor migration to the proximal brain regions [23]. During neuronal apoptosis, lysophosphatidylcholine and nucleotide signaling promote microglial precursor entry into the brain [22,24]. Slc7a7, a Leu/Arg transporter necessary for microglial brain colonization, has recently been discovered to label a macrophage sub-lineage that migrates to the brain to become microglia [25]. Mobility is a basic property of macrophages. Although factors affecting the mobility of microglia/macrophages have been extensively studied, studies have primarily been carried out in vitro or under pathological conditions [26][27][28]. How the motility of macrophages affects their brain colonization during animal development is less studied.
In this study, we generated dock8 mutants by CRISPR/Cas9 using zebrafish as a model organism to determine whether dock8 mutation affects microglia colonization in the zebrafish brain. Our results suggest that dock8 deficiency attenuates microglia colonization via the DOCK8-Cdc42 pathway by impairing microglia motility in zebrafish early larvae, although the overall the number of macrophages was not affected.

RESULTS
dock8 is expressed in macrophages in zebrafish embryos To test whether dock8 could play a role in microglia colonization, we first examined whether dock8 was expressed in macrophages. Expression of dock8 was detected by FISH in Tg(mpeg1:eGFP) embryos. As shown in Fig. S1, dock8 positive signals largely overlapped with eGFP + macrophages from 1-2 dpf (Fig. S1D). However, dock8 was hardly detected in macrophages by FISH after 3 dpf (data not shown). We speculate that the expression of dock8 in macrophages might decrease as the fish grows. Indeed, we found that dock8 was expressed in macrophages at 5 dpf by analyzing published single-cell sequencing data, although at a lower level compared with that at 1 dpf ( Fig. S1A-C) [44,45].

Establishment of the dock8 mutant
To generate a zebrafish dock8 mutant, we disrupted the DHR2 domain which is essential for Dock8 function [34]. We first performed multiple sequence alignment for the DHR2 domain using online tools (http://multalin.toulouse.inra.fr/multalin/). As shown in Fig. S2, the DHR2 domain of zebrafish Dock8 shares over 80% identity with the DHR2 domain of human and mouse. We next designed gRNAs to target exon 39 or exon 45 of the dock8 gene, where the codons represented conserved amino acids in the DHR2 domain of the dock8 gene.
We generated dock8 mutants by co-injection of Cas9 protein and sgRNA into zebrafish embryos at the single-cell stage. Two mutant alleles, with mutations in exon 39 and 45, were recovered. The dock8 −2bp allele lost 2 bp in exon 39 and the dock8 −15, +5bp allele lost 15 bp but gained an extra 5 bp in exon 45 (Fig. 1A). To check whether these two mutant alleles broke the Dock8 protein, we analyzed the codon sequences and found that premature stop codons were introduced in both mutant alleles (Fig. 1B), suggesting possible disruption of Dock8 functions.
dock8 deficiency reduces the number of microglia in zebrafish larvae To determine whether dock8 mutation affects microglia colonization, we first investigated the signal of neutral red (NR), which labels microglia lysosomes in the developing brain at 4 dpf when the microglia population has been well established in the brain [46]. The average number of NR signals in dock8 −15,+5bp mutant embryos (~15 cells) was largely reduced compared with that of sibling embryos (~32 cells) (Fig. 2B, E). To further confirm the reduction of microglia in dock8 mutants, we examined another microglial marker: apolipoprotein Eb (apoeb) by WISH and the macrophage marker mpeg1 using mpeg1 transgenic fish. A similar reduction in the apoeb signal and mpeg1-DsRedx + macrophages/ microglia was observed in the brain of dock8 −15,+5bp mutant embryos at 4 dpf ( Fig. 2C, D, F, G). Interestingly, macrophages in caudal hematopoietic tissue (CHT) showed no obvious differences between mutants and siblings suggesting that the microglia defect in mutants was not due to an overall reduction of macrophages (Fig. S1E).
To confirm that the reduction of microglia at early developmental stages was indeed caused by the dock8 deficiency, we performed a rescue experiment. The in vitro-synthesized dock8 mRNA failed to rescue the dock8 −15,+5bp mutants, likely because the in vitro-synthesized mRNA could not last to 4 dpf. We then examined microglia in dock8 −2bp mutants to determine whether similar defects could be observed in a second mutant allele. Our results showed that both NR and apoeb signals were reduced in dock8 −2bp mutants at 4 dpf (Fig. 3). More importantly, similar microglia defects were reproduced in the compound dock8 −15,+5bp/−2bp mutants at 4 dpf (Fig. 3). These results suggest that the reduced microglia are indeed due to dock8 mutations.
dock8 deficiency impairs the motility of macrophages in zebrafish larvae Some studies have reported that Dock8 regulates macrophage/ microglia migration under pathological conditions [33,43]. We hypothesized that dock8 deficiency affected microglia colonization by regulating motility of macrophages. To test this hypothesis, we first observed the basic motility of macrophages. Interestingly, we found dock8 mutant macrophages tend to move more in close range and exhibited slower migration speeds compared with siblings in the yolk sac at 3 dpf ( Fig. 4 and Video S 1). These studies indicate that dock8 deficiency impairs the motility of macrophages and could consequently affect their migration to the brain in zebrafish larvae.
dock8 deficiency reduces microglia colonization in early zebrafish larvae To confirm that impaired macrophage motility of dock8 mutants leads to microglia paucity at early developmental stages, we directly monitored the colonization of microglia from 2.5-3 dpf by live imaging. As shown in Fig. 5 and Video S 2, the average number of GFP positive microglia migrating into the brain in dock8 −15,+5bp mutant embryos was significantly less than those in sibling embryos. These studies suggest that dock8 deficiency reduces microglia colonization by impairing its motility in early zebrafish larvae.
cdc42 acts downstream of dock8 to affect microglia colonization in zebrafish larvae Dock8 is known to mediate Cdc42 activation and subsequent migration of DCs, T cells, and macrophages [33][34][35]. We then tested whether microglia colonization is regulated by the DOCK8-CDC42 module in zebrafish. There are three cdc42 homologs, cdc42, cdc42l, and cdc42l2, in zebrafish. Interestingly, analysis of published single-cell sequencing data showed that only cdc42 and cdc42l were readily detected in embryonic macrophages (Fig. S3) [44,45]. We hypothesized that cdc42 and cdc42l, but not cdc42l2, could be involved in macrophages. Therefore, we generated deletions of exon 3 of cdc42 and of exon 2 of cdc42l. The resulting cdc42 mutant lost 20 bp and the cdc42l mutant lost 29 bp. These deletions resulted in a premature stop codon in Cdc42 and Cdc42l (Fig. S4).
To determine whether cdc42 or cdc42l affected microglia colonization, we investigated the NR signal in these mutants. NR signals in cdc42 -/or cdc42l -/single mutants were similar to wildtype embryos at 3 dpf suggesting that the number of microglia was normal in single mutants (Fig. 6B). Interestingly, NR signals in cdc42 -/-cdc42l +/-(~17.66 cells), cdc42 +/-cdc42l -/-(~12.72 cells), and cdc42l -/-cdc42l -/embryos (~2.7 cells) were significantly reduced compared with that of double wild-type embryos (~28.89 cells) (Fig. 6). We further confirm a reduction of microglia in cdc42 and cdc42l mutants by WISH (Fig. S5). A reduction of the mpeg1 signal was observed in the brains of cdc42 -/-cdc42l +/+ , cdc42 -/-cdc42l +/-, cdc42 +/-cdc42l -/-, and cdc42l -/-cdc42l -/embryos. By contrast, there were no obvious differences between mutant and sibling macrophages in the body (Fig. S5), suggesting that the microglia defect in the mutants was not due to an overall reduction of macrophages. We noticed that cdc42 -/-cdc42l +/+ only showed a slight reduction of mpeg1 signals but not NR staining. This could have been a result of different staining methods or different batches of fish. Taken together, these results suggest that cdc42 and cdc42l redundantly regulate early microglia colonization.

DISCUSSION
In this study, we reveal that dock8 mutation affects microglia colonization in zebrafish early larvae. We further showed that cdc42/cdc42l mutants have similar defects in microglia colonization suggesting that the DOCK8-CDC42 pathway regulates microglia colonization. Until now, little has been known of the factors upstream of Dock8 to regulate microglia colonization. The Il34-Csf1r pathway has been reported to regulate the migration and colonization of microglial precursors in zebrafish [23]. Interestingly, CSF-1, a ligand of CSF1r, activates Rac and Cdc42 [47]. In addition, Vav, as a Rac-GEF, is activated by CSF-1 to participate in cytoskeletal changes in osteoclasts [47]. Importantly, Cdc42 regulates CSF-1-induced polarization of macrophages [48]. It would be interesting to explore whether Csf1r functions upstream of Dock8 to regulate the migration and colonization of microglial precursors.
Our investigations revealed that dock8 deficiency impaired microglia and/or macrophage motility through the DOCK8-CDC42 pathway. However, how DOCK8 regulates CDC42 to affect these processes remains unclear. Interestingly, Cdc42, via DOCK8, binds effector molecules, such as myotonic dystrophy kinase-related Cdc42-binding kinase (MRCK), through facilitated myosin II (MLC2) phosphorylation to regulate macrophage migration [33]. In addition, CDC42-GTP generated by DOCK8 binds WASp to regulate actin polymerization, T cell transendothelial migration, and homing to lymph nodes [49]. Therefore, we speculate that DOCK8 might regulate CDC42 to affect migration colonization through the above mechanisms.
Our results show a 50% reduction in microglia numbers in dock8 mutants at 4 dpf, but the number of peripheral macrophages was not affected, suggesting dock8 is not indispensable for early macrophage development. Interestingly, the recovery of microglia numbers in dock8 mutants suggests that the colonization of microglia is only temporarily affected. This result suggests that complementary machinery exists to promote the colonization of microglia in dock8 mutants. We speculate that other DOCK family proteins could play important roles, as Dock proteins are known to regulate the actin cytoskeleton, cell adhesion and migration [50]. Recent reports have suggested that DOCK10 and DOCK11 function as Cdc42 and/or Rac1 GEFs to modulate the mobility of C, F, I Representative images (C) and quantification of 4 dpf (F) and 6 dpf (I) apoeb WISH signals of dock8 -15,+5bp mutants and siblings. D, G, J Representative images (D) and quantification of 4 dpf (G) and 6 dpf (J) mpeg1-dsredx positive cells in optic tectum of dock8 -15,+5bp mutants and siblings. Group sizes were at least n = 30 zebrafish embryos. Each dot represents one larva. White dashed lines indicate the optic tectum. Scale bar = 100 µm. Data were analyzed by unpaired Student's t-tests. ns P > 0.05; **** P ≤ 0.0001. Fig. 3 dock8 is the causative gene of microglia deficiency. A Representative images and quantification of 4 dpf NR signals of dock8 -2bp mutants, dock8 -15,+5bp/-2bp compound mutants and siblings. B Representative images and quantification of 4 dpf apoeb WISH signals of dock8 -2bp mutants, dock8 -15,+5bp/-2bp compound mutants and siblings. Group sizes were at least n = 30 zebrafish embryos. Each dot represents one larva. White dashed lines indicate the optic tectum. Scale bar = 100 µm. Data were analyzed by unpaired Student's t-tests. ** P ≤ 0.01; **** P ≤ 0.0001. macrophages and/or microglia [51,52]. Therefore, DOCK8 could work together with other DOCK family members to regulate microglia colonization. We also noticed that the phenotype of microglia colonization is more severe in the cdc42/cdc42l double mutant compared with the dock8 mutant. One possible reason is that other DOCK proteins, such as DOCK10 and DOCK11 [52], function in parallel with DOCK8 to modulate CDC42 activities together. Another possibility is there are other signaling events to activate CDC42. Further studies are warranted to reveal other machineries in addition to the Dock8-Cdc42 pathway.

Neutral red staining
Neutral red solution (Cat#G1315, Solarbio) was diluted 1:1000 in egg water with 0.003% PTU. Zebrafish embryos were incubated in this solution for 5-6 h at 28.5°C in the dark. Images were taken with a fluorescence stereomicroscope (zoom V16) using a 100x objective. Quantification was performed manually.
In vitro synthesis of antisense RNA probe, whole-mount in situ hybridization, and antibody staining Antisense DIG labeled RNA probe of dock8 was synthesized in vitro according to the standard protocol. Primers for dock8 probe synthesis were as follows: XmaI-FP: ctgcagcccgggtcatgcaggagttgggtgag, XbaI-RP: ggccgctctagatcacagcatccggtttctcc. The apoeb probe was described previously [20,60]. The mpeg1 probe was a gift from Professor Zilong Wen of Hong Kong University of Science and Technology. Whole mount in situ (WISH) was performed as described previously [61] with NBT/BCIP (11697471001, Roche) or BM purple AP (11442074001, Roche) or TSA TM Plus Cyanine 3 System (NEL744001KT, Perkin-Elmer). Images were captured with a fluorescence stereomicroscope (zoom V16) using a 100x or 40x objective. Quantification was performed manually.

Time-lapse imaging and cell tracking analysis
Time-lapse imaging was performed as previously reported with some modifications [22]. Embryos were anesthetized in 0.01% tricaine (A5040; Sigma), mounted in 1% low-melting agarose and covered with egg water (0.01% tricaine and PTU). Then embryos were imaged on a Zeiss LSM 880 confocal microscope with a 10x or 20x objective in a 28.5°C thermal chamber. Z-step size was set at 3 µm; usually, 28-30 Z steps for embryos at 2.5-3 dpf or 13-18 Z steps for embryos at 3 dpf are used at 2-3 min intervals. The total observation time was 12-16 h for the fish brain or 4 h for the fish yolk using the 488 nm and/or 561 nm laser. Images were processed with Zeiss ZEN or Imaris software.