Foxg1 and Retinoic Acid Signaling Regulate Zonal Patterning in the Developing Olfactory Epithelium

ABSTRACT

Odor information processing begins in the olfactory epithelium (OE), which in mice is spatially divided into two zones: the dorsomedial zone (D‐zone), responsible for innate aversive behaviors, and the ventrolateral zone (V‐zone), associated with learning‐dependent behaviors. This zonal organization provides the structural framework for olfactory circuit function. However, the mechanisms driving OE zonal specification remain unclear. To investigate the initial segregation of the OE zones, we examined the role of Foxg1, a forkhead transcription factor expressed in the V‐zone throughout life. Conditional deletion of Foxg1 in Sox2‐expressing OE stem cells, coupled with lineage tracing, revealed ectopic localization of Foxg1‐lineage cells in the D‐zone, without altering their regional molecular profile. These results demonstrate that Foxg1 is essential for zonal segregation but is dispensable for zone‐specific molecular identity. We further revealed retinoic acid (RA) as an upstream morphogen regulating D‐zone‐specific gene expression. RA signaling is tightly confined to the D‐zone, ensuring OE regional identity. These findings suggest that the establishment of D‐ and V‐zones is driven by interactions between morphogenic signal and transcriptional program involving Foxg1, providing a molecular basis for understanding the formation of innate and learned olfactory circuits.

1. Introduction

Olfaction, the sense of smell, has evolved to detect chemical cues essential for navigation, foraging, reproduction, and recognition of beneficial or harmful stimuli (Ache and Young 2005; Jones and Roper 1997; Wyatt 2014; Wyatt 2010). In vertebrates, the processing of odor information begins with its detection in the olfactory epithelium (OE) (Buck and Axel 1991). Signals detected in the OE are transmitted to the olfactory bulb (OB) and relayed to higher brain regions, including the olfactory cortex, amygdala, and hippocampus, where they generate odor perception and behavioral responses (Takeuchi and Sakano 2014). In mice, the OE is spatially organized into two distinct zones: the dorsomedial zone (D‐zone), also referred to as zone 1, and the ventrolateral zone (V‐zone), which encompasses zones 2–4 (Gussing and Bohm 2004; Alenius and Bohm 1997). These zones are also characterized by unique molecular profiles and gene expression patterns (Ressler et al. 1993). Studies using D‐zone‐depleted (ΔD) mice and zone‐specific olfactory receptor (OR) knockouts have demonstrated that the D‐zone primarily mediates innate aversive responses and social behaviors, whereas the V‐zone and D‐zone contribute to learned, conditional odor responses (Kobayakawa et al. 2007; Jones et al. 2008; Matsuo et al. 2015; Dias and Ressler 2014; Cho et al. 2011).

Beyond their distinct behavioral roles, the D‐ and V‐zones of the OE exhibit marked differences in biological processes. These include variations in olfactory sensory neuron (OSN) turnover (Vedin et al. 2009), sensitivity to toxins (Vedin et al. 2004), age‐related declines in regenerative capacity (Genter and Ali 1998), and the speed of reconstitution following injury (Ishikura et al. 2023). Notably, studies on ΔD mice have revealed that the projection sites in the OB, normally targeted by axons from the D‐zone, remain vacant and are not dominated by axons from the V‐zone (Kobayakawa et al. 2007). This observation contrasts with other sensory systems, where axonal competition typically leads to neighboring neurons occupying vacant territories (Kobayakawa et al. 2007; Van der Loos and Woolsey 1973; Henschke et al. 2018). The absence of V‐zone axonal invasion into these unoccupied domains suggests that olfactory circuit formation mediating innate and learned behaviors is governed by robust, intrinsic regulatory mechanisms. However, the precise mechanisms underpinning the establishment of the zonal structure remain poorly understood, leaving a significant gap in our understanding of olfactory circuit assembly and function.

Foxg1, a member of the forkhead family transcription factors, is expressed in multiple neural tissues during embryonic development. Extensive research has shown that Foxg1 plays pivotal roles in cell proliferation (Xuan et al. 1995), differentiation (Hanashima et al. 2002; Martynoga et al. 2005), patterning (Hanashima et al. 2007; Muzio and Mallamaci 2005; Kumamoto et al. 2013), and fate determination (Hanashima et al. 2004; Cargnin et al. 2018; Hou et al. 2019) within the forebrain. In the peripheral nervous system, Foxg1 serves as a key marker of the placodes (Hatini et al. 1999), and its loss leads to severe defects across multiple sensory systems, including the OE, taste buds, and inner ear (Fan et al. 2019; Duggan et al. 2008; Ding et al. 2020). Additionally, Foxg1 is evolutionarily conserved and plays a critical role in defining the adenohypophysis and olfactory placode regions within the pre‐placodal ectoderm of jawed vertebrates (Poncelet and Shimeld 2020; Schlosser et al. 2014; Smith et al. 2015). In both mice and zebrafish, Foxg1 deficiency impairs neurogenesis and tissue growth, underscoring its conserved role in OE development (Duggan et al. 2008; Kawauchi, Kim, et al. 2009; Kawauchi, Santos, et al. 2009; Garaffo et al. 2015). Notably, in mice, Foxg1 exhibits specific expression in the V‐zone of the OE in mice from embryonic stages through adulthood (Duggan et al. 2008; Kawauchi, Kim, et al. 2009; Vedin et al. 2009), suggesting its potential role as a master regulator of OE zonal formation, extending beyond its broader evolutionary role in olfactory development.

In this study, we investigated the role of Foxg1 in OE zone formation using constitutive knockout mice and conditional Foxg1 deletions driven by the Sox2‐CreER mouse line to target the developing OE. Our findings identify Foxg1 as a key transcription factor required for OE zone segregation, while also highlighting the involvement of upstream extrinsic signals in this process. Specifically, expression and functional analyses revealed retinoic acid (RA) signaling as a critical regulator of zone organization. These results suggest that the formation of D‐ and V‐zones is orchestrated through cooperative interactions between morphogenic signals and transcriptional programs, providing valuable insights into the molecular mechanisms underlying innate and learned olfactory circuit formation.

2. Materials and Methods

2.1. Animals

All animal experiments were conducted in accordance with institutional guidelines and were approved by the Institutional Animal Care and Use Committee of Waseda University. Mice were housed in the Animal Housing Facility at Waseda University (Tokyo, Japan). Foxg1 ^−/− knockout mice (Foxg1 ^LacZ/LacZ ) were obtained by intercrossing Foxg1 ^LacZ/+ mice (Xuan et al. 1995). Foxg1 conditional knockout mice were generated by crossing Sox2 ^CreER/+ (Sankoda et al. 2021) with Foxg1 ^{loxP‐Flpe/loxP‐Flpe} ; Rosa26 ^{CAG‐FRT‐stop‐FRT‐EGFP/CAG‐FRT‐stop‐FRT‐EGFP} homozygous animals (Miyoshi and Fishell 2012; Sousa et al. 2009). The noon of the vaginal plug was designated as E0.5.

2.2. Tamoxifen Administration

(Z)‐4‐Hydroxytamoxifen (4‐OHT) solution was prepared in 100% ethanol and diluted with corn oil to a final concentration of 1 mg/mL. Pregnant females were administered 4‐OHT (H7904, Sigma) by intraperitoneal injection. A total of 0.1 mg 4‐OHT per mouse was administered.

2.3. All‐Trans‐Retinoic Acid Treatment

All‐trans‐RA (R2625, Sigma; ATRA) was suspended in dimethyl sulfoxide (DMSO) (25 mg/mL). Pregnant mice were administered either ATRA (100 mg/kg of body weight) or DMSO (control) every 12 h from E9.5 to E11.5 by oral administration using feeding needles (Natsume Seisakusho, 0655582661).

2.4. Tissue Preparation

Tissues were fixed in 4% paraformaldehyde (PFA) in phosphate‐buffered saline (PBS) at 4°C for 1 h. Mouse embryos over E15.5 were transcardially perfused with 4% PFA in PBS. After cryoprotection through a series of 10%, 20%, and 30% sucrose in PBS, tissue samples were embedded in Tissue‐Tek O.C.T compound (Thermo) and stored at −80°C until sectioning. Sections were made using a cryostat (Microm HM550; Carl Zeiss) at 20 μm for immunohistochemistry (IHC) and 12 μm for in situ hybridization (ISH) and collected on MAS‐coated glass slides (Matsunami, S9441).

2.5. In Situ Hybridization

Riboprobes for Foxg1, Acsm4, LacZ, Rarb, Bmp4, Six3, Fgf8, Raldh2, Raldh3, Cyp26a1, Cyp26b1, and Olfr16 were generated based on sequences provided by Allen Brain Atlas (https://mouse.brain‐map.org/static/atlas) and as previously described (Hatini et al. 1994; Hatini et al. 1999). RNA probes were synthesized as described previously (Hou et al. 2017) using T3, T7, or Sp6 RNA polymerase (Roche 11031163001, 10881767001, 10810274001) with digoxigenin (DIG) (Roche 11277073910) or Fluorescein (Roche, 11685619910) labeling mix in the presence of an RNase inhibitor (Roche, 03335399001). The oligonucleotide sequences used for the construction of riboprobes are listed in Table 1.

TABLE 1.

List of primer sequences.

Gene	Primer sequences
	Forward	Reverse
Acsm4	AAGGCTGCCAATGTGCTC	TTCTCCTTTGCGATTGGC
Raldh3	GCCAGTTGGAGACCCCTT	TGCGGTGTCCTGCACTTA
Bmp4	CCTGCAGCGATCCAGTCT	GCCCAATCTCCACTCCCT
Rarb	GACCTTGAGGAACCAACAAAAG	ACAACCTCGGTGTCTTGGTTAT
Raldh2	GTATATGGGAGCCCTCATCAAG	CCTCTTAGGAGTTCTTCTGGGG
Cyp26a1	CCCAAGGGCTGGAATGTTA	CCACGGGACTGTAGTAGAGACA
Cyp26b1	GCAATCTTTTTCCTCTCTCTCTTCG	CTTCAGCTCCTGCATGGTCA
Six3	AGAGTTGTCCATGTTCCAGTTG	CTGATTTCGGTTTGTTCTAGGG
Fgf8	CAGGTCCTGGCCAACAAG	AGCTCCCGCTGGATTCCT
Olfr16	GCGGGATTCCTGAAGTGATAG	GCAAGGGAGTGATGATAGTGTAG

For in situ hybridization (ISH), slides were dried at 37°C for at least 15 min, then immersed in 1× DEPC‐PBS for 10 min, followed by 4% PFA/PBS for 10 min, and subsequently washed in 1× DEPC‐PBS three times for 5 min each. Sections were then prehybridized in hybridization buffer (50% formamide, 2× SSC, 1× Denhardt's solution, 20% dextran sulfate, 0.5 mg/mL baker's yeast RNA [Sigma R6750], 0.5 mg/mL ssDNA) for 2 h at room temperature. Following prehybridization, sections were hybridized with hybridization buffer with probes (1 μg/mL) at 65°C overnight (O/N). Sections were then washed with 2× SSC/50% formamide at 60°C for 30 min, 0.2× SSC/50% formamide at 42°C for 30 min, and 0.2× SSC twice for 10 min at room temperature. After washing, sections were soaked in Buffer I (100 mM Tris‐HCl, pH 7.4, 150 mM NaCl), or 10 min at room temperature, then in Blocking solution (1% blocking reagent in Buffer I) for 1 h at room temperature.

For chromogenic single‐color ISH, sections were incubated with anti‐Digoxigenin‐AP antibody (Roche 11093274910; 1:500) in Buffer I for 30 min at room temperature after blocking. Sections were then washed with Buffer I three times for 15 min each, followed by incubation in Buffer II (100 mM NaCl, 50 mM MgCl₂, 100 mM Tris–HCl, pH 9.5) for 5 min at room temperature. The color reaction was performed in Buffer II containing NBT, BCIP and 1 mM levamisole O/N at room temperature. Sections were mounted using CC/Mount tissue mounting medium (Sigma). Images were acquired using a BZ‐X710 All‐in‐One fluorescent microscope (KEYENCE).

Fluorescent single‐color and two‐color ISH were performed as previously described (Yamauchi et al. 2022). Following blocking, sections were incubated O/N at 4°C with anti‐DIG‐ or anti‐Fluorescein‐POD antibodies (Roche 11207733910 or 11426346910) (1:1000) in Buffer I. Sections were then washed three times with 1xPBS for 10 min at room temperature. After washing, sections were incubated in Tris buffer (50 mM Tris–HCl, pH 7.5) for 5 min, followed by incubation in reaction buffer containing 10 μM CF tyramide (Biotium, 9217192173) and 3 μg/mL glucose oxidase in 2% BSA for 5 min at room temperature. The glucose oxidase‐based fluorochromized tyramide reaction was initiated by adding 5 μg/mL β‐D‐glucose, followed by incubation for 1 h at room temperature. For two‐color ISH, after POD inactivation with 3% H₂O₂/PBS, the same procedure starting from the blocking step was repeated.

2.6. Immunohistochemistry

After drying, slides were soaked in 1× PBS twice for 5 min. Sections were permeabilized with 0.3% Triton X‐100/PBS for 20 min. Slides were washed with PBS and incubated with a blocking buffer at room temperature for 15 min. The primary antibody solution was then applied, and the slides were incubated O/N at 4°C. After incubation, the slides were washed three times with PBS and subsequently incubated with a secondary antibody solution at room temperature for 1 h. Sections were treated with DAPI (1:10,000, DOJINDO) for 5 min at room temperature for nuclei staining. Sections were mounted with Polyvinyl alcohol mounting medium with DABCO, antifade (Supelco). The following primary antibodies were used: rabbit anti‐Sox2 (1:2000; Abcam), mouse anti‐HuC/D (1:1000; Invitrogen), chicken anti‐Green Fluorescent Protein (GFP) (1:1000; Abcam), goat anti‐NQO1 (1:500; Abcam), rat anti‐Ctip2 (1:1000; Abcam), rabbit anti‐Foxg1 (1:500; Abcam), rabbit anti‐Caspase3 (1:500; Cell signaling technology) and rabbit anti‐OMP (1:5000; Abcam). Secondary antibodies conjugated with Alexa Fluor dye (ThermoFisher) were used at 1:500 dilutions, except donkey anti‐chicken‐488 (Jackson ImmunoResearch) was used at a 1:500 dilution. Images were acquired using a BZ‐X710 All‐in‐One fluorescent microscope (KEYENCE).

2.7. Dual ISH and Immunohistochemistry (ISH‐IHC)

For combined ISH and IHC, fluorescent single‐color ISH was performed first. Following the glucose oxidase‐based fluorochromized tyramide reaction, slides were washed in 1× PBS and processed for IHC, beginning from the blocking step.

2.8. LacZ Staining

For β‐galactosidase (LacZ) staining, slides were incubated in LacZ staining solution (0.1 M PBS, 0.02% NP‐40, 1 mg/mL 5‐bromo‐4‐chloro‐3‐indolyl‐β‐D‐galactoside (X‐gal), 2 mM MgCl₂, 0.01% sodium deoxycholate, 5 mM potassium ferricyanide, 5 mM potassium ferrocyanide) at 37°C for 1 h. For LacZ‐IHC double staining, IHC was performed following LacZ staining.

2.9. Quantification

Quantifications were performed manually using 2–3 sections per sample. For the analysis of GFP‐positive cells shown in Figure 3d, 5 Foxg1‐cHET and 5 Foxg1‐cKO embryos were analyzed. For the quantification presented in Figure S2c, a total of 10 embryos were analyzed.

FIGURE 3.

Ectopic distribution of Foxg1‐lineage cells upon conditional Foxg1 deletion. (a) Schematic diagram showing conditional deletion of Foxg1 gene in olfactory placode stem cells. Foxg1‐cHET indicates Foxg1 conditional heterozygote of Sox2 ^CreER/+ ; Foxg1^{+/flox‐Flpe} ; RCE genotype embryos, Foxg1‐cKO indicates Foxg1 conditional KO of Sox2 ^CreER/+ ; Foxg1 ^{LacZ/flox‐Flpe} ; RCE genotype embryos. In both Foxg1‐cKO mice and Foxg1‐cHET mice, GFP marks the progeny of Foxg1‐lineage cells following tamoxifen induction. (b) Foxg1 and Sox2 IHC in E9.5 Foxg1 ^LacZ/+ mouse embryos. Enlarged views of the boxed regions are shown in bottom panels. (c) Left: Schematic diagram of OE zones in E15.5 wildtype embryo. Right panels: GFP and NQO1 IHC in E9.5 tamoxifen (4‐OHT) injected Foxg1‐cHET and Foxg1‐cKO mouse embryos analyzed at E15.5. (c′) indicates enlarged views of the boxed regions in (c). Red and blue boxes correspond to D‐zone and V‐zone, respectively. (d) Quantifications of GFP+ cell numbers within the NQO1+ epithelium of Foxg1‐cHET and Foxg1‐cKO embryos administered with 4‐OHT at E9.5 and analyzed at E15.5. Each data point represents an individual embryo, 5 Foxg1‐cHET and 5 Foxg1‐cKO embryos were analyzed (***p = 0.0011; Welch's t test). (e) Co‐labeling with GFP IHC and LacZ staining in E15.5 Foxg1‐cKO mice administered 4‐OHT at E9.5. The right panels show enlarged views of the boxed regions, with LacZ staining converted to magenta pseudocolor. Arrowhead shows double‐positive cells. Scale bars: (b) 100 μm, (c, e) 200 μm.

3. Results

3.1. Olfactory Epithelium Zone Specification Begins at E10.5

The olfactory epithelium (OE), derived from the olfactory placode (OP, Figure 1a), consists of two distinct regions: the dorsomedial zone (D‐zone) and the ventrolateral zone (V‐zone) (Figure 1b). Several genes have been reported to exhibit zone‐specific expression, including Acsm4 and NQO1 in the D‐zone, and Foxg1, Bcl11b, OCAM in the V‐zone (Oka et al. 2003; Gussing and Bohm 2004; Vedin et al. 2009; Enomoto et al. 2019). We first sought to identify the onset of zone specification by analyzing the expression of Foxg1 and Acsm4, both of which have been reported to be expressed from early embryonic development (Oka et al. 2003; Kawauchi, Kim, et al. 2009). ISH with specific probes revealed Foxg1 expression in the olfactory placode as early as E9.5, whereas Acsm4 was undetectable at this stage (Figure 1c). By E10.5, Foxg1 was expressed in the ventrolateral region of the olfactory pit but was absent in the dorsomedial region (Figure 1c, arrowhead). At the same stage, Acsm4 expression became detectable in the dorsal region, where Foxg1 was absent (Figure 1c). By E11.5, Foxg1 was observed in all OE regions except the dorsomedial area (Figure 1c) and was also detected in the vomeronasal organ (VNO) and the respiratory epithelium (RE) (Figures 1d and 2b), while Acsm4 expression was confined to the dorsomedial region (Figure 1c,d). Double fluorescent ISH using specific probes for Foxg1 and Acsm4 revealed predominantly complementary expression patterns within the OE, including mosaic expression near the border between the D‐ and V‐zones (Figure 1d″). By E15.5, Foxg1 and Acsm4 delineated an expression boundary between the D‐zone and V‐zone (Figure 1e), indicating the establishment of the D‐ and V‐zones and maintained into adulthood (6 weeks) (Figure 1b). These results indicate that the complementary expression of Foxg1 and Acsm4 begins from E10.5, marking the onset of zonal specification in the OE. Notably, the expression pattern also indicated that the zone structure develops prior to the formation of the olfactory bulb (Miller et al. 2010), implying that zone specification is governed by a molecular network intrinsic to the OE.

FIGURE 1.

Spatiotemporal expression of Foxg1 and Acsm4 in the developing olfactory epithelium. (a) Schematic diagram showing the formation of olfactory epithelium (OE), vomeronasal organ (VNO), respiratory epithelium (RE) from olfactory placode (OP). FB: forebrain. (b) Left panel: Schematic diagram of a 6‐week (w)‐old mouse coronal section of the OE. OB: olfactory bulb. D‐zone: dorsomedial zone, V‐zone: ventrolateral zone. Right panels: LacZ reporter staining of Foxg1 ^lacZ/+ heterozygote mouse OE representing Foxg1 transcription and Acsm4 in situ hybridization (ISH). (c) Foxg1 and Acsm4 ISH of embryonic day (E) 9.5, E10.5, and E11.5 embryo OE coronal sections. (d) Double fluorescent ISH of Foxg1 (green) and Acsm4 (magenta) at E11.5. (d′, d″, d‴) indicates enlarged views of the boxed regions indicated in (d) with DAPI staining (white). (b) and (d) were flipped horizontally to match the medial/lateral orientation to other figures. (e) Left panel: Schematic diagram of E15.5 mouse coronal section at OE levels. Right panels: Foxg1 and Acsm4 ISH of E15.5 embryo. Scale bars: (c, d) 100 μm, (b) 500 μm, (d′, d″, d‴) 20 μm, (e) 200 μm.

FIGURE 2.

Constitutive loss of Foxg1 leads to mis‐segregation of OE zones. (a) Immunohistochemistry (IHC) of stem cell marker Sox2 (magenta) and neuron marker HuC/D (green) in E12.5 Foxg1 ^LacZ/+ control and Foxg1 ^LacZ/LacZ knockout (KO) embryos. Dashed lines indicate the boundary between HuC/D⁺ OE region and HuC/D⁻ RE region. (b) Left panels: ISH of Acsm4, Raldh3, Bmp4, and LacZ in E11.5 Foxg1 ^LacZ/+ and Foxg1 ^LacZ/LacZ KO mice. Right: Schematic diagram indicating regional markers. (c, d) ISH of lacZ, Six3, and Fgf8 at E10.5 (c) and E11.5 (d) in Foxg1 ^LacZ/+ and Foxg1 ^LacZ/LacZ mice. Scale bars: 100 μm.

3.2. Constitutive Loss of Foxg1 Leads to Mis‐Segregation of OE Zones

Foxg1 is a master regulator of forebrain development, which plays a critical role in cerebral cortex patterning (Du et al. 2019). Previous studies have shown that Foxg1 regulates self‐renewal of OE cells through cyclin‐dependent kinase inhibitor (CKI) p21Cip1 inhibition (Kawauchi, Kim, et al. 2009). Loss of Foxg1 leads to severely underdeveloped OE, and OE is no longer detectable at E13.5 (Kawauchi, Kim, et al. 2009). To investigate whether Foxg1 contributes to zone specification in the OE, we generated Foxg1 constitutive knockout (KO) mice by intercrossing Foxg1 heterozygote mice carrying a LacZ reporter knocked into the Foxg1 locus (Xuan et al. 1995). Immunohistochemistry for Sox2 revealed that despite the overall reduction of the OE in the Foxg1 ^lacZ/lacZ KO mice, Sox2‐positive stem cells were detected throughout the OE (Figure 2a). Similarly, immunolabeling with neuronal marker HuC/D indicated the presence of both neurogenic and non‐neurogenic epithelium in the Foxg1 KO mice (Figure 2a, border indicated by dashed lines). The expression of Bmp4, a marker for the non‐neurogenic RE, indicated that both the RE and OE were induced from the olfactory placode, similar to what was observed in the control mice (Figure 2b) (Kawauchi, Kim, et al. 2009; Forni et al. 2013). To assess zone formation in the OE of Foxg1 KO embryos, we analyzed the expression of Acsm4 as a marker for D‐zone cell identity and Raldh3 as a marker for V‐zone cell fate (Figure 2b) (Login et al. 2015). The results revealed that the expression of Acsm4 was not detected in the OE in Foxg1 KO embryos, while Raldh3 expression was confined to a limited region of the ventrolateral OE (Figure 2b). Given that Raldh3 is expressed in both the V‐zone and the RE, we further examined the expression of Bmp4, which is restricted to the RE (Forni et al. 2013). Overlap between the expression of Raldh3 and Bmp4 indicated that Raldh3 is not expressed in the OE in Foxg1 KO embryos. Moreover, lacZ expression, reflecting the Foxg1 promoter activity, indicated the activation of Foxg1 promoter throughout the OE, despite the absence of functional Foxg1 protein (Figure 2b).

In the developing telencephalon, Foxg1 expression is induced by Fgf8 in the presence of Six3 (Shimamura and Rubenstein 1997; Kobayashi et al. 2002). To investigate whether this regulatory network is conserved in the OE, we examined the spatiotemporal expression of Fgf8 and Six3 in the developing OE (Figure 2c). At E10.5, Fgf8 was detected at the rim of the olfactory pit in Foxg1 ^lacZ/+ heterozygote control embryos, while Six3 expression was localized to the ventral region and absent in the dorsal region (upper panels, Figure 2c). By E11.5, Fgf8 expression became restricted to the ventral‐most region near the nasal fin, and Six3 expression was localized to the ventrolateral region of the OE, overlapping with that of Foxg1 expression. This suggested that the Six3‐Fgf8‐Foxg1 regulatory network is conserved in the OE (upper panels, Figure 2c,d).

To further test this hypothesis, we analyzed the expression of Six3 and Fgf8 in Foxg1 KO mice to determine whether their expression accounts for the expanded activity of the Foxg1 promoter. Consistent with the findings at E11.5 (Figure 2b), at E10.5, lacZ expression was detected throughout the olfactory pit in Foxg1 ^lacZ/lacZ KO embryos (Figure 2c). Furthermore, while Fgf8 expression was localized to the rim of the nasal pit, its expression expanded dorsally as compared with the Foxg1 ^lacZ/+ controls. Similarly, Six3, which is typically restricted ventrally, was expressed throughout the olfactory pit, encompassing the dorsal region (Figure 2c). At E11.5, both Fgf8 and Six3 exhibited dorsal expansion, with expression observed throughout the OE (Figure 2d). These changes in expression patterns align with the activation of the Foxg1 promoter based on lacZ reporter expression, supporting the notion that the Fgf8‐Six3‐Foxg1 regulatory network is conserved in the OE. These findings demonstrate that the loss of Foxg1 leads to a mis‐segregation of OE zones, despite maintained neurogenesis.

3.3. Foxg1 Is Necessary for Correct Zone Specification

The significant impact of Foxg1 removal on OE neurogenesis and tissue growth complicated precise assessment of its role in zone formation. To address this limitation, we utilized conditional Foxg1 KO mice to examine the cell‐autonomous role of Foxg1 in zonal formation, avoiding confounding effects of widespread developmental defects. Conditional deletion of Foxg1 was achieved by crossing Sox2 ^CreER/+ mice (Sankoda et al. 2021) carrying a lacZ allele in the Foxg1 locus (Sox2 ^CreER/+; Foxg1 ^lacZ/+ double heterozygotes) with floxed Foxg1 mice (Foxg1 ^{flox‐Flpe/flox‐Flpe} ) carrying a Flp‐dependent EGFP reporter allele (R26R ^{CAG‐FRT‐stop‐FRT‐EGFP/CAG‐FRT‐stop‐FRT‐EGFP} : RCE) (Miyoshi and Fishell 2012) (Figure 3a). This approach enabled Cre‐dependent deletion of Foxg1 and simultaneous transcription of the Flpe recombinase induced by tamoxifen, resulting in GFP expression specifically in Foxg1‐deleted, Foxg1‐lineage cells (Figure 3a). The genotypes were designated as follows: Foxg1‐cKO (Sox2 ^CreER/+ ; Foxg1 ^{LacZ/flox‐Flpe} ; RCE) and Foxg1‐cHET(Sox2 ^CreER/+ ; Foxg1 ^{+/flox‐Flpe} ; RCE) as controls. In both Foxg1‐cKO mice and Foxg1‐cHET mice, GFP marks the progeny of Foxg1‐lineage cells after tamoxifen induction.

During embryogenesis, Sox2 is expressed in OE stem cells (Figures 2a and 3b) (Li et al. 2022), overlapping with Foxg1‐expressing populations throughout the olfactory placode at E9.5 (Figure 3b). To ensure recombination prior to zone‐specific Foxg1 and Acsm4 expression, tamoxifen was administered at E9.5, before zonal structures were established (Figure 1c). At E15.5, a stage when zone structures are well defined (Figure 1e), immunolabeling for NQO1 revealed that in Foxg1‐cHET embryos, very few GFP‐positive cells (2.2 ± 0.52, n = 5 embryos) were present in the NQO1‐positive D‐zone. In contrast, Foxg1‐cKO embryos exhibited a significantly increased number of GFP‐positive cells within the D‐zone (33.0 ± 3.97, n = 5 embryos, p = 0.0011, Figure 3c,d). These findings indicate that Foxg1‐lineage cells predominantly contribute to the V‐zone, which was unexpected given the broad expression of Foxg1 and Sox2 throughout the olfactory placode (Figure 3b). To further investigate this, we performed short‐term fate mapping of Foxg1‐expressing lineage cells using both Foxg1 ^lacZ/+ heterozygous and Foxg1‐cHet embryos to assess temporal changes in their spatial distribution. We first examined whether the D‐ and V‐zones originate from a common progenitor population using Foxg1 ^lacZ/+ embryos. Although Foxg1 transcripts were not detected in the D‐zone at E10.5 (Figure 1c), LacZ staining was observed in this region at the same stage (Figure S1a), indicating perdurance of reporter activity from earlier Foxg1‐expressing progenitors. This suggests that both the D‐ and V‐zones are derived from an early Foxg1‐expressing progenitor population. To further characterize the spatial distribution of Foxg1‐lineage cells over time, we analyzed GFP‐positive cells in Foxg1‐cHET embryos at E11.5 and E13.5 (Figure S1b). At both stages, a small number of GFP‐positive cells were detected within the D‐zone (open arrowheads: GFP‐positive cells in the D‐zone; filled arrowheads: GFP‐positive cells in the V‐zone; Figure S1b). These findings suggest that Foxg1‐lineage cells initially populate the D‐zone but are subsequently excluded.

To explore the potential mechanism underlying this exclusion, we assessed cell death during this developmental window using immunohistochemical staining for cleaved caspase‐3. However, we did not observe co‐localization of GFP and caspase‐3 within individual cells (Figure S2a), and caspase‐3‐positive cells showed no consistent spatial distribution between the D‐ and V‐zones (Figure S2b,c). These data argue against cell death as the primary mechanism for the exclusion of Foxg1‐lineage cells. Nonetheless, given the transient activation of caspase‐3 during apoptosis, we cannot completely rule out the contribution of cell death to this process. Collectively, these results indicate that while both the D‐ and V‐zones originate from Foxg1‐expressing progenitors, Foxg1‐lineage cells are excluded from the D‐zone as development proceeds. Moreover, cell‐autonomous loss of Foxg1 permits these lineage cells to persist within the D‐zone.

GFP‐positive cells located in the D‐zone of Foxg1‐cKO embryos acquire D‐zone characteristics, as indicated by the overlapping expression of GFP and NQO1 (filled arrowheads, Figure 3c′). Notably, ectopic NQO1 expression was not observed in the V‐zone of Foxg1‐cKO OE (Figure 3c′). Furthermore, LacZ was detected in GFP‐positive cells located in the V‐zone, indicating active transcription from the Foxg1 promoter in Foxg1‐cKO cells (arrowheads, Figure 3e, right panel). These results demonstrate that the molecular identity of OE cells is not altered by the loss of Foxg1, resulting in a molecular profile consistent with their spatial location within the OE (Figure 3c′,e). To further determine whether the ectopic GFP‐positive cells observed in the D‐zone of Foxg1‐cKO mice adopt the molecular and functional characteristics of D‐zone neurons, we analyzed these mice at postnatal day 7 (P7). Since Foxg1‐cKO mice exhibit perinatal lethality due to the broad KO of Foxg1 in multiple neural tissues, P7 was selected as the latest viable time point for analysis. Since the olfactory circuit is established by P7 (Wu et al. 2018), this stage represents a sufficiently mature developmental time point to assess the functional differentiation of OSNs. We first examined whether GFP‐positive cells in the D‐zone had differentiated into mature OSNs and expressed markers indicative of D‐zone identity. Immunohistochemical analysis revealed that these cells expressed NQO1, a marker for D‐zone OSNs, as well as olfactory marker protein (OMP), a marker of mature OSNs, indicating that the ectopic cells had acquired the molecular identity of mature D‐zone OSNs (Figure 4b). To further assess their functional identity, we analyzed the expression of the OR gene, Olfr16 (also known as MOR23), which is expressed in a small subset of D‐zone OSNs (Vassalli et al. 2002). Co‐localization of GFP and MOR23 confirmed that these ectopic cells are capable of expressing D‐zone‐specific ORs (Figure 4c). Collectively, these findings suggest that Foxg1‐lineage cells ectopically located in the D‐zone of Foxg1‐cKO mice acquire both molecular and functional characteristics of mature D‐zone OSNs.

FIGURE 4.

Zonal distributions are not altered in Foxg1‐cKO mice at P7. (a) Left: Experimental scheme and timeline for analysis shown in (b), (c), and (d). Right: Schematic diagram of OE zones and corresponding molecular markers at P7. (b–d) Left panels: Schematic diagram of OE zones. Boxed regions indicate areas shown at higher magnification in the corresponding right panels. (b) IHC for GFP, NQO1, and OMP in Foxg1‐cKO mice. (c) Dual ISH‐IHC for MOR23 fluorescence ISH and GFP IHC in Foxg1‐cKO mice. (d) IHC for GFP and Bcl11b in Foxg1‐cHET and Foxg1‐cKO mice at P7. Arrowheads indicate double‐positive cells. Scale bars: (b, c) 20 μm, (d) 50 μm.

To further validate the phenotype of Foxg1‐cKO mice, we examined the expression of Bcl11b, a transcription factor known to regulate the proper zonal distribution of ORs by repressing Class I OR expression (Enomoto et al. 2019). Analysis of Bcl11b expression in the V‐zone of Foxg1‐cHET and Foxg1‐cKO mice revealed that GFP‐positive cells in both genotypes maintained Bcl11b expression (Figure 4d), suggesting that the zonal regulatory network remains intact in the absence of Foxg1.

Collectively, these results highlight the critical role of Foxg1 in lineage‐dependent cell positioning, ensuring OE zonal segregation. Additionally, the data suggest the presence of upstream signals encoding positional information within the OE.

3.4. RA Signaling as a Candidate for Encoding Positional Information in the OE

To investigate upstream signals encoding positional information in the OE, we focused on secreted signals for two key reasons. First, the expression patterns of Foxg1 and Acsm4 suggest that the boundary between the D‐zone and the V‐zone is gradually established (Figure 1c–e). Second, zone‐specific gene expression begins after the major morphological event of invagination (Figure 1a,c). These dynamic morphological changes alter the position and size of the OE, making secreted factors particularly well suited for regulatory roles. Given that Foxg1 initiates zone‐specific expression by E10.5 (Figure 1c), we focused on secreted signals that exhibit zone‐specific expression during the critical E9.5–E10.5 developmental window. Candidate genes were selected using the Allen Brain Atlas (ABA, https://mouse.brain‐map.org/static/atlas) and Gene Expression Database (GXD, https://www.informatics.jax.org/expression.shtml). Among these, we identified Rarb, a receptor for RA, which exhibited expression in and around the dorsomedial OE, where Acsm4 is highly expressed (Figure 5a,b). Rarb functions both as a receptor and transcription factor in response to retinoic acid (RA) the active form of vitamin A, a well‐established morphogen during embryonic development (Rhinn and Dollé 2012). RA is synthesized by specific cell populations expressing one of the three retinaldehyde dehydrogenases (Raldh1, Raldh2, or Raldh3) and acts as a ligand for nuclear RA receptors (RARs), which regulate the transcriptional activity of target genes.

FIGURE 5.

Spatiotemporal expression of RA receptor RARb and RA‐synthesizing enzymes Raldh1, Raldh2, and Raldh3 in the developing OE. (a) ISH of Retinoic acid receptor b (Rarb), retinaldehyde dehydrogenases 1, 2 and 3 (Raldh1, 2, 3) in E9.5, E10.5, E11.5 wildtype embryos. (b) Double fluorescent ISH of Acsm4 (magenta) and RARb (green) at E11.5. (b′) shows enlarged views of the boxed regions. Scale bars: (a, b) 100 μm, (b′) 20 μm.

To evaluate RA signaling during zone specification, we analyzed the expression patterns of Rarb and the RA‐synthesizing enzymes Raldh2 and Raldh3, previously reported to be expressed in the OE (Paschaki et al. 2013) from E9.5 to E11.5. ISH revealed that at E9.5, Rarb was expressed in the mesenchyme surrounding the olfactory placode, with stronger lateral expression (Figure 5a). By E10.5, Rarb expression became prominent within the D‐zone and adjacent mesenchyme (arrowhead and arrow, top middle panel in Figure 5a). At E11.5, Rarb expression concentrated in the D‐zone and dorsal/lateral mesenchyme (arrowhead and arrow, top right panel in Figure 5a). Double fluorescent ISH revealed co‐expression of Acsm4 and Rarb within the same cells (arrowheads, Figure 5b′). Raldh2, in turn, was detected at E10.5 in the mesenchyme adjacent to the D‐zone of the olfactory pit and persisted at E11.5 in the mesenchyme surrounding the D‐zone and nasal fin (arrows, Figure 5a). In contrast, Raldh3 showed broad expression across the olfactory placode at E9.5, with stronger medial expression. By E10.5, Raldh3 localized to the dorsal OE, and by E11.5, its expression was observed throughout the OE, excluding the D‐zone and vomeronasal organ (VNO). These expression patterns suggest that RA is selectively synthesized in and around the D‐zone and received within the D‐zone by E10.5, coinciding with the onset of zone‐specific expression. Based on these findings, we proposed that RA signaling is a strong candidate for encoding positional information in OE cells.

3.5. RA Signaling Is Tightly Regulated in the Developing OE

To assess the role of RA in zone specification, we manipulated RA levels by administering ATRA at 100 mg/kg from E9.5 to E11.5 (Figure 6a). Previous studies have shown that high concentrations of RA are rapidly metabolized within 24 hour (Eckhoff et al. 1989); therefore, to maintain elevated RA levels, we administered ATRA every 12 hour. High RA exposure resulted in partially resorbed fetuses, while surviving embryos exhibited variable phenotypes, including abnormally small body size or normal appearance. Analyses were restricted to embryos with normal body size (n = 4). Histological examination revealed abnormal facial morphology in RA‐treated embryos, such as a deeper midline cleft and disoriented OE (arrow, Figure 6c). While RA has been shown to suppress neurogenesis in explant cultures (Paschaki et al. 2013), HuC/D immunolabeling showed that primary neurogenesis and RE induction were preserved in RA‐treated embryos (Figure 6b).

FIGURE 6.

Gain‐of‐function analysis using all‐trans retinoic acid (ATRA). (a) Experimental scheme showing the time course of ATRA administration. (b) HuC/D IHC and Foxg1 and Acsm4 ISH in E11.5 DMSO or ATRA treated embryos. (c, d) ISH of (c) Rarb, (d) Cyp26a1, and Cyp26b1 in E11.5 of DMSO or ATRA treated embryos. Scale bars: (b–d) 100 μm, (c, d) half or full section 400 μm.

To assess the effects of RA on zone specification, we analyzed the expression patterns of Foxg1 and Acsm4 in RA‐treated embryos. At E11.5, Acsm4 expression remained localized to the dorsomedial region, while Foxg1 was restricted to the ventrolateral region, indicating that high RA treatment did not disrupt zone formation (bottom panels, Figure 6b). To confirm the efficacy of RA treatment, we examined Rarb expression, which also serves as a readout of active RA signaling (Balkan et al. 1992). In RA‐treated embryos, ectopic expression of Rarb was detected in the facial mesenchyme and telencephalon, indicating widespread RA signaling (asterisks, Figure 6c). However, within the OE, Rarb remained confined to the dorsomedial region, consistent with Acsm4 expression in both RA‐treated and control embryos (Figure 6c). This suggests that RA signaling is tightly regulated and confined to the D‐zone, explaining the preservation of zone specification despite high RA exposure.

To investigate the mechanism underlying this restricted RA signaling, we examined whether upregulated RA degradation in nonactivated regions or the inherent responsiveness of the D‐zone to RA signaling accounts for this restriction. We analyzed the expression of RA‐degrading enzymes Cyp26a1 and Cyp26b1 (Figure 6d). In control embryos, Cyp26a1 was not expressed in the OE (asterisk, upper panel in Figure 6d); however, in RA‐treated embryos, Cyp26a1 was detected in the dorsomedial region, coinciding with Rarb and Acsm4 expression (arrowheads, lower panel in Figure 6d). In contrast, Cyp26b1 exhibited ectopic expression in the facial mesenchyme (asterisks, lower panel in Figure 6d), but was absent in the OE (Figure 6d). Given that Cyp26a1 is a downstream target of RA (Loudig et al. 2005), but not Cyp26b1, the selective induction of Cyp26a1 in the D‐zone suggests that RA activity in the OE is not restricted by increased degradation in the V‐zone, but rather by a D‐zone restricted receptive field for RA signaling. Furthermore, the upregulation of Cyp26a1 following high RA exposure indicated that RA concentration is tightly regulated within D‐zone cells through negative feedback mechanisms.

In summary, elevating RA levels through a teratogenic dose of ATRA does not disrupt OE zone specification. Instead, these findings underscore that the receptive field for RA signaling in the OE is tightly regulated and aligns with D‐zone parcellation, supporting the hypothesis that RA signaling provides positional information critical for establishing the OE zonal structure.

4. Discussion

Foxg1 is an evolutionarily conserved gene with a well‐established role in olfactory placode development (Poncelet and Shimeld 2020; Schlosser et al. 2014; Smith et al. 2015). Previous studies have demonstrated its critical role in olfactory neurogenesis in mice and zebrafish, suggesting its conserved function in olfactory neurogenesis across vertebrates (Duggan et al. 2008; Kawauchi, Kim, et al. 2009; Kawauchi, Santos, et al. 2009; Garaffo et al. 2015). Despite its importance in olfactory neurogenesis, the expression of Foxg1 in the mouse OE is restricted to the ventrolateral region (V‐zone) starting at E10.5 (Figure 1c), immediately following the onset of neurogenesis. The functional significance of this spatially restricted expression has remained unclear.

In this study, we present evidence supporting the role of Foxg1 in establishing zonal organization during the earliest phase of olfactory neurogenesis. Our findings reveal that Foxg1, in cooperation with retinoic acid (RA), plays a pivotal role in defining the zonal composition of the OE (Figure 7). Analysis of Foxg1‐cKO embryos revealed ectopic localization of Foxg1‐lineage cells within the D‐zone (bottom panel, Figure 7), identifying Foxg1 as a key transcriptional regulator of zonal organization in the developing OE. Notably, these Foxg1‐cKO cells exhibited persistent expression of D‐ and V‐zone markers even at P7, suggesting that cells acquire zonal identity independently of Foxg1. However, Foxg1 constitutive KO embryos showed no expression of Acsm4 (Figure 2b), which appears contradictory. Given that the OE in Foxg1 null embryos is severely underdeveloped in both size and the number of HuC/D‐positive neurons (Figure 2a), the loss of Acsm4 expression likely reflects impaired neurogenesis rather than a direct role of Foxg1 in zonal specification. Together, the earliest spatial restriction of Foxg1 expression and its demonstrated role in zonal segregation suggest that its spatially regulated expression is crucial for establishing the OE zonal structure.

FIGURE 7.

Model for the establishment of zonal identity in the developing OE. Top panels: Schematic diagram of olfactory epithelium induction from the olfactory placodes. Upon invagination, the dorsomedial olfactory epithelium is exposed to retinoic acid (RA) synthesized in the overlying mesenchymal tissues, leading to the induction of RARβ receptor and RA signaling responsible for Acsm4 expression. Bottom panels: Progression of D‐ and V‐zone segregation in the olfactory epithelium of Foxg1‐cHet and Foxg1‐cKO embryos. Sox2 ^CreER/+ mediated recombination of the floxed Foxg1 allele (illustrated in Figure 3a) marks Foxg1‐lineage cells with GFP. At E9.5, olfactory placode cells express Six3 and Sox2 (Purcell et al. 2005; Storm et al. 2006). Following E9.5 tamoxifen injection, Foxg1‐expressing lineage cells are progressively excluded from the D‐zone in Foxg1‐cHet embryos. In Foxg1‐cKO, loss of Foxg1 expression allows Foxg1‐lineage GFP cells to invade the D‐zone and express NQO1.

Our findings indicate that both the D‐ and V‐ zones of the OE originate from a common population of Foxg1‐expressing progenitor cells (Figure S1a). Interestingly, although these progenitors initially give rise to cells in both zones, expression of Foxg1 appears to restrict its lineage contribution predominantly to the V‐zone (Figure 3c,d). This restriction becomes progressively pronounced over developmental time, transitioning from a mixed to a sharply defined boundary between zones. Analysis using caspase‐3 suggests that cell death is not the primary mechanism underlying this lineage restriction (Figure S2a,b). This spatial distribution raises the possibility that the segregation of Foxg1‐lineage cells is regulated by coordinated signaling pathways and transcriptional networks. Foxg1 and RA signaling are strong candidates for mediating this process; however, further analysis will be necessary to elucidate the precise mechanisms underlying this lineage sorting. Given that both zones arise from a shared progenitor pool, the restriction of Foxg1‐lineage cells to the V‐zone may serve a critical developmental function. One plausible hypothesis is that this segregation enables the establishment of zone‐specific neurogenic programs. Foxg1 is known to regulate progenitor cell proliferation in the forebrain (Xuan et al. 1995; Hettige et al. 2022) and has also been shown to regulate self‐renewal of OE cells through inhibition of p21 ^Cip1 (Kawauchi, Kim, et al. 2009). Such regulation may underlie regional differences in the timing and rate of neurogenesis. Previous studies have shown that the topographic projection of olfactory senseory neuron (OSN) axons is established through the sequential maturation and axonal projection of OSNs along the dorsomedial‐to‐ventrolateral axis (Takeuchi et al. 2010; Eerdunfu et al. 2017). Thus, it is plausible that Foxg1 contributes to early regional patterning of the OE, providing a foundation for zone‐specific neurogenesis and subsequent functional organization.

Zonal structure also plays an essential role in odorant receptor (OR) choice. OR genes are classified into two classes: Class I and Class II, based on the homology of their amino acid sequences. These classes exhibit specific expression patterns within the OE. Class I OR‐expressing OSNs are restricted to the D‐zone, whereas Class II OR‐expressing OSNs are found in both the D‐ and V‐zones. Recent studies have identified the transcription factor Bcl11b as a key regulator of the V‐zone, functioning by repressing Class I OR expression, thereby ensuring the proper zonal distribution of ORs (Enomoto et al. 2019). Foxg1‐cKO cells located in the V‐zone retain Bcl11b expression (Figure 4d). This observation suggests that Foxg1 may not directly regulate Bcl11b expression in the developing OE, implying the existence of multiple regulatory networks that govern the establishment of zonal identity.

Olfaction is a critical sensory modality for both vertebrate and non‐vertebrate species, enabling the detection of environmental signals without direct contact and regulating innate behaviors. Olfactory perception also facilitates the encoding of complex information through combinatorial odor molecules, shaping species‐specific behaviors across generations (Buck and Axel 1991; Dias and Ressler 2014; Deshe et al. 2023; Remy 2010). While both innate and learning‐dependent circuits have been identified in various species (Kobayakawa et al. 2007; Ferrero et al. 2011; Semmelhack and Wang 2009; Suh et al. 2004; Roussel et al. 2014), the presence of zone structures within the OE appears unique to terrestrial mammals, including rats, mice, goats, macaques, and common marmosets (Octura et al. 2018; Ressler et al. 1993; Strotmann et al. 1994; Horowitz et al. 2014; Kurihara et al. 2022). This structural organization is absent in other vertebrates such as frogs, zebrafish, and chickens (Bear et al. 2016; Kishida et al. 2015), raising intriguing questions about the evolutionary processes that led to the emergence of zonal structures.

The ability to detect and differentiate odors is largely influenced by the number and diversity of OR genes, which constitute the largest multigene family in vertebrates (Niimura and Nei 2007; Niimura 2009; Niimura et al. 2014). Terrestrial mammals, in particular, have undergone a significant expansion of the OR gene repertoire, reflecting their adaptation to diverse and dynamic habitats (Niimura and Nei 2007; Niimura 2009; Niimura et al. 2014; Go and Niimura 2008). The evolutionary dynamics of OR genes, characterized by a birth‐and‐death process, have enabled species to optimize their response to complex chemical environments (Hayden et al. 2010; Liao et al. 2024). However, incorporating newly evolved OR genes into relatively fixed olfactory circuits poses a challenge. A potential breakthrough to address this is the spatial segregation of olfactory receptive areas into distinct zones, organized based on the functional roles of odors—whether processed by innate or learned circuits. This zonal organization allows the olfactory system to maintain stable, hard‐wired circuits for innate behaviors while providing flexible neural architecture to accommodate learned, conditional responses to odors. Such an arrangement allows the optimization of OR gene repertoires, enabling terrestrial mammals to detect and respond to a broad range of chemical stimuli, tailored to their specific habitats and lifestyles. The spatial organization of the OE underscores the evolutionary significance of zonal structures in balancing innate and learned behaviors in response to environmental demands.

A regulatory network involving morphogens and transcription factors is well documented in the patterning of various sensory organs (O'Sullivan et al. 2023; Plas et al. 2008; Glover et al. 2023; Villeneuve et al. 2024). However, such a network has not been characterized in the olfactory system. In this study, we provide the first evidence of cooperative interactions between the morphogen RA and the transcription factor Foxg1 in regulating the patterning of the OE. Our findings demonstrate that RA encodes positional information along the dorsomedial–ventrolateral axis, which is interpreted by Foxg1 to determine D‐zone or V‐zone identities. Notably, while no transcription factor specific to the D‐zone had been identified during early developmental stages, our expression analysis revealed that Rarb, a receptor for RA that also functions as a transcription factor, exhibits D‐zone‐specific expression beginning at E10.5. This positions Rarb as the earliest known transcription factor to exhibit D‐zone‐specific expression.

Our data further show that Foxg1 and Rarb serve as key patterning determinants in response to RA signaling during zone formation. Despite elevated RA signaling induced by high‐dose ATRA administration, the OE's zonal patterning remained intact. This resilience is likely due to the restricted RA receptive field within the dorsomedial OE, suggesting an upstream mechanism regulating RA reception. Interestingly, the establishment of the RA receptive field occurs shortly after the olfactory placode is identified, suggesting a potential role of mechanical forces during invagination—a distinctive feature of placodes—in shaping the receptive field. This hypothesis aligns with observations in other sensory systems, such as hair follicle and dental placodes, where mechanical forces and changes in cell shape influence patterning (Villeneuve et al. 2024; Yamamoto et al. 2015). Together, these findings suggest a unique mechanism in which RA signaling and transcriptional regulation integrate spatial cues and mechanical processes to drive OE patterning. Future investigations into the interplay between mechanical cues, signaling pathways, and transcriptional networks will be key to unraveling the complex regulatory processes underlying OE development. These insights could further illuminate the implications of OE patterning in sensory function, species‐specific adaptation, and the maintenance of neural circuits throughout life.

Author Contributions

A.K. and C.H. designed research; A.K. performed research; A.K. and C.H. analyzed data; and A.K. and C.H. wrote the paper.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1: Supporting Information.

Kuriyama, A. , and Hanashima C.. 2025. “Foxg1 and Retinoic Acid Signaling Regulate Zonal Patterning in the Developing Olfactory Epithelium.” Development, Growth & Differentiation 67, no. 6: 314–330. 10.1111/dgd.70020.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.