The wild-type embryos employed in the present study were procured through natural mating of the AB strain zebrafish. Post-fertilization, embryos were incubated in a Holtfreter’s solution at an ambient temperature of 28.5°C. Embryonic staging was conducted according to discernible morphological landmarks. For the purpose of visualizing pharyngeal pouch cells, transgenic lines Tg(sox17:GFP) and Tg(nkx2.3-mCherry) were utilized. Tg(sox10:GFP; nkx2.3-mCherry) and Tg(fli1:GFP; nkx2.3-mCherry) lines were employed to delineate both the pharyngeal arch cells and pouch cells.

Utilizing the CRISPR/Cas9 genome-editing technique, single guide RNAs (sgRNAs) were engineered to target exon 1 of cxcl12a and exon 2 of cxcr4b. The sequence corresponding to the sgRNA target in cxcl12a exon 1 was: 5’-GGT​CGC​CAT​TCA​TGC​ACC​GAT​TTC​CAA-3’. For cxcr4b exon 2, the sequence was: 5’-CTT​ACT​GTG​CCG​GCA​TCC​GG-3’. Subsequent to the CRISPR/Cas9-mediated manipulation, frameshift mutations were observed—specifically, a four-base-pair deletion in cxcl12a and a thirteen-base-pair deletion in cxcr4b. These deletions resulted in premature stop codons in the respective mRNAs. Homozygous mutants of cxcl12a ^−/− and cxcr4b ^−/− were isolated through the screening of progeny from heterozygous mutant parents.

For cxcl12a and cxcl12b, capped messenger RNAs (mRNAs) were synthesized in vitro from the linearized plasmid templates employing the mMessage mMachine kit (Ambion). Control morpholinos, constituting the sequence (5’-CCT​CTT​ACC​TCA​GTT​ACA​ATT​TAT​A-3’), were obtained along with ATG morpholinos (MOs) specifically targeting cxcl12b (5'-CGC​TAC​TAC​TTT​GCT​ATC​CAT​GCC​A-3') and MO targeting the start codon of cxcl12a (5'-ATC​ACT​TTG​AGA​TCC​ATG​TTT​GCA-3'). These were procured from Gene Tools (Philomath, OR, United States), were microinjected in accordance with previously established protocols (Dambly-Chaudiere et al., 2007; Hollway et al., 2007; Nair and Schilling, 2008).

Whole-mount in situ hybridization assays were executed in strict adherence to previously published protocols (Ning et al., 2013). In dual-color fluorescence in situ hybridization, anti-Digoxigenin-POD (11633716001, Roche) and anti-Fluorescein-POD (11426346910, Roche) were deployed as the primary antibodies to discern Digoxigenin-labeled mCherry probes and Fluorescein-labeled cxcl12a or cxcr4b probes, respectively. Subsequent analysis was performed utilizing the Perkin Elmer TAS fluorescein system (NEL701A001KT) as delineated in earlier publications (Mao et al., 2019). The composite in situ hybridization images display the following annotations: upper left quadrant designates the genotype, upper right indicates developmental stage, lower left specifies the probe in use, and the lower right reveals the aggregate count of representative samples alongside detected samples.

Transcript levels of cxcl12a, cxcr4b, hand2, and sox10 were ascertained from a composite pool of 30 embryos at specified developmental intervals. Total RNA was extracted employing TRIzol reagent (Invitrogen, Beijing, China) and purified to an A260/A280 optical density ratio of 2.0. Subsequently, 2 µg of each sample was subjected to reverse transcription utilizing 5X RT MasterMix (Abcam, Beijing, China, G592). Quantitative PCR assays were conducted employing TB Green Premix Ex Taq II (Takara, RR820A) within an Mx3000P real-time PCR system (Stratagene). Expression levels of the genes under investigation were quantified via the 2^−ΔΔCT method and were normalized relative to β-actin transcripts. The fold-change in mRNA expression is presented in relation to wild-type embryo controls.

Zebrafish embryos were anaesthetized and positioned in a 2% low melting point agarose matrix (0815, AMRESCO) immersed in egg water. Temporally specific live imaging was executed employing a Nikon A1R⁺ confocal microscope. All resultant confocal images were processed and analyzed utilizing Nikon NIS-Elements AR 4.13.00 software.

Embryos were initially fixed in a 4% paraformaldehyde solution overnight and subsequently rinsed with a 0.1% Tween-20 infused distilled water until transparency was achieved. The specimens were then stained with an Alcian blue buffer solution (comprising 0.015% Alcian Blue, 80% ethanol, and 20% acetic acid) overnight at ambient temperature. A destaining process ensued, utilizing ethanol as the solvent. Subsequently, the embryos were enzymatically treated with a 0.5% trypsin solution (0458, AMRESCO) in a supersaturated borax medium until the specimens attained a softer consistency. Post-trypsin treatment, embryos were incubated in a 1% KOH solution for 30 min, followed by rinsing in 0.1% Tween-20 infused water. Lastly, embryos were subjected to a gradient of glycerol concentrations and subsequently imaged via microscopy.

The In Situ Cell Death Detection Kit, TMR Red (Roche, 12156792910), was employed for the TUNEL assays in accordance with the manufacturer’s guidelines. Embryos harvested at specified stages were fixed in 4% paraformaldehyde at 4°C overnight, followed by a triple wash in PBST and proteinase K-mediated permeabilization. The permeabilized embryos were subsequently re-fixed in 4% paraformaldehyde for 30 min at room temperature, washed thrice in PBST, and incubated overnight at 4°C in a TUNEL-labeling milieu. Following triple PBST washes, the embryos were mounted in 2% low-melting-point agarose and visualized through confocal microscopy.

For BrdU incorporation assays, embryos were exposed to a 10 mM BrdU solution at designated stages for a duration of 20 min and were subsequently harvested. Incorporated BrdU and GFP were detected using anti-BrdU antibodies (1:1,000; B5002, Sigma) and rabbit anti-GFP antibodies (1:1,000, ab183734, Abcam) via whole-mount immunostaining.

Data were subjected to analytical scrutiny utilizing GraphPad Prism 9 software and are articulated as the mean ± standard deviation (SD). Comparisons between control and experimental cohorts were conducted through an unpaired, two-tailed Student’s t-test. Levels of statistical significance were demarcated as follows: p < 0.05 was considered statistically significant and designated with a single asterisk (*), p < 0.01 was deemed highly significant and marked with double asterisks (**), and p < 0.001 was categorized as exceedingly significant and annotated with triple asterisks (***).

Previous investigations have elucidated the pivotal role of cxcl12b in cranial neural crest cell (CNCC) migration and neurocranium morphogenesis, originating from the pharyngeal endoderm and arch mesenchyme (Olesnicky Killian et al., 2009). Additionally, dmrt2b has been identified as an upstream regulator of cxcl12b, thereby influencing pharyngeal neural crest cell ontogenesis (Li et al., 2018). In light of these findings, our study aims to interrogate the influence of cxcl12a and its cognate receptor, cxcr4b—homologous counterparts to cxcl12b and cxcr4a in zebrafish—on craniofacial development. An initial examination of the spatiotemporal expression patterns of cxcl12a and cxcr4b was conducted employing whole-mount in situ hybridization (WISH) experiments at distinct developmental stages. Intriguingly, while no expression of cxcl12a was discerned in CNCCs within the timeframe of 14–17 h post-fertilization (hpf) (from ZFIN. org), specific expression was observed in the pharyngeal region commencing at 18 hpf, coinciding with the manifestation of the first pharyngeal pouch (Figure 1A). During subsequent developmental intervals, ranging 24 hpf to 72 hpf, there was a conspicuous enrichment of cxcl12a expression in the pharyngeal sector (Figure 1B). To further delineate the precise anatomical locus of cxcl12a expression within the pharyngeal domain, dual-color fluorescence in situ hybridization assays were deployed to examine the exact expression of cxcl12a in the Tg (nkx2.3: mCherry) transgenic embryos. The outcomes, illustrated in Figure 1C unequivocally established co-localization of cxcl12a transcripts with mCherry-tagged pharyngeal pouch cells at 36 hpf. In contrast, cxcr4b exhibited expression within the pharyngeal region but was conspicuously absent in the pharyngeal pouches (Supplementary Figures S1A–C). Collectively, these data cogently substantiate that cxcl12a is distinctly localized within the pharyngeal pouches.

In pursuit of further delineating the contributory roles of cxcl12a and cxcr4b in craniofacial ontogenesis, targeted mutations in both genes were engendered via the CRISPR/Cas9 genomic editing system (Supplementary Figures S2A, B). Notably, aberrations in nanos3 localization, a marker for the primordial germ cells, were discernibly manifested in homozygous cxcl12a and cxcr4b mutants (Supplementary Figure S2C) (Boldajipour et al., 2008). Subsequent evaluations employing whole-mount in situ hybridization (WISH) and quantitative real-time polymerase chain reaction (qRT-PCR) revealed precipitous decrements in the expression levels of both cxcl12a and cxcr4b within these mutants (Supplementary Figures S2D, E). These observations rigorously authenticate these mutants as bona fide loss-of-function strains.

A comprehensive phenotypic scrutiny revealed that cxcl12a mutants exhibited symptoms of pericardial edema coupled with mandibular hypoplasia at 5 days post-fertilization (dpf), anomalies amenable to amelioration via cxcl12a mRNA microinjection (Figure 2A). In contrast, cxcr4b ablation failed to induce any overt morphological perturbations in the resultant mutants (Figure 2A). Alcian blue histochemical analyses revealed a substantive attenuation in the structural integrity of both Meckel’s cartilage and palatoquadrate, as well as a marked diminution in the number of ceratobranchial cartilages in cxcl12a mutants, thereby implicating compromised chondrogenic differentiation (Figure 2B). Remarkably, these phenotypic aberrations could be substantially rectified through cxcl12a mRNA microinjection (Figure 2B). Concomitantly, cxcr4b mutants exhibited no such craniofacial cartilage deformities (Figure 2B). Furthermore, gene expression assays targeting the extracellular primary cartilage matrix marker, col2a1a, undertaken at 72 hpf, ascertained a notable reduction in its expression levels exclusively in cxcl12a mutants, a condition reversible upon cxcl12a mRNA supplementation (Figure 2B).

To augment our understanding of cxcl12a's involvement in craniofacial cartilaginous morphogenesis, an antisense cxcl12a MO targeting cxcl12a mRNA—a MO previously validated—was employed to attenuate cxcl12a expression (Gamba et al., 2010; Jiang et al., 2019). To assess the efficacy of this MO-mediated silencing, a plasmid harboring both the target sequence of the cxcl12a MO and eGFP was synthesized. MO efficiency was subsequently substantiated through the quantification of fluorescence intensity (Supplementary Figure S3A). Perturbations in nanos3 expression were also evident following cxcl12a knockdown (Supplementary Figure S3B), corroborating both the specificity and efficacy of the cxcl12a MO. Consonant with observations in cxcl12a mutants, morphants displayed manifest aberrations in pharyngeal cartilage formation, defects that were ameliorable by cxcl12a mRNA injection (Supplementary Figure S2D). Collectively, Based on these data, these findings validate that the observed craniofacial cartilage malformations in either cxcl12a mutants and morphants emanate from cxcl12a transcriptomic attenuation, effectively obviating off-target effect of CRISPR/Cas9 effects.

Previously, the Cxcl12b-Cxcr4a axis has been delineated as requisite for CNCC migration and patterning during craniofacial ontogenesis (Olesnicky Killian et al., 2009). Given this, we postulated that cxcl12a and cxcl12b might exert synergistic functions in this developmental process. To interrogate this hypothesis, we deployed a previously characterized cxcl12b MO to suppress cxcl12b expression in cxcl12a mutants (Cha et al., 2012; Nguyen et al., 2014). The efficaciousness of cxcl12b MO was ascertained via fluorescence intensity metrics and in situ hybridization employing a cxcl12b-specific probe (Supplementary Figures S3A, C). Evaluation of cartilaginous elements in cxcl12b morphants revealed a pronounced diminution or outright absence of the ethmoid plate, coupled with trabecular fusion, corroborating prior literature consistent with previous report (Supplementary Figure S3D) (Olesnicky Killian et al., 2009). Remarkably, when cxcl12b MO was introduced into cxcl12a mutants, the resultant phenotypic distortions in pharyngeal cartilage were exacerbated; specifically, both Meckel’s cartilage and palatoquadrate were gravely malformed, and ceratobranchial cartilages were virtually obliterated (Figures 2A, B). Intriguingly, the simultaneous microinjection of cxcl12a and cxcl12b mRNA substantively restored craniofacial cartilage integrity in double-knockdown morphants (Figures 2A, B). Thus, our findings affirm that cxcl12a principally orchestrates craniofacial cartilage formation, whilst cxcl12b downregulation independently induces ethmoid plate atrophy and trabecular fusion.

It is universally acknowledged that cranial neural crest cells (CNCCs) originate from the neural plate and subsequently undergo an epithelial-to-mesenchymal (ETM) transition. Following this transformation, CNCCs traverse toward pharyngeal arches and eventually differentiate into the chondrocytes constituting facial cartilage (Dash and Trainor, 2020). In order to elucidate the ramifications of cxcl12a deletion on CNCC activity, we ascertained the expression profiles of specific neural crest cell (NCC) markers across temporal stages. Remarkably, the expression of the neural crest cell specification marker foxd3 remained congruent among control embryos, cxcl12a mutants and cxcl12a ^{cxcl12b MO} at the 5-somite stage (Figure 3). This uniformity in expressions trends indicates that cxcl12a is inconsequential to the induction and specification of the early neural crest cells.

To further scrutinize the impact on NCC migration, we employed dlx2a and crestin as migration markers. Within the time frame of 18–24 h post-fertilization (hpf) in control embryos, both dlx2a and crestin transcripts were profusely localized within the three neural crest cell groups. Nonetheless, only a marginal decrement in the expression levels of dlx2a and crestin was discernible in cxcl12a mutants at both 18 hpf and 24 hpf (Figure 3A). Conversely, injection of cxcl12b MO into cxcl12a mutants exacerbated the impediments to NCC migration (Figure 3A). In vivo imaging of the Tg(sox10-EGFP) line at 24 hpf corroborated that arch structures in control embryos and cxcl12a mutants were comparable, whereas the number of CNCCs in cxcl12a ^{cxcl12b MO} organisms was conspicuously reduced, aligning with prior studies (Figure 3B). Collectively, these findings substantiate the notion that the abrogation of cxcl12a exerts a minimal influence on CNCC migratory patterns.

Extant literature delineates that the post-migratory CNCCs in the arches ultimately differentiate into cartilaginous chondrocytes (Hirata and Hall, 2000; Clouthier and Schilling, 2004), a process modulated by pharyngeal pouches serving as pivotal signaling centers (Piotrowski and Nusslein-Volhard, 2000; Couly et al., 2002; Crump et al., 2004). Given the localized expression of cxcl12a in these pharyngeal pouches post-CNCC migration (Figure 1C), we scrutinized the repercussions of cxcl12a depletion on these cells by assessing the expression profiles of hand2 and sox10 at 36 hpf. Notably, a quantitative decline in CNCCs was observed in both cxcl12a mutants and morphants in comparison to these control embryos, a reduction that could be successfully mitigated through the exogenous administration of cxcl12a mRNA (Figure 4A and Supplementary Figure S3E). To corroborate these observations, we employed quantitative reverse-transcription polymerase chain reaction (RT-qPCR) and whole-mount in situ hybridization (WISH) analyses, both of which yielded results in consonance with the initial assessments (Supplementary Figure S4). In line with these findings, comparable results were observed in cxcl12a morphants in Tg(nkx2.3:mCherry;fli1:GFP) or Tg(nkx2.3:mCherry; sox10:GFP) background, characterized by GFP expression in CNCCs at 36 hpf. These morphants exhibited disarrayed and abbreviated pharyngeal arch structures subsequent to cxcl12a knock-down (Figures 4B, C). Moreover, pharyngeal pouch patterning was exacerbated in cxcl12a ^{cxcl12b MO} mutants (Figures 4A–C).

To discern whether this observed phenotype could be attributed to alterations in CNCC apoptosis or proliferation, TUNEL staining and Bromodeoxyuridine (BrdU) incorporation assays were conducted on the Tg(sox10:EGFP) transgenic reporter line. Intriguingly, apoptosis levels in chondrogenic cells remained invariant between cxcl12a mutants and control embryos (Figures 4E, F). However, a pronounced diminution in the proliferation of BrdU⁺ CNCCs was detected following cxcl12a depletion (Figures 4G, H). Further compounding this effect, injection of cxcl12b MO into cxcl12a mutants culminated in even more severe pharyngeal arch malformations, attributable to a substantially diminished proliferative capacity of CNCCs (Figure 4 and Supplementary Figure S4).

Collectively, these empirical findings substantiate the indispensable role of cxcl12a in fostering CNCC proliferation subsequent to their migration.

Previous empirical evidence presented in Figures 4B, C elucidated that microinjection of cxcl12a MO in transgenic embryos with the Tg(nkx2.3:mCherry;fli1:GFP) or Tg(nkx2.3:mCherry;sox10:EGFP) background engendered a discernible diminution in mCherry-expressing pharyngeal pouches at 36 hpf. Given the co-localization of cxcl12a expression within the pharyngeal pouch, we conducted a meticulous examination of pharyngeal endoderm marker nkx2.3 and cv2 in both control and cxcl12a mutant embryos, either in the presence or absence of cxcl12b MO injections. Corroborating our earlier observations, the patterning of the pharyngeal pouches appeared perturbed in cxcl12a mutants consistent with our observation (Figure 5A).

To substantiate our preliminary findings, we deployed transgenic embryos featuring the Tg(nkx2.3-mCherry) and Tg(sox17-GFP) genetic backgrounds to serve as indicators of pharyngeal pouch cells. Our analysis revealed that the structural integrity of the pharyngeal pouch was compromised upon cxcl12a MO injection, manifesting as disorganized patterning (Figures 5B–D). Additionally, concomitant microinjection of cxcl12a MO with cxcl12b MO resulted in exacerbated perturbations in pharyngeal pouch morphology (Figures 5B–D). In summation, these empirical observations validate the cooperative functionality of cxcl12a and cxcl12b in orchestrating the formation of the pharyngeal pouch.