Foxg1 Regulates the Development of Microvillous Cells in the Olfactory Epithelium

Abstract

Foxg1 is a transcription factor that plays important roles in the development of the nervous and sensory systems. In the olfactory epithelium, it is known that Foxg1 is necessary for the development of subsets of olfactory sensory neurons, though its precise role remains undefined. In this study, we show that in zebrafish Foxg1a is required for the formation of microvillous olfactory sensory neurons, as well as regulation of subsets of sensory neurons and progenitor cells in the embryonic olfactory epithelium. Using the foxg1a^a266 zebrafish line, we found that sox10 and trpc2b expression are both absent from the olfactory epithelium in these mutants, suggesting that there are defects in the development of microvillous olfactory sensory neurons. Another major class of olfactory sensory neurons in the olfactory epithelium, ciliated cells, appear to develop normally in foxg1a^a266 mutants. Additional groups of neurons, such as Islet1-postive sensory neurons and mechanosensory rod cells are decreased but not absent in foxg1a^a266 mutant embryos. Progenitor cell populations in the olfactory epithelia are differentially affected, with an increase in the number of Sox2-positive cells and a decrease in Sox3-positive cells in foxg1a^a266 mutant embryos as compared to heterozygous sibling controls. Finally, we show that foxg1^a266 embryos do not form olfactory epithelium neuron projections to the brain. These results show a nuanced role for Foxg1 in the developing olfactory epithelium.

Graphical Abstract

1. Introduction

Cellular heterogeneity in sensory tissues is essential for nuanced discrimination of a wide variety of stimuli. Sensory systems rely on this heterogeneity for increased acuity of sensation in modalities such as vision (Balla et al., 2025), taste (Roper and Chaudhari, 2017), hearing (Shrestha et al., 2023), and smell (Tepe et al., 2018). Cell types are defined partially by how they specifically receive sensory information through specialized receptors (Khalil et al., 2025). Vision relies on a variety of cells such as cones for the perception of color and rods for photosensitivity (Thoreson and Dacey, 2019). Gustatory responses depend on diversity in taste receptor cell types for which respond to distinct tastants to enhance survival through caloric modulation to impact health (Doyle et al., 2023). The inner ear depends on differences in hair cells and sensory sensations to transduce hearing and vestibular responses (Burns et al., 2015). Olfactory diversity arises from populations of olfactory sensory neurons (OSNs) that express a dramatically broad range of receptors allowing them to respond to a massive variety of odorants (Lizbinski and Dacks, 2018).

Olfaction requires this heterogeneity to distinguish odorants involved with biological functions, including feeding, mating, and survival (Barnard et al., 2025; Beauséjour et al., 2025; Wei et al., 2025). The OSNs in the olfactory epithelium (OE) detect chemical cues present in the environment. Within the OE, differences in olfactory receptor expression aids in the organism’s ability to identify kin (Biechl et al., 2016), pursue prey (Hughes et al., 2010), and escape danger (Masuda et al., 2024). Determining how diverse sensory cell populations arise in the OE is important for understanding the development of the olfactory system and how these distinct modalities are disrupted during disease.

Zebrafish present a distinct advantage to the study of olfactory system development through their external fertilization, quick development, and the easy accessibility of the OE. Several types of cells exist within the teleost OE, with ciliated and microvillous cells among the most numerous broad classes of OSNs (Villamayor et al., 2021). Ciliated cells are distinguished by the presence of microtubule-based apical cilia, which aid in water movement to increase odor detection, to create odor fields (Reiten et al., 2017), and sense odors like amines, food extract, and bile acids (DeMaria et al., 2013). Microvillous cells, which contain apical microvilli rather than cilia, respond to amino acids (DeMaria et al., 2013) as well as modulate survival of other olfactory neurons (Genovese and Tizzano, 2018; Lemons et al., 2020). Within zebrafish, the high-mobility group transcription factor SRY-box transcription factor 10 (Sox10) is required for the development of olfactory placode-derived microvillous OSNs (Saxena et al., 2013, Aguillon et al., 2018).

One marker known for its early expression in the OE is the winged-helix transcription factor forkhead box G1 (Foxg1). In addition to the OE, Foxg1 has been shown to contribute to brain development through the control of neuronal differentiation (Hou et al., 2020). This neurogenic function also occurs in the OE, with Foxg1 promoting olfactory neuronal differentiation (Garaffo et al., 2015; Kawauchi et al., 2009). Further, this process is conserved since Foxg1 is required for the differentiation of a subset of neurons the OE in mice and zebrafish (Duggan et al., 2008). Yet, Foxg1 plays a pleiotropic role in development (Schäffner et al., 2023). Within these structurally distinct cell types, heterogeneity is found by the expression of one or a few of the large family of odorant receptors (Calvo-Ochoa and Byrd-Jacobs, 2019; Michel and Derbidge, 1997). In mouse, Foxg1 aids in dorsal-ventral zoning of the OE, driving cells toward a ventrolateral fate (Kuriyama and Hanashima, 2025). However, Foxg1 restriction to a heterogenous subset of cells in is preceded by its generalized expression within the developing olfactory placode (Duggan et al., 2008; Kuriyama and Hanashima, 2025). With the connection of olfactory function and neurodevelopmental disorders (Sweat and Cheetham, 2024), it is important to investigate the precise function of Foxg1 within the developing OE.

This study investigates the role of Foxg1a during OE development using an established foxg1a mutant zebrafish line, which was created during a screen for schizophrenia related genes (foxg1a^a266; Thyme et al., 2019). The foxg1a^a266 allele contains a large deletion in the single foxg1a exon and is predicted to be a null mutation (Thyme et al., 2019). Previous research in rodents suggests that Foxg1 may act upstream of Sox10 in the developing OE (Duggan et al., 2008), though the connection of Foxg1 to Sox10 has not been thoroughly investigated. Here, we demonstrate in zebrafish that Foxg1a function is required for the formation of sox10-positive cells in the developing OE and the subsequent formation of a subset of OSNs. In foxg1a^a266 mutant embryos, we find a loss of the microvillous cell marker trpc2, a decrease in the number of OSNs labeled with Calretinin and Tg(sox10:GFP), a near loss in rod cells, and reduction in Islet-1-positive sensory neurons. In addition, we find changes in the populations of Sox2 and Sox3 OE progenitor cell populations in foxg1a^a266 mutants. This suggests that Foxg1 regulates the development of a range of neuronal and progenitor cell types in the OE. These findings suggest human patients with Foxg1 syndrome (Brimble et al., 2025) could be assessed for potential olfactory deficits.

2. Materials and Methods

2.1. Zebrafish Care and Maintenance

Zebrafish lines used in this study are as follows: foxg1a^a266 (Thyme et al., 2019), Tg(sox10:EGFP)^ba2 (Dutton et al., 2009), and Tg(prim:lyn₂mCherry) (J. Wang et al., 2018). Adult zebrafish were maintained in a facility with a 14-/10-hour light/dark cycle at 28°C. Embryos and larvae were grown in E3 embryo medium (14.97 mM NaCl, 500 μM KCL, 42 μM Na2HPO4, 150 μMKH2PO4, 1 mM CaCl2 dihydrate, 1 mM MgSO4, 0.714 mM NaHCO3, pH 7.2) at 28°C and in accordance with standard protocols (Kimmel et al., 1995). The fish used in this study were treated with tricaine (Syndel) before fixation using 4% paraformaldehyde (PFA) in phosphate buffered saline (PBS) (Thermo Fisher). All experiments were performed according to McGraw Laboratory protocol #445344, approved by the UMKC IACUC.

2.2. Immunohistochemistry

Whole mount immunohistochemistry was adapted from established protocols (Ungos et al., 2003). Briefly, embryos were fixed in 4% PFA at 4°C overnight or at room temperature for 1 hour. All washes used 1x PBS/0.1% Triton X 100 (PBSTx). Samples were washed before a 1-hour incubation in an antibody block solution (1x PBS, 0.2% Triton-X 100, 2mg/ml BSA, 1% DMSO, 0.02% sodium azide, 2% Goat Serum), then washed and incubated in primary antibody (concentrations are antibody specific) at 4°C overnight. Following primary antibody incubation, samples were washed in PBSTx and then incubated overnight in secondary antibody (1:1000 in blocking solution) at 4°C, washed, and labeled with 30μM DAPI (Thermo Fisher) in PSB, and then stored in 50% Glycerol/PBS. Primary antibodies used were as follows: α-Calretinin (mouse monoclonal, 1:100, Sigma-Aldrich), α-Acetylated Tubulin (mouse monoclonal, 1:1000, Sigma), α-HuC/HuD (Elavl3; mouse monoclonal, 1:200, Invitrogen), α-Sox2 (rabbit polyclonal, 1:100, GeneTex), α-Sox3 (rabbit polyclonal, 1:100, GeneTex), and α-Islet1 (mouse monoclonal, 1:100, Millipore Sigma). Secondary antibodies used are goat α-rabbit Alexa-647 (1:1000, Invitrogen), goat α-mouse Alexa-568 (1:1000, Invitrogen), and goat α-mouse Alexa-647 (1:1000, Invitrogen). Samples using α-Acetylated tubulin were co-stained with Alexa Fluor 568 phalloidin (Invitrogen). Antibodies for Sox2, Sox3, and Islet1 were fixed for one hour, followed by a water wash overnight, and two nights of primary incubation, all steps at room temperature.

2.3. BrdU Incorporation

Cellular proliferation was assessed using incorporation of 5-Bromo-2’-dioxyuridine (BrdU; Millipore Sigma). BrdU incorporation protocols were adapted from previously established protocols. Briefly, 1-,2-,3- or 4-day old zebrafish were transferred into 6-well dishes containing 10mM BrdU in E3 embryo medium for a pulse of BrdU incorporation (30 min at 4°C followed by 28°C for1 hour), embryos were fixed with 4% PFA over night at 4°C. Following fixation in 4% PFA, samples were washed using 1x PBS-0.1% Tween-20 1% DMSO (PBDT) and transferred to 100% methanol at −20°C for future use. After serial rehydration of samples using PBDT, a 5-minute incubation with 10μM/ml Proteinase K was preformed, followed by a 15-minute post-fixation using 4% PFA. Samples were washed with PBDT before a 1-hour incubation in 1N HCl at room temperature. Samples were washed with PBDT and placed in a blocking solution of 10% goat serum in PBDT for 1 hour, followed by incubation in 1:100 α-BrdU antibody solution (BD Biosciences) overnight at 4°C. Post-primary antibody incubation, embryos were washed with PBDT before 4°C incubation in a secondary antibody overnight in the dark. Finally, after secondary antibody was washed using 1x PBS/ 0.1% TritonX-100, the samples were labeled with 30μM DAPI (Thermo Fisher) in PSB then stored in 50% Glycerol/PBS.

2.4. TUNEL Labeling

Cell death was assessed through adapting the Click-IT Plus TUNEL (terminal deoxynucleotidyl transferase dUTP nick end labeling) protocols (Thermo Fisher). Briefly, embryos were fixed in 4% PFA for 1 hour at room temperature, washed twice with PBTw, permeabilized with 10μg/ml Proteinase K for 5 minutes and refixed with 4% PFA. Once washed from fixative, samples were rinsed with deionized H₂O, incubated in TdT reaction buffer for 10 minutes at 37°C, then in TdT reaction mixture (TdT reaction buffer, EdUTP, and TdT enzyme) before rinsing with deionized H₂O, washed with 3% BSA in PBS, and rinsed in PBS. Following this, samples were incubated in Click-It TUNEL reaction cocktail (1x Click-It Plus TUNEL reaction buffer, Copper protectant, Alexa Fluor picolyl azide, 1x Click-It Plus TUNEL Reaction buffer additive) at 37°C protected from light, then washed with 3% BSA in PBS, and rinsed with PBS. Samples were stored in 50% Glycerol/PBS.

2.5. Whole Mount in situ Hybridization

Whole mount RNA in situ hybridization (WISH) was carried out using established protocols (Thisse & Thisse, 2008). Briefly, embryos or larvae were fixed in 4% PFA in PBS overnight at 4°C, then stored in methanol at −20°C until processed. Samples were put through these steps: they were serially washed out of methanol in 1x PBS/0.1% Tween-20 (PBTw), incubated in 10μg/ml Proteinase K for 5–10 minutes, post-fixed in 4% PFA for 15 minutes, washed in PBTw, pre-hybridized at 65°C for 1 hour in hybridization mix (50% formamide, 5x saline-sodium citrate (SSC), 5 μg/ml heparin, 9.25 mM citric acid, and 0.1% Tween20), hybridization occurred at 65°C overnight with 1:200 probe diluted in hybridization mix, serial washed from hybridization to 2x SSC (66%, 33%, 2x SSC), washed with 0.2X SCC then PBTw, blocked with 2% goat serum (GS) and 20mg/ml bovine serum albumin (BSA), incubated overnight at 4°C in α-digoxygenin-AP Fab fragments (α-DIG-AP; 1:10,000; Millipore Sigma) to detect the probes, followed by washes in PBTw, washed in coloration buffer (100 mM Tris-HCl pH 9.5, 50 mM MgCl2, 100 mM NaCl, 0.1% Tween-20), and incubated in coloration buffer with coloration substrate (4.5mg/ml nitro blue tetrazolium chloride and 3.5mg/ml 5-bromo-4-chloro-3-indolyl-phosphate, toluidine-salt, Millipore Sigma). Following the coloration incubation, the samples were cleared using methanol and stored in 50% glycerol/PBS for imaging. The probes used in this study were: foxg1a, eya1, and six1b. Using a PCR-based protocol, the cDNA of interest is amplified by PCR with a reverse primer containing a T7 RNA binding sequence (CCAAGCTTCTAATACGACTCACTATAGGGAGA). Following PCR amplification of probe template, antisense RNA probes were generated with T7 RNA polymerase and digoxygenin RNA labeling mix.

2.6. HCR fluorescent RNA in situ hybridization

Hybridization chain reaction fluorescent in situ hybridization (FISH) was performed according to the manufacturer’s zebrafish-specific protocol (Molecular Instruments). Embryos were fixed in 4% PFA followed by storage in methanol at −20°C until processing. Samples were then serially washed out of methanol using PBTw, permeabilized using 10μM/ml Proteinase K in PBS, post-fixing in 4% PFA for 15 minutes, and washed in PBTw. Samples were prehybridized at 37°C for 30 minutes using hybridization buffer (Molecular Instruments) and hybridized at 37°C overnight in 2pmol probe in 500μl hybridization buffer. At room temperature, samples were washed 5x SSC. Separately, 30pmols of fluorescent hairpins (Molecular Instruments) 1 and 2 were prepared in the dark by heating to 95°C for 1 minute and cooled to room temperature for 30 minutes before adding both to 500μl amplification buffer. Samples were incubated overnight at room temperature in the hairpin/amplification buffer, then washed with 5x SSC, labeled with DAPI, and mounted in Fluorescent Mounting Media (EMD Millipore) to prevent fading. Probes used in this study were foxg1a, sox10, foxd3 trpc2b, ompb, sox2, and sox3.

2.7. Image Acquisition

For imaging of RNA in situ hybridization and immunohistochemistry samples, fixed and processed embryos and larvae were decapitated and the head placed with 50% glycerol/PBS either dorsal side or rostral side up toward the coverslip within a well manufactured from placement of a square of vacuum grease to angle olfactory epithelium dorsally. Images for embryos processed for hybridization chain reaction fluorescent in situ hybridization were obtained within 3 days to prevent signal loss. Samples used for intensity measurements were taken at the same laser settings between heterozygous and foxg1a^a266 homozygous clutchmates. Images were collected using a Zeiss 510 meta confocal microscope using Zen 2009 software, a Nikon-AXR resonant scanning confocal microscope using NIS-Elements software, or a Zeiss Imager D.2 compound microscope with Zen Pro 2.3 software.

2.8. Image Analysis

Images were processed and quantified using FIJI software (Schindelin et al., 2012). Olfactory epithelia were defined by the position of DAPI-positive cells in the OE. Cell numbers were manually collected using the FIJI plugin “Cell Counter”. Briefly, images were opened to specific channels and cells of the olfactory epithelium were counted by hand for every 7–8 images per 1um slices to capture discrete cells throughout the sample. Fluorescent quantification occurred on embryos imaged at the same fluorescent intensity. Average intensity was measured by drawing an ROI on DAPI-positive z-stacks and measuring intensity of z-stacks on relevant channels. Images had their brightness and contrast adjusted with Adobe Photoshop to enhance visual differences for publication and no images were altered for analysis.

2.8. Statistical analysis

All experimental data are expressed as mean +/− standard deviation. The data were analyzed using GraphPad Prism software (GraphPad Software, La Jolla, CA, USA) and data were subjected to an F-test for variance as well as tests for normality (Anderson-Darling, D’Agostino-Pearson, Shapiro-Wilk, and Kolmogorov-Smirnov). If the data had equal variance and passed normality, an Unpaired t-test was used. If only variances were significantly different, a Welch’s test was used. If the data failed any of the normality tests, a Mann-Whitney non-parametric tests was used. The probability level used for statistical significance was p < 0.05.

3. Results

3.1. Sox10-positive olfactory epithelial cell populations foxg1a^a266 zebrafish

While Foxg1 research has primarily investigated the developing forebrain, its role in craniofacial development has also been studied, showing changes in cellular identity of jaw cartilage and bone in mouse models (Compagnucci and Depew, 2020). Additionally, it is known that signaling in the forebrain influences craniofacial development (Marcucio et al., 2005). Using an established zebrafish foxg1a mutant line (foxg1a^a266) (Thyme et al., 2019) we evaluated the formation of early jaw structures. At 5 days post fertilization (dpf) in contrast to heterozygous sibling controls (Figure 1A–A”, C–C”), foxg1a^a266 mutant larvae (Fig 1B–B”) develop longer Meckle’s cartilage, which are displaced dorsally, and significantly narrow jaws, In part, the developmental defects seen in the jaws of foxg1a^a266 mutants, may be downstream of forebrain formation, which is significantly reduced as compared to heterozygous siblings (Danesin and Houart, 2012; Hou et al., 2020); (Figure 1A”, B”, and C”’, yellow dotted line). Interestingly, when using the Tg(sox10:eGFP) transgenic line (sox10:GFP) to mark jaw cartilage, we additionally found that sox10 expression is dramatically reduced in foxg1a^a266 embryos olfactory epithelia (1A”, B”, and C””, purple arrows) despite expression within cartilage. Decreased sox10 expression has been suggested in the olfactory placode of foxg1 mutant mice (Compagnucci and Depew, 2020; Duggan et al., 2008), though precise role of Foxg1 in regulating Sox10 during olfactory system development has not been described.

Figure 1.

Confocal projections of heads labeled with α-Elavl3 (magenta) in Tg(sox10:eGFP) embryos (green) and DAPI (grey) at 5dpf. Lateral view of heterozygous sibling (A) and foxg1a^a266 (B) heads, Meckle’s cartilage angle denoted by magenta dashed line and view of heterozygous sibling (A) and foxg1a^a266 (B) heads, Meckle’s length by green dashed lines. Ventral view of heterozygous sibling (A’) and foxg1a^a266 (B’) heads, Meckle’s cartilage width denoted by orange dashed line. Dorsal view of heterozygous sibling (A”) and foxg1a^a266 (B”) heads, olfactory epithelium (oe) indicated by purple arrows and forebrain (fb) by yellow dashed line. (C) Quantification of Meckel’s cartilage length in μm, (C’) jaw angle, (C”) Meckel’s cartilage length in μm, forebrain area in μm², and (C””) oe sox10:GFP fluorescence intensity in arbitrary units (A.U.). n=7–10 embryos per condition. Statistical analysis: Unpaired t-test for C, C’, C”’; Welch’s t-test for C””; Mann-Whitney test for C”. Scale bar=100μm. DIC images of foxg1a RNA in situ hybridization in wild-type embryos at 28hpf (D), 48hpf (E), 72hpf (F), and 96hpf (G) in the oe (magenta arrowheads) and fb. n=7–8 embryos per condition Scale bar= 20μm. Confocal projections of foxg1a HCR in situ hybridization (magenta) in Tg(sox10:eGPF) embryos (green), labeled with DAPI (gray) at 28hpf (H), 48hpf (I), 72hpf (J), and 96hpf (K). Insets show regions of overlap between foxg1a expression and sox10:GFP (blue arrowheads). n=7–8 embryos per condition Scale bar= 20μm.

To define the expression patterns of foxg1a and sox10:GFP-labeling in the developing OE, we performed whole mount RNA in situ hybridization (WISH) for foxg1a. In wild-type embryos, foxg1a is expressed in the OE from 28 hours post fertilization (hpf) with progressive restriction to a subset of cells at 96hpf (Fig. 1D–G). Expression of foxg1a occurred in expected areas where it influences development, including the forebrain and eye (Umeda et al., 2024). Further, using hybridization chain reaction (HCR) fluorescent RNA in situ hybridization (FISH) for foxg1a, we can see overlap of foxg1a expression with sox10:GFP-expressing cells between 28–96hpf (Fig. 1H–K). Next, we evaluated the expression of sox10 using HCR FISH between heterozygous siblings and foxg1a^a266 embryos from 1dpf through 4dpf. Overall, heterozygous siblings (Fig. 2A–A”, C, D–D”, F, G–G”, I, J–J” and L) express higher levels of fluorescence for sox10 in the OE compared to foxg1a^a266 embryos (Fig. 2B–B”, E–E”, H–H”, and K–K”); the expression patterns are cellular and overlap with sox10:GFP in heterozygous siblings (yellow arrowheads) as compared to diffuse in foxg1a^a266 mutants.

Figure 2.

Confocal projections of sox10 HCR in situ hybridization (magenta), Tg(sox10:eGFP) (green), and DAPI labeling (gray) in the oe (white dashed lines). At 1dpf in heterozygous sibling (A-A”’) and foxg1a^a266 (B-B”’) oe, insets (yellow dashed lines) show regions of overlap between sox10 expression and sox10:GFP (yellow arrowheads). (C) Quantification of sox10 fluorescence intensity at 1dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. At 2dpf in heterozygous sibling (D-D”’) and foxg1a^a266 (E-E”’) oe, insets (yellow dashed lines) show regions of overlap between sox10 expression and sox10:GFP (yellow arrowheads). (F) Quantification of sox10 fluorescence intensity at 2dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. At 3dpf in heterozygous sibling (G-G”’) and foxg1a^a266 (H-H”’) oe, insets (yellow dashed lines) show regions of overlap between sox10 expression and sox10:GFP (yellow arrowheads). (I) Quantification of sox10 fluorescence intensity at 3dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. At 4dpf in heterozygous sibling (J-J”’) and foxg1a^a266 (K-K”’) oe, insets (yellow dashed lines) show regions of overlap between sox10 expression and sox10:GFP (yellow arrowheads). (L) Quantification of sox10 fluorescence intensity at 4dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. n=7–10 embryos per condition. Statistical analysis: Unpaired t-test for F, I; Welch’s t-test for C; Mann-Whitney for L. Scale bar=20μm. Olfactory epithelium (oe).

Live lineage tracing of cell migration identified the preplacode ectoderm as the progenitor pool for sox10:GFP-positive cells in the OE (Aguillon et al., 2018). When we used live imaging to investigate sox10:GFP-expressing cells in heterozygous sibling and foxg1a^a266 mutant embryo OE, we find sox10:GFP-positive presumptive neural crest cells streaming over the anterior most aspect of the developing head (Movie 1,2; Fig 3A–B”’, cyan arrows). It has been noted that early placode cells do not mix with sox10:GFP-positive cranial neural crest cells (Harden et al., 2012). In our zebrafish, between 20 and 37hpf in heterozygous sibling embryos, sox10:GFP-postive cells are seen to arise in the forebrain area and in the OE (labeled in magenta with Tg(prim: lyn2-mCherry) for placodal cells) (Movie 1, Fig. 3A–A”’); these GFP-positive cells are largely absent in foxg1a^a266 mutant embryos (Movie 2, Fig. 3B–B”’). When we assessed placode markers using WISH, we found no gross difference expression patterns between heterozygous and foxg1a^a266 embryos for either eya1 (Fig. 3C–C’) or six1b (Fig. 3D–D’). This suggests that the effect of Foxg1 on the OE occurs after the initial olfactory placode formation. Further, when we assessed another cranial neural crest marker foxD3 to identify overlap with our sox10:GFP transgene expression, we identified overlap of HCR foxd3 expression with GFP in presumptive neural crest cells surrounding the OE (Fig. 1E–F’, yellow arrows) in both heterozygous siblings (Fig. 1E–E’) and foxg1a^a266 mutants (Fig. 1F–F’). In heterozygous siblings (Fig. 1E–E’), while foxd3 expression converges with GFP expression outside the OE, there does not appear to be any co-expression of with sox10:GFP in the OE. These data further suggest that early sox10 expression in the OE is likely not neural crest derived. Foxg1 thus appears to control these placode-derived Sox10-positive cells after initial formation of olfactory placodes.

Video 1.

Download video file ^{(1MB, avi)}

Timelapse imaging of Tg(sox10:eGFP) expression in Tg(prim:lyn2mCherry) placodes within heterozygous sibling zebrafish from 20–37hpf. Tg(sox10:eGFP) labeled cells appear and migrate across the embryonic head and begin to appear in the olfactory epithelium (magenta) at approximately minute 375. Scale bar=20μm

Video 2.

Download video file ^{(931.2KB, avi)}

Timelapse imaging of Tg(sox10:eGFP) expression in Tg(prim:lyn2mCherry) placodes within foxg1a^a266 zebrafish from 20–37hpf. Tg(sox10:eGFP) labeled cells appear and migrate across the embryonic head and do not populate the olfactory epithelium (magenta) or forebrain. Scale bar=20μm.

Figure 3.

Still confocal projections from time lapse images of Tg(sox10:eGFP) (green) and Tg(prim:lyn2mCherry) (magenta) from 20–37hpf in heterozygous sibling (A-A”’) and foxg1a^a266 (B-B”’) embryos. Yellow arrows indicate oe sox10:GFP+ cells, blue arrowheads indicate sox10:GFP+ cells in the forebrain, and orange arrowheads indicate sox10:GFP+ presumptive neural crest cells. n=10 heterozygous sibling and 8 foxg1^a266 embryos. DIC images of eya1 RNA in situ hybridization at 24hpf in heterozygous sibling (C) and foxg1^a266 (C’) embryos. n=13 heterozygous sibling and 12 foxg1^a266 embryos. DIC images of six1b RNA in situ hybridization at 24hpf in heterozygous sibling (D) and foxg1^a266 (D’) embryos. n=53 heterozygous sibling and 11 foxg1^a266 embryos. Confocal projections of foxd3 HCR in situ hybridization (magenta), sox10:GFP (green) and DAPI labeling (gray) at 28hpf in in heterozygous sibling (E-E’) and foxg1^a266 (F-F’) embryos, yellow arrowheads indicate foxd3 expression in sox10:GFP+ neural crest cells. White dashes lines indicate the oe. n=8 heterozygous sibling and 7 foxg1^a266 embryos. Scale bar=20μm. Olfactory epithelium (oe).

3.2. Neuronal populations in foxg1a^a266 embryos

Foxg1 has a known role in regulating olfactory sensory neuronal (OSN) differentiation in the OE (Garaffo et al., 2015; Kawauchi et al., 2009).To determine if the defects in the foxg1a^a266 mutants are specific to the development of sox10:GFP-positive OSNs in the OE we fixed embryos at 1–4dpf and labeled them with the pan neuronal marker α-Elavl3 antibody. Overall, we find no difference in OE cell numbers between heterozygous siblings (Fig. 4A–A’, C, D–D’, F, G–G’, I, J–J’, and L) and foxg1a^a266 mutants (Fig. 4B–B’, E–E’, H–H’, K–K’) using DAPI-positive nuclei to assess cell number. Compared to heterozygous siblings (Fig. 4A”, D”, G”, and J”), in the OE of foxg1a^a266 embryos, there is a significant near absence of sox10:GFP-positive cells (Fig. 4B”, C’, E”, F’, H”, I’, K”, and L’). When evaluating olfactory epithelium neurons, compared to heterozygous controls (Figure 4A”’, D”’, G”’, and J”’), we found a significant decrease in the population of α-Elavl3-positve cells at all timepoints in foxg1a^a266 embryos (Fig. 4B”’, C”, E”’, F”, H”’, I”, K”’, and L”), suggesting that Foxg1a function is required for the development of a subset of OSNs.

Figure 4.

Confocal projections of oe (dashed white lines) labeled with DAPI (gray), α-Elavl3 antibody (magenta), and Tg(sox10:eGFP) (green). α-Elavl3 and sox10:GFP label cells in the oe at1dpf in heterozygous sibling (A-A”’) and foxg1a^a266 (B-B”’) embryos. Quantification of DAPI+ cells (C), sox10:GFP+ cells (C’), and α-Elavl3+ cells (C”) at 1dpf in heterozygous sibling and foxg1^a266 oe. α-Elavl3 and sox10:GFP label cells in the oe at 2dpf in heterozygous sibling (D-D”) and foxg1a^a266 (E-E”’) embryos. Quantification of DAPI+ cells (F), sox10:GFP+ cells (F’), and α-Elavl3+ cells (F”) at 2dpf in heterozygous sibling and foxg1^a266 oe. α-Elavl3 and sox10:GFP label cells in the oe at 3dpf in heterozygous sibling (G-G”) and foxg1a^a266 (H-H”’) embryos. Quantification of DAPI+ cells (I), sox10:GFP+ cells (I’), and α-Elavl3+ cells (I”) at 3dpf in heterozygous sibling and foxg1^a266 oe. α-Elavl3 and sox10:GFP label cells in the oe at 4dpf in heterozygous sibling (J-J”) and foxg1a^a266 (K-K”’) embryos. Quantification of DAPI+ cells (L), sox10:GFP+ cells (L’), and α-Elavl3+ cells (L”) at 4dpf in heterozygous sibling and foxg1^a266 oe. n=7–10 embryos per condition. Statistical analysis: Unpaired t-tests for C, C’, F, F’, I, I’, L’; Welch’s t-tests for C”, F”, I”, L, L”’. Scale bar=20μm. Olfactory epithelium (oe).

Previous research established that Sox10 is required for microvillous OSN development in zebrafish (Saxena et al., 2013). We next sought to determine if microvillous cells are broadly lost within the developing OE of foxg1a^a266 embryos. Using HCR FISH to assess the expression of the microvillous OSN marker trpc2b from 2–4dpf, we found that while heterozygous siblings exhibit cellular expression that largely colocalizes with Tg(sox10:eGFP) cells (Fig. 5A–A”’, D–D”’, and G–G”’), in foxg1a^a266 zebrafish (Fig. 5B–B”’, E–E”’, H–H”’), there is no discernable trpc2b expression. Analysis of fluorescent intensity shows a significant decrease for foxg1a^a266 embryos compared to heterozygous siblings(Fig. 5C, F, & I). Further, when assessing colocalization of trpc2b with foxg1a at 4dpf, we found that all trpc2b labeled cells also express foxg1a (Fig. 5J–J””, arrows—white indicating without sox10:GFP co-labeling, and yellow arrows co-labeled with sox10:GFP, while only 36% of foxg1a-expressing cells are also trpc2b-positive). In agreement with previously reported data (Thyme et al., 2019), there is diffuse expression of fox1a mRNA in the foxg1a^a266 mutants (Fig. 5K–K”), while expression of trpc2b and sox10:GFP remains absent (Fig. 5K”’–K””). Our data for co-expression of trpc2b and foxg1a in heterozygous siblings at 4dpf suggests a possible broader role for Foxg1 beyond regulating the development of sox10:GFP-positive microvillous cells. Early research suggested a decrease in microvillous OSNs for morpholino-induced knockdown of foxg1a using an antibody for Calretinin (Duggan et al., 2008), but more recent research suggests that α-Calretinin labels both microvillous and ciliated OSNs (Kress et al., 2015). Therefore, we investigated Calretinin to assess a broader impact of Foxg1a loss of function. Between 2–4dpf, heterozygous sibling embryos develop an increasing number of α-Calretinin-labeled and sox10:GFP-positive cells (Figure 6A–A’, D–D’, and G–G’). We found that foxg1a^a266 embryos (Fig. 6B–B’, E–E’, and H–H’) have fewer calretinin positive cells (Fig. 7C, F, and I). A subset of α-Calretinin-positive cells are co-labeled with Tg(sox10:eGFP) in control embryos which are missing in foxg1a^a266 embryos (Fig. 6C”, F”, and I”). Together, this data suggests that Foxg1a is required for the formation of a subset of OSNs, specifically microvillous cells. However, this data does not preclude other OSN cell types.

Figure 5.

Confocal projections of oe (dashed white lines) showing trpc2b HCR in situ hybridization (magenta), Tg(sox10:eGFP) (green), and DAPI (gray). At 2dpf in heterozygous sibling (A-A”’) and foxg1a^a266 (B-B”’) oe, insets (yellow dashed lines) show regions of overlap between sox10:GFP-labeling and trpc2b expression (yellow arrowheads). (C) Quantification of trpc2b fluorescence intensity at 2dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. At 3dpf in heterozygous sibling (D-D”’) and foxg1a^a266 (E-E”’) oe, insets (yellow dashed lines) show regions of overlap between sox10:GFP-labeling and trpc2b expression (yellow arrowheads). (F) Quantification of trpc2b fluorescence intensity at 3dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. At 4dpf in heterozygous sibling (G-G”’) and foxg1a^a266 (H-H”’) oe, insets (yellow dashed lines) show regions of overlap between sox10:GFP-labeling and trpc2b expression (yellow arrowheads). (I) Quantification of trpc2b fluorescence intensity at 4dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. n=7–10 embryos per condition. Statistical analysis: Unpaired t-test for F; Welch’s t-test for C, I. Confocal projections of oe (dashed white lines) showing foxg1a HCR in situ hybridization (cyan), trpc2b HCR in situ hybridization (magenta), and Tg(sox10:eGFP) (green) at 4dpf in heterozygous sibling (J-J””) and foxg1a^a266 (K-K””) larvae. Insets (yellow dashed lines) show regions of overlap between foxg1a, trpc2, and sox10:GFP (yellow arrowheads) or foxg1a and trpc2 (white arrowheads). n=7 heterozygous sibling and 8 foxg1^a266 embryos. Scale bar=20μm. Olfactory epithelium (oe).

Figure 6.

Confocal projections of α-Calretinin antibody labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (dashed white lines). α-Calretinin and sox10:GFP label cells in the oe at 2dpf in heterozygous sibling (A-A”) and foxg1a^a266 (B-B”) embryos. Quantification of α-Calretinin+ cells (C) and α-Calretinin /sox10:GFP+ cells (C’) at 2dpf in heterozygous sibling and foxg1^a266 oe. α-Calretinin and sox10:GFP label cells in the oe at 3dpf in heterozygous sibling (D-D”) and foxg1a^a266 (E-E”) embryos. Quantification of α-Calretinin+ cells (F) and α-Calretinin /sox10:GFP+ cells (F’) at 3dpf in heterozygous sibling and foxg1^a266 oe. α-Calretinin and sox10:GFP label cells in the oe at 4dpf in heterozygous sibling (G-G”) and foxg1a^a266 (H-H”) embryos. Quantification of α-Calretinin+ cells (I) and α-Calretinin /sox10:GFP+ cells (I’) at 2dpf in heterozygous sibling and foxg1^a266 oe. Statistical analysis: Welch’s t-test for F, I; Mann-Whitney for C, C’, F’, I’. Scale bar=20μm. Olfactory epithelium (oe).

Figure 7.

Confocal projections of oe (dashed white lines) showing ompb HCR in situ hybridization (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) at 4dpf in heterozygous sibling (A-A”’) and foxg1a^a266 (B-B”’) embryos. (C) Quantification of ompb fluorescence intensity at 4dpf in arbitrary units (A.U.) between heterozygous sibling and foxg1a^a266 oe. n=8 for all conditions. Confocal projects of α-Acetylated tubulin antibody labeling (yellow), phalloidin (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in oe (white dashed lines) at 4dpf in heterozygous sibling (D-D”’) and foxg1a^a266 (E-E”’) embryos. Insets (yellow dashed lines) show cilia (D’,E’; white brackets) and rod cells (D”-D”’, E”-E”’). Quantification of nasal cavity in μm² (F; blue dashed lines in D-E), cilium length in μm (G), phalloidin+ rods (H), and rod length in μm (I). n=8–10 embryos per condition. Statistical analysis: Unpaired t-test for C; Welch’s t-test for F; Mann-Whitney for G, H, I. Scale bar=20μm. Olfactory epithelium (oe).

The other major OSN cell type are ciliated cells, which express the marker omp (Sato et al., 2005). Knockdown of foxg1a may lead to the early loss of omp labeling (Duggan et al., 2008), however, this may be a nonspecific artifact of morpholino use. To determine if ciliated cells are influenced by Foxg1a in the OE, we performed HCR FISH using ompb at 4dpf. The fluorescent intensity did not differ between heterozygous sibling (Fig. 7A–A”, C) and foxg1a^a266 (Fig. 7B–B”) embryos. While not necessarily affecting the differentiation of ciliated cells, Foxg1 may affect morphological features such as cilia length or nasal cavity size, which can be changed when ciliated cells are impacted (Rayamajhi et al., 2024). Therefore, we labeled fish with α-Acetylated tubulin antibody and phalloidin to label actin. At 4dpf, heterozygous sibling and foxg1a^a266 embryos have no difference in nasal cavity size or average cilia length (Fig. 7D–D’, E–E’, F, and G), though the increase in the variance of nasal cavity area with foxg1a^a266 mutants is likely due to the cranial architecture differences from forebrain loss. Additionally, previous work identified a rare population of olfactory neurons which may act as mechanosensory cells, called rod cells, that are characterized by a single actin protrusion (Cheung et al., 2021); a subset of these cells co-express Tg(sox10:eGFP), though their development is not altered in sox10 mutant zebrafish (Cheung et al., 2021). The heterozygous controls contain phalloidin-positive rod cells that often co-label with Tg(sox10:eGFP) (Fig. 7D”–D”’, magenta arrows); these phalloidin-positive rod cells are significantly decreased in foxg1a^a266 embryos (Fig. 7E”–E”, H) but not entirely lost. The length of phalloidin-positive rods is also significantly shorter in foxg1a^a266 mutants compared to heterozygous siblings (Fig. 7I). This loss of a rare cell type that often co-labels with sox10:GFP further supports a role for Foxg1a upstream of Sox10 in the OE.

3.3. Cell population maintenance

We next asked if loss of Foxg1a function alters cell survival and proliferation in the OE, as Foxg1 is important for these cellular processes in other contexts. To investigate proliferation, we pulsed BrdU for 1 hour in embryos ranging from 1–4dpf. For all time points, BrdU incorporation was not significantly different between heterozygous sibling (Fig. 8A–A’, C, D–D’, F, G–G’, I, J–J’, and L) and foxg1a^a266 (Fig. 8B–B’, E–E’, H–H’, K–K’) embryos. A TUNEL assay was performed to identify cell death. At no time points identified from 1–4dpf did foxg1a^a266 embryos (Fig. 8N–N’, Q–Q’, T–T’, W–W’) differ in cell death compared to heterozygous siblings (Fig. 8M–M’, O, P–P’, R, S–S’, U, V–V’, and X). This suggests that the decrease in OSNs found in foxg1a^a266 OE is not primarily the result of cell death. Together these data suggest that the OE of foxg1a^a266 embryos might have a reduction of cell specification and differentiation rather than changes on overall cell number due to altered proliferation or survival.

Figure 8.

Confocal projects of α-BrdU antibody labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (white dashed lines). BrdU and sox10:GFP labeling at 1dpf in heterozygous sibling (A-A’) and foxg1a^a266 (B-B’) oe. Quantification of BrdU+ at 1dpf (C). BrdU and sox10:GFP labeling at 2dpf in heterozygous sibling (D-D’) and foxg1a^a266 (E-E’) oe. Quantification of BrdU+ at 2dpf (F). BrdU and sox10:GFP labeling at 3dpf in heterozygous sibling (G-G’) and foxg1a^a266 (H-H’) oe. Quantification of BrdU+ at 3dpf (I). BrdU and sox10:GFP labeling at 4dpf in heterozygous sibling (J-J’) and foxg1a^a266 (K-K’) oe. Quantification of BrdU+ at 4dpf (L). Confocal projects of TUNEL labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (white dashed lines). TUNEL and sox10:GFP labeling at 1dpf in heterozygous sibling (M-M’) and foxg1a^a266 (N-N’) oe. Quantification of TUNEL+ at 1dpf (O). TUNEL and sox10:GFP labeling at 2dpf in heterozygous sibling (P-P’) and foxg1a^a266 Q-Q’) oe. Quantification of TUNEL+ at 2dpf (R). TUNEL and sox10:GFP labeling at 3dpf in heterozygous sibling (S-S’) and foxg1a^a266 (T-T’) oe. Quantification of TUNEL+ at 3dpf (U). TUNEL and sox10:GFP labeling at 4dpf in heterozygous sibling (V-V’) and foxg1a^a266 (W-W’) oe. Quantification of TUNEL+ at 1dpf (X).). n=7–10 embryos per condition. Statistical analysis: Unpaired t-test for all except X; Mann-Whitney for X. Scale bar=20μm. Olfactory epithelium (oe).

Foxg1 is known to maintain progenitor populations in the telencephalon and disruption of Foxg1 leads to inappropriate specification of ventral tissues (Martynoga et al., 2005). Further, Foxg1 in the OE supports the early specification of ventrolateral OE tissues (Kuriyama and Hanashima, 2025). To better understand how non-neuronal OE cells may be affected in foxg1a^a266 embryos, we analyzed two early progenitor cell markers: Sox2 and Sox3. While Sox3 labeling is not well characterized within the OE, research suggests that both Sox2 and Sox3 are downstream of Eya1 and Six1 in placodal tissues (Schlosser et al., 2008). In other placodal tissue such as the otic placode, Sox2 and Sox3 have non-redundant roles in sensory and neuronal cell lineages (Gou et al., 2018). In line with this literature, we found differential changes within these stem cell markers; in contrast to their heterozygous siblings between 1–4dpf (Fig. 8A–A’, C, D–D’, F, G–G’, I, J–J’, and L), α-Sox2 antibody labeling was significantly increased in foxg1a^a266 embryos (Fig. 9 B–B’, E–E’, H–H’, K–K’)). However, compared to their heterozygous siblings (Fig. 9N–N’, P, Q–Q’, S, T–T’, V, W–W’, and Y), foxg1a^a266 embryos displayed a significant decrease in α-Sox3 antibody-labeled cells between 1–4dpf (Fig. 9O–O’, R–R’, U–U’, X–X’). To determine if there is overlap of stem cells markers and sox10:GFP during development of the OE, we examined wild-type embryos at 1dpf using HCR FISH, we found only a subset of sox10:GFP+ cells had sox2+ puncta (11.4 %± 5.7; Fig. 9M–M”, white arrows), while all sox10:GFP+ cells had some puncta of sox3 (100.0% ± 0.0; Fig. 9Z–Z”, yellow arrows). The overlap of sox3 and sox10:GFP expression in the developing OE, as well as the decrease of a-Sox3+ cells in foxg1a^a266 mutants suggests that development of this population is sensitive to Foxg1a function. Previous research has suggested that the lack of Foxg1, particularly in the telencephalon, depletes the progenitor pool and leads to a decrease in neural cells (Xuan et al., 1995). Our data shows that Foxg1 impacts two different progenitor pools in a differential manner within the OE. This indicates a nuanced role for Foxg1a in specification of cell types for olfactory tissues.

Figure 9.

Confocal projects of α-Sox2 antibody labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (white dashed lines). α-Sox2 and sox10:GFP labeling at 1dpf in heterozygous sibling (A-A’) and foxg1a^a266 (B-B’) oe. Quantification of α-Sox2+ at 1dpf (C). α-Sox2 and sox10:GFP labeling at 2dpf in heterozygous sibling (D-D’) and foxg1a^a266 (E-E’) oe. Quantification of α-Sox2+ at 2dpf (F). α-Sox2 and sox10:GFP labeling at 3dpf in heterozygous sibling (G-G’) and foxg1a^a266 (H-H’) oe. Quantification of α-Sox2+ at 3dpf (I). α-Sox2 and sox10:GFP labeling at 4dpf in heterozygous sibling (J-J’) and foxg1a^a266 (K-K’) oe. Quantification of α-Sox2+ at 4dpf (L). Confocal projections of sox2 HCR in situ hybridization (cyan), foxg1a HCR in situ hybridization (magenta), Tg(sox10:eGFP), and DAPI (gray) at 1dpf in wild-type oe (M-M”; white dashed lines). Insets (yellow dasked lines) show overlap of foxg1a and sox10:GFP expression (white arrowheads). n=8 embryos. Confocal projects of α-Sox3 antibody labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (white dashed lines). α-Sox3 and sox10:GFP labeling at 1dpf in heterozygous sibling (N-N’) and foxg1a^a266 (O-O’) oe. Quantification of α-Sox3+ at 1dpf (P). α-Sox3 and sox10:GFP labeling at 2dpf in heterozygous sibling (Q-Q’) and foxg1a^a266 (R-R’) oe. Quantification of α-Sox3+ at 2dpf (S). α-Sox3 and sox10:GFP labeling at 3dpf in heterozygous sibling (T-T’) and foxg1a^a266 (U-U’) oe. Quantification of α-Sox3+ at 3dpf (V). α-Sox3 and sox10:GFP labeling at 4dpf in heterozygous sibling (W-W’) and foxg1a^a266 (X-X’) oe. Quantification of α-Sox3+ at 4dpf (Y). Confocal projections of sox3 HCR in situ hybridization (cyan), foxg1a HCR in situ hybridization (magenta), Tg(sox10:eGFP), and DAPI (gray) at 1dpf in wild-type oe (Z-Z”; white dashed lines). Insets (yellow dasked lines) show overlap of foxg1a, sox3 and sox10:GFP expression (yellow arrowheads). n=8 embryos. n=8 embryos. n=7–9 embryos per condition. Statistical analysis: Unpaired t-test for all except C, Y; Mann-Whitney for C, Y. Scale bar=20μm. Olfactory epithelium (oe).

Another progenitor cell marker that plays an important role in the formation of the OE is Pax6 (Collinson et al., 2003). Classically, Pax6-positive cells have been described in the OE as non-neuronal (Davis & Reed, 1996). Pax6 often co-labels with Sox2 and plays a role in maintaining progenitor status (Guo et al., 2010). At 4dpf, we found α-Pax6 antibody labeling in heterozygous siblings (Fig. 10A–A’, C) and a significant decrease in Pax6-positive cells in the OE for foxg1a^a266 embryos (Fig. 10B–B’), in a pattern more closely resembling our Sox3 results. Notably, there are remaining Pax6 positive cells within the forebrain heterozygous and foxg1a^a266 embryos (yellow arrowheads). Pax6 plays a pleiotropic role where a dosage decrease of the protein within the brain has effects of regionalization and development of neural tissue (Grindley et al., 1997; Kozmik & Kozmikova, 2024).

Figure 10.

Confocal projects of α-Calretinin antibody labeling (magenta), α-Pax6 antibody labeling (cyan), Tg(sox10:eGFP) (green), and DAPI (gray) in the oe (yellow arrow heads) at 2dpf in in heterozygous sibling (A-A’) and foxg1a^a266 (B-B’) embryos, yellow arrows indicate α-Pax6 + fb cells. Quantification of α-Pax6+ cells in heterozygous sibling and foxg1a^a266 oe (C). Confocal projects of α-Islet1 antibody labeling (magenta), Tg(sox10:eGFP) (green), and DAPI (gray) at 2dpf in in heterozygous sibling (D-D”) and foxg1a^a266 (E-E”) embryos. Insets (white dashed lines) show α-Islet1+ cells in the oe (yellow arrowheads) and the fb (blue arrows) Quantification of α-Islet1+ cells in heterozygous sibling and foxg1a^a266 oe (F) and the percentage of α-Islet1+/sox10:GFP+ cells in the fb. Confocal projects of α-Calretinin antibody labeling (magenta), α-Acetylated tubulin antibody labeling (cyan), Tg(sox10:eGFP) (green), and DAPI (gray) in the fb at 2dpf in in heterozygous sibling (G-G”’) and foxg1a^a266 (H-H”’) embryos. Insets (dashed yellow lines) show axons in the presumptive glomeruli, yellow arrows indicate overlap of α-Acetylated tubulin, α-Calretinin, and sox10:GFP. Quantification of α-Calretinin area in μm² (I) and percent over lap of sox10:GFP/α-Calretinin+ projections (I’) in heterozygous sibling and foxg1a^a266 fb. n=7–10 embryos per condition. Statistical analysis: Unpaired t-test for C; Mann-Whitney for F, F’, I, I’. Scale bar=20μm. Olfactory epithelium (oe) and forebrain (fb).

Heterogenous populations of neurons contribute to the OE, including Islet-1-positive cells (Siddiqi et al., 2021) and Calretinin-positive cells (Kohwi et al., 2007; Kress et al., 2015). Islet-1 cells originating from the olfactory placode form migratory neurons including gonadotropin releasing hormone 1 (GnRH1) positive neurons in mice (Taroc et al., 2020). From work in our lab, we know that foxg1a plays a role in regulating Islet-1 cells in the posterior lateral line mechanosensory system (Bell et al., 2024). Therefore, we investigated α-Islet-1 antibody labeling in the OE. For heterozygous siblings at 2dpf, there was a subset of α-Islet-1-positive cells medial to the OE (Fig. 10D–D”, yellow arrows) while foxg1a^a266 embryos have a decreased α-Islet-1-positive cell OE population (Fig. 10E–E”, yellow arrows, F). Islet-1-positive GnRH1 (GnRH3 in zebrafish) neurons migrate using vomeronasal axons to the forebrain (Wray, 2010) and GnRH neurons rely on axon guidance cues for maturation during development (van Battum et al., 2025). Within the forebrain (fb), Islet-1-positve cells co-label with sox10:GFP-positive cells (Fig. 10D, blue arrows); these cells are absent in foxg1a^a266 embryos compared to their heterozygous siblings (Fig. 10F). This led us to ask if neural projections are affected in foxg1a^a266 embryos.

Within the OE, it is known that calcium binding proteins expression extends into axonal projections (Kress et al., 2015). We thus investigated the presumptive glomeruli for α-Calretinin-positive cells alongside sox10:GFP expression at 2dpf. α-Calretinin- and sox1:GFP-positive projections were robust within heterozygous siblings (Fig. 10G–G”’), while significantly reduced in foxg1a^a266 embryos (Fig. 10H–H”). α-Calretinin presumptive glomeruli were decreased or missing within foxg1a^a266 embryos (Fig. 10H”) with no Calretinin- and sox10:GFP- double positive projections (Fig. 10H”’) compared to heterozygous siblings (Fig. 10G”–G”’, I, and I’). α-Acetylated tubulin antibody labeling was present in both heterozygous siblings (Fig. 10G–G’) and foxg1a^a266 (Fig. 10H–H’). These data suggest that some axon extension may occur in foxg1a^a266 mutants, but not specifically from α-Calretinin- and sox:GFP-labeled OSNs. This perhaps indicates specific changes in axon outgrowth or guidance for a subset of neural projections. Overall, the data within this study shows Foxg1 as a pleiotropic player within the OE, affecting progenitor cells, neuronal development, and neuronal projections.

4. Discussion

Foxg1 is a key factor regulating neural development (Hettige and Ernst, 2019; Hou et al., 2020; Kersigo et al., 2011), though there remain many unanswered questions about the role of Foxg1 in the developing embryo, including how Foxg1 regulates Sox10 within the forming olfactory epithelium. In this paper, we have confirmed that Fox1a regulates sox10 expression in the OE as suggested in previous literature (Compagnucci and Depew, 2020; Duggan et al., 2008) and we describe a loss of microvillous OSNs in the foxg1a^a266 mutants.

Sox10 has often been described in the context of cranial neural crest development (Pingault et al., 2022), though in the OE lineage tracing suggests that Tg(sox10:eGFP)-positive cells arise from pre-placodal ectoderm (Aguillon et al., 2018). Our data, using cranial neural crest marker foxd3, shows that there is no major expression in the OE and supports this claim. Placodal markers were not different between foxg1a^a266 embryos and their heterozygous siblings, suggesting that Foxg1 does not broadly influence olfactory placodal development and its role arises downstream of initial placode formation.

Foxg1 is a conserved factor that plays pleiotropic roles in neural tissue (Kumamoto and Hanashima, 2017). We would therefore expect neural markers to change when Foxg1 function is compromised. In this study, we show a decrease in number of cells labeled with neural marker Elavl3 within the OE of foxg1a^a266 embryos from 1dpf to 4dpf. We do not see the loss of the entire neurogenic niche, since we have a remaining population of Elavl3-positve cells. This decrease in Elavl3-positive cells then must affect a subset of the neurogenic population. These data indicate that Foxg1 is affecting the proper production of neurons, aligning with established literature (Duggan et al., 2008; Garaffo et al., 2015; Kawauchi et al., 2009), but suggests that Foxg1 may only impact some neuron types.

In this paper, we focus on the two major OSN types: ciliated and microvillous. Previous literature suggests that Tg(sox10:eGFP)-labeled cells produce microvillous OSNs (Aguillon et al., 2018; Saxena et al., 2013). Here, we show that Foxg1 acts upstream of trpc2b and sox10 to specify the development of microvillous cells and in the absence of Foxg1a function, these cells fail to form. However, Foxg1 may have a broader effect in the OE than the loss of sox10-positive and microvillous cells since it acts upstream of many neural factors in the developing brain (Cargnin et al., 2018; Vezzali et al., 2016; Weise et al., 2019). Further, only approximately 37% of foxg1a expression occurred with trpc2b expression in cells for 4dpf heterozygous siblings indicating that Foxg1 affects a broader set of cells than microvillous. Yet, when we investigated ompb expression, we found no difference in absence of Foxg1a function, suggesting cell type specific regulation in the developing OE. Genes that affect ciliated cells can affect broad morphology, such as altering the size of the nasal cavity as found following the loss of Foxj1 in the developing zebrafish OE (Rayamajhi et al., 2024). However, we find no change in olfactory cell cilia length nor nasal cavity size in foxg1a^a266 mutants as compared to control embryos. The shape of the epithelial opening may have altered downstream of the loss of the forebrain structure, but this did not appear to majorly affect the area. These data lead us to conclude that foxg1 does not play a role in ciliated OSNs in the OE. Though we did find alterations in another OE cell type, the rod cells that are proposed to act as mechanosensory cells, a subset of which express Tg(sox10:eGFP) (Cheung et al., 2021). We find control embryos have an increased number and longer actin projections of rod cells at 2dpf than the foxg1a^a266 mutants. While Cheung et al. (2021) reports that only around 64% of rod cells express Tg(sox10:Lifeact-mRFPruby) and sox10 mutants did not have significant changes in rod cells in their study, our evidence suggests that Foxg1 is responsible for a large subset of rod cells that are likely also sox10:GFP-positive. Rod cells in foxg1a^a266 embryos are less abundant and their rod length is affected, which may indicate either a role within this remaining population of cells, or of surrounding cells that may affect rod cells. The loss of this large portion of rod cells that may be sox10:GFP-positive further reinforces the upstream effect Foxg1 has on Sox10 in the OE.

This loss of a subset of OE cells may be tied to the role of Foxg1 in regulating cell behaviors (Martynoga et al., 2005). We showed that there is no overall difference in cell death, proliferation, or total cell number in foxg1a^a266 embryos compared to controls. The lack of cell death agrees with a recent study showing that α-cleaved Caspase 3 antibody labeling is not changed in Foxg1 mutant mouse OE (Kuriyama and Hanashima, 2025). With the lack of difference in BrdU-labeled proliferation, we pursued other cell lineages that may be affected by Foxg1 that can explain the loss of neurons in our mutants without altering overall cell number. Previous research suggests that Foxg1 regulates neuronal differentiation through solidifying their cell fate, as seen in dorsal telencephalic neurons (Mall et al., 2017). The population of sox10:GFP-positive cells that are lost in the OE of foxg1a^a266 mutants may have altered cell fate and progenitor identity.

Resident progenitor cell pools in the OE allow tissue plasticity, homeostasis, and regeneration (Y.-Z. Wang et al., 2011). SoxB1 factors Sox2 and Sox3 are well known for their similar roles in maintaining neural precursor cell lineage (Sarkar and Hochedlinger, 2013). In the OE, Sox2 has been well studied for its roles in regulating development and regeneration (Kang et al., 2025; Mercurio, 2023) while a potential function for Sox3 remains relatively unexplored despite its expression in the OE. We assessed the effect of Foxg1 on Sox2 or Sox3 progenitor cells in the embryonic zebrafish OE. Interestingly, we see differential changes in these two cellular pools. Sox3 cells are significantly reduced in foxg1a^a266 embryos, while the Sox2 population of cells is increased. This finding opposes research that suggests Foxg1 acts proportionally to Sox2 (Duggan et al., 2008; Garaffo et al., 2015; L. Wang et al., 2018). However, a knockout of Foxg1 in mice exhibited OE expression of Sox2 within their knockout despite decreased OE size (Kuriyama and Hanashima, 2025). The decrease in Sox3 positive cells correlate with the decrease in neuronal markers, while the increase in Sox2 may be explained by the stalling of differentiation or by an increase in glial fate decision (Hoffmann et al., 2014; Mercurio et al., 2019). Our findings that there is also a decrease in Pax6-positive cells in foxg1a^a266 mutants corroborates this glial hypothesis, since Pax6 accelerates neuronal differentiation and Sox2 expands neuronal progenitors (Packard et al., 2016) and neuronal populations are decreased. It is unknown if the remaining population of Pax6 cells co-express Sox2 or are sustentacular cells(Guo et al., 2010). Our study demonstrates how Foxg1 may influence OE cells by its role in programing progenitor pool cell fate.

These progenitor cell populations have pleiotropic roles that affect tissue surrounding the OE. For instance, changes in Pax6 may impact neuronal migration (Zhang et al., 2025). Based on the pleiotropic effect that Foxg1 has on neural tissue (Wong et al., 2019), we investigated the role that Foxg1 played in migratory sensory cells and circuit formation for the OE. Work from our lab showed that foxg1a^a266 zebrafish have decreased Islet1-positive neuromast cells (Bell et al., 2024), providing a role for Foxg1 upstream of Islet-1 in the posterior lateral line. Derived from the olfactory placode, Islet1/2 contributes to GnRH migratory neuronal identity (Taroc et al., 2020). This population appears to be separate from sox10:GFP-positive cells (Aguillon et al., 2018). In this paper, we show a decrease in Islet1-positive cells on the medial side of the OE. These cells show similarities to Islet-positive cells that are GnRH-3 cells in zebrafish (Aguillon et al., 2018). Olfactory placode derived GnRH-3 cells migrate along axons to contribute to the hypothalamus and courtship behavior (Cui et al., 2025). A study mouse knocking down somatostatin in the OE decreased the number of GnRH and Islet-1 neurons in the OE and forebrain (Murakami et al., 2024), and it has been asserted that Foxg1 acts upstream of Somatostatin-expressing interneurons (Chen et al., 2019). Several of these factors influence neurite outgrowth including Islet-1 (Tanaka et al., 2011) and somatostatin (Ferriero et al., 1994). We also note a loss of Islet-1- and sox10:GFP-positive cells in the forebrain. The decrease in OE Islet1-positive cells provides evidence to broader consequences of Foxg1 in the OE than only sox10:GFP-positive cells. Finally, we identified neuronal projections from the OE from sox10:GFP- and Calretinin-positive cells. We know from previous research that Foxg1 is important for cortical-cortical projections (Cargnin et al., 2018), retinal projections (Pratt et al., 2004; Tian et al., 2008), and for OSNs (Garaffo et al., 2015). We found a loss of sox10:GFP-positive projections to the olfactory bulb, as well as a decrease of calretinin positive projections, implying failure of glomeruli formation in foxg1a^a266 embryos for cells influenced by Foxg1. Among the many roles that Foxg1 plays, we show that the proper formation of neural projections is among those affected in foxg1a^a266 mutant embryos. Since the loss Foxg1 is well known to negatively impact the brain development through differentiation as well as neuronal migration (Hanashima et al., 2004; Kumamoto et al., 2013; Lin et al., 2025), and improper function has caused hypoplastic olfactory bulbs in human infants (Jang et al., 2021), the disruption in glomeruli formation may be due to Foxg1’s role in the olfactory bulb rather than neuronal pathfinding. Further research into circuit formation in foxg1a^a266 zebrafish is warranted.

5. Conclusions

Foxg1 is an important transcription factor with pleiotropic effects that influence development of neural tissues. In this study, we show through analysis of the foxg1a^a266 mutant zebrafish line that Foxg1 is vital for the formation of microvillous OSNs. The loss of these cells is likely due to Foxg1’s role in regulating neural stem cell fate since we see a decrease in neural markers. Similarly, circuit formation for the subset of affected cells may be disrupted, with decreased neuronal projections from the OE and Islet1-positive OE cells. While roles for Foxg1 have previously been established in proliferation and cell cycle, we see no changes in proliferation or cell death, or overall cell number, using this model. Future work may be required to parse out effects on other heterogenous cell types in the zebrafish OE. While we see no change in ciliated OSNs, there may be subtle changes in other subclasses of OSNs or olfactory glial cells, which requires future investigation. Further, this change affects the progenitor cell pool for the OE by increasing the number of Sox2-expressing cells while decreasing the number of Sox3-expressing and Pax6-expressing cells. Future study is required to analyze how these progenitor populations affect different heterogenous cell populations in the olfactory epithelium. Our work shows a gap in knowledge where Foxg1 affects Sox10 positive microvillous OSNs through its role in progenitor pool regulation, downstream of placode formation. These findings show an important connection for nuanced Foxg1 function in the OE which may aid in better addressing complications in rare human diseases such as FOXG1 Syndrome.

Table 1.

Reagents and resources

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal α-Calretinin (clone 6B8.2)	Millipore Sigma	Cat# ZMS1073, RRID:AB_9425
Mouse monoclonal α-acetyl-alpha Tubulin	Millipore Sigma	Cat# MABT868, RRID:AB_2819178
Mouse monoclonal α-HuC/HuD (Elavl3) (16A11)	Thermo Fisher Scientific	Cat# A-21272, RRID:AB_2535822
Rabbit polyclonal α-Islet1	GeneTex	Cat# GTX128201, RRID:AB_2868422
Rabbit polyclonal α-Sox2	GeneTex	Cat# GTX124477, RRID:AB_1117806
Rabbit polyclonal α-Sox3	GeneTex	Cat# GTX129235, RRID:AB_2885934
Rabbit monoclonal α-BrdU	Thermo Fisher Scientific	Cat# MA5-46930, RRID:AB_2938002
Goat α-mouse Alexa-568	Thermo Fisher Scientific	Cat. # A-11004; RRID:AB_2534072
Goat α-mouse Alexa fluor-647	Thermo Fisher Scientific	Cat. # A-21236; RRID:AB_2535805
Goat α-rabbit Alexa fluor-647	Thermo Fisher Scientific	Cat. # A-21244; RRID:AB_2535812
α-digoxigenin-AP, fab fragments	Millipore Sigma	Cat. # 11093274910; RRID:AB_2734716
Chemicals, Peptides, and Recombinant Proteins
DAPI	Thermo Fisher Scientific	D1306
Phalloidin Alexa Fluor-568	Thermo Fisher Scientific	A12380
BrdU (5-Bromo-2’-deoxyuridine)	Millipore Sigma	19–160
Click-iT Plus TUNEL Assay Kit	Thermo Fisher Scientific	C10618
Experimental Models: Organisms/Strains
Zebrafish: Wildtype *AB	ZIRC http://zebrafish.org	ZFIN:ZDB-GENO-960809-7
Zebrafish: foxg1aa266	(Thyme et al., 2019)	RRID: ZDB-FISH-200507-6
Zebrafish: Tg(sox10:EGFP)ba2	(Dutton et al., 2009)	ZFIN:ZDB-TGCONSTRCT-090819-1
Zebrafish: Tg(prim:lyn2-mCherry)	(J. Wang et al., 2018)	ZFIN:ZDB-TGCONSTRCT-190412-1
Oligonucleotides
foxg1a in situ hybridization probe template primers  Forward:5’TGCTGGTGGTTGTTGCTACT3’  Reverse:5’CCAAGCTTCTAATACGACTCACTATAGGGAGATGACTAGGGTAAGCCGACGA3’	IDT	ZFIN:ZDB-GENE-040718-155
six1b in situ hybridization probe template primers  Forward:5’TTCTTTCGGGTTTACGCAGG3’  Reverse:5’CCAAGCTTCTAATACGACTCACTATAGGGAGAGAGCTGATGTGGGTGTCCTTG3’	IDT	ZFIN: ZDB-GENE-040426-2308
eya1 in situ hybridization probe	(Sahly et al., 1999)	ZDB-GENE-990712-18
foxg1a RNA FISH Probe	Molecular Instruments (Bell et al., 2024)	NC_007119.7
sox10 RNA FISH Probe	Molecular Instruments	NM_131875.1
sox2 RNA FISH Probe	Molecular Instruments	NM_213118.1
sox3 RNA FISH Probe	Molecular Instruments	NM_001001811.2
trpc2b RNA FISH Probe	Molecular Instruments	NM_001030166.2
foxd3 RNA FISH Probe	Molecular Instruments	NM_131290.2
ompb RNA FISH Probe	Molecular Instruments	NM_173281.2
Software and algorithms
PRISM	GraphPad	www.graphpad.com
FIJI	(Schindelin et al., 2012)	https://imagej.nih.gov/ij/