The Indirect Role of Fibroblast Growth Factor-8 in Defining Neurogenic Niches of the Olfactory/GnRH Systems

Bmp4 is expressed in the developing nasal mesenchyme and is coexpressed with FGF8 in the RE. A–D, E11.5, parasagittal sections, orientation indicated in A. R, Rostral; D, dorsal; V, ventral; C, caudal. Use dashed line as reference to separate non-neurogenic and neurogenic OP areas. A, Bmp4 in situ hybridization showed expression along the RE (black arrowheads) and in rostral, dorsal, and ventral nasal mesenchyme (white arrowheads). Bmp4 expression was not detected in the caudal nasal mesenchyme (asterisk) facing the neurogenic VNO (compare to B). B, Fgf8^nullLacz/WT; X-gal enzymatic reaction (blue) and HuC/D immunostaining (brown) highlights the occurrence of neurogenesis mainly caudal to Fgf8-expressing areas (blue). Fgf8 was expressed in areas where Bmp4 expression was detected (compare black arrows in A and B). C, Immunostaining for p-SMAD 1,5,8 (brown) detected active BMP4 downstream signaling in areas of Bmp4 expression (A, C, white arrowheads) and along the RE (black arrowheads). Decreased p-SMAD 1,5,8 immunoreactivity was detected in area facing the VNO (asterisk). D, MSX1/2 immunolabeling (dark brown) revealed a pattern similar to p-SMAD (compare with C) with high MSX1/2 levels in the RE (black arrowheads) and mesenchyme proximal to Bmp4 sources (white arrowheads; compare with A). No expression was found in the mesenchyme facing the neurogenic VNO (asterisk). E, F, MSX1/2 immunolabeling (brown) and Fgf8 expression (X-gal enzymatic reaction, blue) revealed co-localization in the RE (black arrowheads). Scale bars: (in A) A–D, 100 μm; F, 100 μm.

Fgf8 genetic lineage cell fate tracing. A–E, Parasagittal (A–C) and coronal (D, E) sections of E11.5 Fgf8Cre/RosaYFP embryos. A, Yellow fluorescent protein (YFP) immunostaining (brown) reveals Fgf8 lineage (YFP expression) in the facial epidermis (Ep) and along the rostral RE. YFP expression was not detected in the OE and VNO, confirming distinct genetic lineage for the Fgf8-expressing epidermis and these structures. B–C, YFP (green) and tubulin+HuCD immunostaining (red). Neurons (red) of the OE (B) and VNO (C) derive from progenitors negative for Fgf8 expression (green). D, Section shows the distinct lineage for neuronal cells of OE, VNO, and migratory mass (MM; red only) and the Fgf8 (YFP+)-expressing epidermal cells (green). Fgf8 lineage was found for the superficial ectoderm forming the epidermis (Ep) along the lateral (LNP) and medial nasal process (MNP). Note that neurons positive for Fgf8 lineage were detected in the developing forebrain (arrowheads, Br). E, Immunostaining against YFP (brown) and tubulin III (gray-blue); use boxed area in F as reference. No neuronal Fgf8 lineage was detected in the lateral or medial neuroepithelium. F–I, E15.5 Fgf8Cre/RosaYFP embryo. F, YFP expression (brown) was detected in the facial epidermis (Ep), rostral RE, and oral mucosa (OM), but not in the OE and VNO. YFP (green) and tubulin+HuCD immunostaining (red) confirmed lack of Fgf8 lineage in the OE (G) and VNO (H). I, YFP expression (brown) was detected in the oral mucosa (OM) and dental enamel (TE), but not the tongue (Tng). Scale bars: A, D, F, I, 100 μm; B, C, E, G, H, 50 μm.

Mesenchymal sources of Nog define neurogenic permissive borders. A–F, X-gal enzymatic reaction (blue) on Nog^nullLacz/WT. A, Msx1/2 expression (brown) decreased as Nog (blue) expression increased along the RE and in the nasal mesenchyme. White arrow depicts potential Bmp signal in inducing Nog expression. B, Nog (blue) was expressed by MSX1-positive cells (brown) along the RE while mesenchymal Nog expression defined MSX1-negative cells in the VNO. C, HuC/D immunostaining (brown) confirmed that Nog was in cells along the superficial epidermis (SE, arrowhead), non-neuronal RE (arrowheads) and in the ventral mesenchyme (asterisk). D, Vomeronasal area; immunostaining against transit amplifying progenitor cell marker ASCL1 (brown) and HuC/D (gray) showed similar expression patterns with respect to Nog sources (blue); HuC/D (black arrow) and ASCL1 (white arrowhead) were rarely in Nog-expressing areas, but were proximal to the Nog-expressing mesenchyme (asterisk). E–I, E11.5 coronal sections, orientation indicated in E. D, Dorsal; V, ventral; L, lateral; M, medial. E, In situ hybridization against Bmp4 highlighted Bmp4 sources in lateral, ventrolateral, and ventromedial mesenchyme (white arrows). Bmp4 was also expressed by mesenchyme proximal to the RE and by the RE (black arrowhead). F, X-gal staining (blue) revealed Nog expression in proximity to the lateral and ventromedial sources of Bmp4 (compare with E) as well as along the Bmp4-expressing RE (black arrowhead). G, Immunostaining highlighting MSX1 expression (black) in response to Bmp signaling in the mesenchyme and rostral RE. Nog expression (blue; red circles) was found in the dorsolateral and ventromedial nasal mesenchyme. H, HuC/D immunostaining (brown) indicated that neuron formation in the dorsolateral OE coincided with Nog expression in the mesenchyme (red arrow), while neuron formation in the VNO occurred proximal to the medioventral source of Nog (lower red circle). I, Immunostaining for the transcription factor AP2α revealed expression mainly in the non-neurogenic epithelium (use F as reference). AP2α levels decreased in proximity of the dorsolateral and ventromedial source of Nog (compare with G, H; lower red circle). J, Schematic representing relation between Bmp4 expression that induces Nog expression (red) in the nasal mesenchyme. By silencing Bmp signaling, Nog expression defines the neurogenic permissive areas of the OP. Markers related to neurogenic (green arrow) or epithelial fate (black arrow) are indicated. Scale bars, 100 μm.

BMP4 directly induced Nog expression in developing nasal mesenchyme. A–F, Coronal nasal explant from Nog^nullLacz/WT. A, B, Control bead soaked in BSA did not induce Nog expression in the surrounding tissue. C–F, BMP4-soaked bead placed on lateral (C, D) or midline (E, F) mesenchymal tissue of explants from Nog-LacZ mice induced Nog expression. Evidence of BMP4-induced expression of Nog was present regardless of bead position (see arrows in C, D, E vs internal negative control arrowheads in C, D). Endogenous levels of Nog served as internal controls in these experiments. BSA-soaked beads placed on Nog-LacZ explant tissue (A, B) as well as BMP4-soaked beads placed on explants from WT mice (G, H) served as negative controls and showed no induction of Nog:β-gal after X-gal staining. Bead is marked by asterisk in all images. High-magnification images are shown in B–H.

GnRH neurons form at the border between non-neuronal and neuronal epithelium in the developing OP. A–C, X-gal and GnRH immunostaining on serial coronal sections of E11.5 Nog^nullLacz/WT embryo highlighting Nog (blue) and GnRH cells (brown). From rostral (A) to caudal (C), GnRH neurons originated only in ventral portions of the OP (circled), adjacent to the ventral mesenchymal source of Nog (asterisk). D, Scatter plot representing distribution of GnRH neurons (red) in four consecutive sections of Nog^nullLacz/WT embryo in relation to mesenchymal source of Nog (blue). E–J, E11.5 coronal sections. E, GnRH neurons (brown) originate in a ∼250 mm area (bracket) that spans from the end of the Nog-expressing epithelium (blue dots) into the developing VNO and faces the mesenchymal source of Nog (asterisk). No GnRH neurons form in the lateral ventral OP (white arrowhead), a region that is not adjacent to Nog-expressing mesenchyme. F, PAX6 (brown) is not expressed in, or is expressed at very low levels along, the GnRH niche (bracket; compare with E). G, AP2α (black) is expressed in the GnRH neurogenic niche. AP2α expression decreased as PAX6 levels increased (compare with F). H, SOX2 (brown) levels increased along the GnRH niche (brackets), in correspondence to the ventral source of Nog (blue, asterisk). I, The GnRH niche (bracket, compare with J) was between the MSX1/Nog-positive epithelium and the MSX1-negative neurogenic VNO. MSX1 levels decreased coincident with mesenchymal Nog expression (asterisk). J, GnRH immunostaining (brown) as reference for H and I. K, Schematic summarizing molecular expression along the epidermal RE (1), transitional area (2), GnRH neurogenic area (3), and VNO (4). Molecules expressed in each area are listed. Color bars on right side indicate environmental factors associated with the transition from epidermal (1) to vomeronasal (4; +, high expression levels; +/−, low/nonuniform expression; −, nonexpressed/below detection). Scale bars, 100 μm.

Fgf8 hypomorphs display complete loss of medioventral neurogenic milieus only in homozygosity. A–F, E11.5 parasagittal sections: HuC/D immunostaining on WT (A, C) and Fgf8^neo/neo hypomorph (B, D). In control animals (A, C), positive neurons were detected in the OE (between white arrowheads) and VNO (black arrowhead). In Fgf8 hypomorphs (B, D) neurogenesis was limited to the OE (white arrowheads), no neurons were detected in the developing VNO (C vs D, black arrowheads). Similar results were observed comparing Fgf8^WT/nullLacz(E) to Fgf8^neo/nullLacz (F). No neurons were detected in the primordial VNO of either of the Fgf8 mutants (B, D, F). G–I, E11.5 immunostaining for GnRH showed comparable neurons (brown; black arrows) proximal to the VNO in wild-type (G), heterozygous Fgf8^WT/neo (H), and Fgf8^WT/null (I) embryos. J, Quantification of GnRH neurons in the nasal pit of all three genotypes showed no statistical differences (Fgf8^WT/WT vs Fgf8^WT/neo, p = 0.4; Fgf8^WT/WT vs Fgf8^{WT/null LacZ}, p = 0.9). Scale bars: A, B, E, F, 100 μm; C, D, G, H, I, 50 μm.

Changes in Fgf8 expression levels alter Bmp4, Nog expression in the mesenchyme and neurogenic pattern in the OP. Left, E11.5 control mice. Right, E11.5 Fgf8^neo/nullLacz hypomorphs. A, B, HuC/D-positive cells (between arrowheads) identify the neurogenic areas in the nasal pit of control (A) and hypomorph (B). C, D, In situ hybridization for Bmp4. Bmp4 expression in the mesenchyme is indicated by arrows. In hypomorphs (D) a ventrocaudal expansion of Bmp4 expression correlated with lack of neuronal markers in the ventral portion of the OP (compare A, C to B, D). In Fgf8 mutants, Bmp4 expression along the RE appeared reduced compared with controls. E, F, p-SMAD 1,5,8 immunostaining on hypomorphs (F) showed a ventrocaudal expansion of mesenchymal Bmp4 signaling juxtaposed to the developing VNO and within the putative VNO (compare arrows). G, H, Nog in situ hybridization showed Nog expression in Fgf8 mutants was absent in the ventral mesenchyme juxtaposed to the VNO, but was detectable in a more dorsal portion of the developing pit (H, arrows). Expansion of Bmp4 expression (C vs D) in the hypomorphs correlated with changes in Nog expression in the nasal area (G vs H). Scale bars: A–D, 100 μm; E–H, 50 μm. I, J, Schematics summarizing results, with Bmp4 (gray) inducing mesenchymal Nog expression (red) that subsequently controls neurogenesis (green). The different pattern of Bmp and Nog expression translates in altered neurogenesis in Fgf8 mutants.

Expression pattern of Bmp antagonists in nasal mesenchyme determines precursor identity and neurogenic potential of the OP. Coronal sections. Orientation indicated in A. D, Dorsal; V, ventral; L, lateral; M, medial. A–L, Sections from Nog^nullLacz/WT control and Fgf8^neo/neo/Nog^nullLacz/WT mice showing neuron formation (HuC/D), Bmp4 signaling (MSX1), PAX6, SOX2, AP2α, and GnRH in relation to mesenchymal Nog (β-gal) expression (blue, arrows). In control animals (A) neuron formation (HuC/D brown, dorsal to dashed line) in the OE and VNO was defined by dorsolateral and medioventral sources of Nog (arrows). In Fgf8 hypomorphs (B) the medioventral source of Nog was found dorsalized with consequent loss of ventral neurogenesis in the VNO. C, D, MSX1/2 immunostaining (black) highlights the relation between mesenchymal Bmp signaling and Nog expression in control (C) and Fgf8 mutants (D). Reduction in MSX1 expression was noted in the RE (arrowheads) in mutant animals (D) compared with controls (C), while altered expression pattern was observed in the ventromedial mesenchyme with reduced expression proximal to the OP and broader expression in the medial mesenchyme. E, F, In control animals (E), PAX6 expression was high in the neurogenic areas of the pit (dorsal to dashed line; compare with A) in both OE and VNO proximal to the Nog sources (arrows) and distal to Bmp4 sources. In Fgf8 mutants (F), PAX6 expression was still detected in the medioventral OP where the VNO normally forms, though neurogenesis did not occur (arrow; compare with A). G, H, SOX2 expression was found to follow mesenchymal Nog expression in both controls (G) and Fgf8 hypomorphs (H). In the latter, high SOX2 levels were only detected in the areas proximal to Nog sources. Thus, Nog expression correlated with the neurogenic pattern. No SOX2 expression was found within, or proximal to, the developing VNO, where the GnRH normally form. I, J, Reduced AP2α expression was observed in the ventromedial RE of Fgf8 mutants (compare I, arrowhead, with J) while no obvious differences were observed in the lateral RE where transition from AP2α+ to HuC/D (compare with B) was similar to the controls (E, A). K, L, In control animals, the GnRH niche (K, arrowhead in circle) was found facing the mesenchymal source of Nog (arrow). In Fgf8 mutants GnRH niche (L, circled) was far from the mesenchymal source of Nog (arrow). Scale bars: (in A) A–J; (in K) K, L, 100 μm.