Skn-1a/Pou2f3 is required for the generation of Trpm5-expressing microvillous cells in the mouse main olfactory epithelium

Abstract

Background: The main olfactory epithelium (MOE) in mammals is a specialized organ to detect odorous molecules in the external environment. The MOE consists of four types of cells: olfactory sensory neurons, supporting cells, basal cells, and microvillous cells. Among these, development and function of microvillous cells remain largely unknown. Recent studies have shown that a population of microvillous cells expresses the monovalent cation channel Trpm5 (transient receptor potential channel M5). To examine functional differentiation of Trpm5-expressing microvillous cells in the MOE, we investigated the expression and function of Skn-1a, a POU (Pit-Oct-Unc) transcription factor required for functional differentiation of Trpm5-expressing sweet, umami, and bitter taste bud cells in oropharyngeal epithelium and solitary chemosensory cells in nasal respiratory epithelium.

Results: Skn-1a is expressed in a subset of basal cells and apical non-neuronal cells in the MOE of embryonic and adult mice. Two-color in situ hybridization revealed that a small population of Skn-1a-expressing cells was co-labeled with Mash1/Ascl1 and that most Skn-1a-expressing cells coexpress Trpm5. To investigate whether Skn-1a has an irreplaceable role in the MOE, we analyzed Skn-1a-deficient mice. In the absence of Skn-1a, olfactory sensory neurons differentiate normally except for a limited defect in terminal differentiation in ectoturbinate 2 of some of MOEs examined. In contrast, the impact of Skn-1a deficiency on Trpm5-expressing microvillous cells is much more striking: Trpm5, villin, and choline acetyltransferase, cell markers previously shown to identify Trpm5-expressing microvillous cells, were no longer detectable in Skn-1a-deficient mice. In addition, quantitative analysis demonstrated that the density of superficial microvillous cells was significantly decreased in Skn-1a-deficient mice.

Conclusion: Skn-1a is expressed in a minority of Mash1-positive olfactory progenitor cells and a majority of Trpm5-expressing microvillous cells in the main olfactory epithelium. Loss-of-function mutation of Skn-1a resulted in complete loss of Trpm5-expressing microvillous cells, whereas most of olfactory sensory neurons differentiated normally. Thus, Skn-1a is a critical regulator for the generation of Trpm5-expressing microvillous cells in the main olfactory epithelium in mice.

Figure 1

Expression of Skn-1a in the developing main olfactory epithelia. (A)In situ hybridization with RNA probes for Skn-1a in coronal sections of mouse MOE at embryonic days 13.5 and 16.5 and postnatal days 0, 7, 14, and 30. The expression of Skn-1a was first detected at embryonic day 13.5 and was observed during subsequent development. The Skn-1a-expressing cells were located in apical, intermediate, and basal positions in the MOE during embryonic stages and were gradually restricted to apical and basal positions in postnatal development. (B) The expression of Skn-1a in the rostral-caudal axis of the MOE at postnatal day 7. Skn-1a expression was observed throughout the MOE, in terms of the rostral-caudal and the dorsal-ventral axis. (C) In the adult MOE, Skn-1a-expressing cells were distributed in graded fashion: low density in the dorsomedial region to high density in the lateral region. Left and right images are higher-magnification images of the dorsomedial and lateral regions (the areas enclosed by the dashed boxes in the center image), respectively. (D)In situ hybridization of signaling molecules in SCCs on coronal sections of adult MOE. Expression of Tas1r3, Tas2r105, Tas2r108, Gnat3, and Plcb2 was not observed. Only the signal of Trpm5 mRNA was detected in the superficial layer of the MOE. Scale bars: 50 μm in A and D, 500 μm in B and C.

Figure 2

Characterization of Skn-1a-expressing cells in the main olfactory epithelium. (A)Skn-1a-expressing cells were characterized using two-color in situ hybridization in coronal sections of the MOE at postnatal day 0 with RNA probes for Skn-1a (green) and OSN progenitor/precursor genes Mash1 (neuronal progenitors), Ngn1 (neuronal precursors), and NeuroD (differentiating/postmitotic neurons). Small populations of Skn-1a-potitive cells and Mash1-positive cells overlapped. The arrowhead indicates a co-labeled cell, and arrows indicate either Skn-1a or Mash1 single-labeled cells. None of Skn-1a-positive cells were co-labeled with Ngn1 and NeuroD (arrows). (B and C)In situ hybridization of Skn-1a (green) with OMP (mature OSNs; B, red) and Trpm5 (Trpm5-positive microvillous cells; C, red) in coronal sections of the adult MOE. Neither apical nor basal Skn-1a-expressing cells (arrows) were co-labeled with OMP signals. Trpm5 signals were co-labeled with apical Skn-1a signals (arrowheads) but not with basal Skn-1a signals (arrow). Scale bars, 25 μm. (D and E) Populations of Skn-1a-expressing cells (D) and Mash1-expressing cells (E) were analyzed by two-color in situ hybridization at postnatal day 30. Quantitative analyses revealed that 8.34 ± 2.82% (mean ± SD) of the Skn-1a-expressing cells coexpressed Mash1 (n = 3), and 77.7 ± 5.95% coexpressed Trpm5 (n = 3). In the OSN-lineage, Mash1-positive olfactory progenitors rarely expressed Skn-1a (1.41 ± 0.564%, n = 3).

Figure 3

Effect of Skn-1a deficiency on the differentiation of olfactory sensory neurons. The impact of Skn-1a deficiency on the OSN differentiation was examined by in situ hybridization using OSN neuronal marker genes Mash1 (neuronal progenitors), Ngn1 (neuronal precursors), NeuroD (differentiating/postmitotic neurons), GAP43 (immature neurons), and OMP (mature neurons) in coronal sections of wild-type and Skn-1a^-/- mice at postnatal day 7. (A) No obvious differences in the expression of marker genes were observed between Skn-1a^-/- and wild-type mice in most cases. (B) Examples of the Skn-1a^-/- mice showing a partial but obvious phenotype of a defective differentiation of OSNs only in the specific region of ectoturbinate 2 at postnatal day 7 (upper panels). Expression of GAP43 and OMP was greatly suppressed, whereas expression of Mash1, Ngn1, and NeuroD was upregulated (lower panels: high magnification images of the dotted boxes). Scale bars, 500 μm.

Figure 4

Expression of Skn-1a and Trpm5 in the MOE of Mash1^-/- embryos. Expression of Skn-1a and Trpm5 in the Mash1^-/- MOE was examined by in situ hybridization at embryonic day 18.5. The MOE of Mash1^-/- embryos appeared smaller and thinner than that of wild-type littermates, as observed previously. Expression of either Skn-1a or Trpm5 was observed in both the wild-type and Mash1^-/- MOE. Higher-magnification images of the dotted boxes are presented to the right of each image. Scale bars, 100 μm.

Figure 5

Effect of Skn-1a deficiency on the functional differentiation of Trpm5-positive microvillous cell. (A)In situ hybridization of Trpm5 on coronal sections of the MOE of wild-type and Skn-1a^-/- mice. The mRNA signal of Trpm5 was absent in Skn-1a^-/- mice. (B and C) Coronal sections of wild-type and Skn-1a^-/- MOE of adult mice were immunostained with an anti-Trpm5 antibody (green) and an anti-villin (B) or anti-ChAT (C) antibody (red). Trpm5-positive cells were villin positive in the microvilli in the wild-type MOE (arrowheads), whereas no immunoreactive signal for Trpm5 or villin was observed in the Skn-1a^-/- MOE. Trpm5-positive cells were co-immunostained with anti-ChAT antibody in wild-type (arrowheads) but not in Skn-1a^-/- mice. Scale bars: 100 μm in A, 10 μm in B and C.

Figure 6

Quantification of microvillous cell density in the most superficial layer of the MOE. (A) Image of an MOE dorsal recess from a ChAT-eGFP mouse, showing ChAT/Trpm5-expressing microvillous cells (GFP⁺) in the most superficial layer, a region above the supporting cell nuclei. (B) A higher-magnification view of the DAPI-stained nuclei in the dorsal MOE. Arrowheads point to nuclei of GFP⁺ microvillous cells. (B’) Overlay of GFP signal onto B. (C) Image of an MOE dorsal recess from an Skn-1a^-/- mouse. Arrows in B and C point to nuclei that do not belong to GFP⁺ microvillous cells. (D) Plot of the averaged density per surface area of DAPI-stained nuclei and GFP⁺ cells in the most superficial layer of the MOE from ChAT-eGFP mice. Counting was conducted from the dorsal recess and septum of the MOE. Approximately 80% of the cells in the area are GFP⁺ microvillous cells. (E) Comparison of averaged nucleus density, showing approximately 73% reduction in the nucleus density of Skn-1a^-/- mice compared with that of ChAT-eGFP mice. Scale bars: 100 μm in A, 20 μm in B-D.