Wnt-Responsive Lgr5⁺ Globose Basal Cells Function as Multipotent Olfactory Epithelium Progenitor Cells

Lgr5 marks GBCs in the olfactory epithelium

We performed quantitative RT-PCR to investigate the spatial and temporal expression patterns of Lgr5 mRNA in neonatal and adult olfactory epithelium (Fig. 1A). Expression of Lgr5 was detected in neonatal olfactory epithelium and at a reduced level in adulthood. Interestingly, the pattern of Lgr5 expression paralleled the expression of Ascl1 (a neural progenitor cell marker) at each time point. This expression pattern suggests a potential role of Lgr5 in olfactory epithelium development and proliferation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Lgr5 marks GBCs in the olfactory epithelium. A, Quantitative analysis of Lgr5 expression in the olfactory epithelium by real-time PCR. B, C, Confocal images of Lgr5-EGFP⁺ cells in the olfactory epithelium of mice at P2 (B) and 3 months (3M) of age (C). Images presented are from olfactory epithelium coronal sections stained with an anti-EGFP antibody to amplify the endogenous EGFP signal. Lgr5-EGFP⁺ cells in the basal layer are marked by arrowheads. D, E, Coimmunohistological staining of EGFP with HBC markers, K14 (D) and ICAM-1 (E). Lgr5-EGFP⁺ cells always rest directly on the ICAM-1⁺ or K14⁺ HBCs. F–I, Coexpression of GBC markers Ascl1 or NeuroD1 with Lgr5-EGFP in P2 (F, H) and adult olfactory epithelium (G, I). Many Lgr5-EGFP-positive cells lost Ascl1 expression (G), and NeuroD1 could hardly be detected in 3-month-old animals (I). Boxed areas are highlighted on the right. Lgr5-EGFP⁺Ascl1⁺ cells and Lgr5-EGFP⁺NeuroD1⁺ cells are marked by arrowheads. J, The percentage of Ascl1⁺ and NeuroD1⁺ cells in Lgr5-EGFP⁺ population. The mean ± SD of three independent experiments is presented. Scale bars: B–I, 50 μm.

To characterize Lgr5 expression in more detail, we took advantage of the Lgr5^{EGFP-Ires-CreERT2} knock-in mouse. In this strain, the endogenous Lgr5 promoter controls expression of EGFP and the modified Cre recombinase CreERT2 to faithfully represent the endogenous Lgr5 expression (Barker et al., 2007). Confocal images showed that the Lgr5-driven EGFP signal (referred to as Lgr5-EGFP hereafter) was detectable at E12 (data not shown). Around P2, the majority of Lgr5-EGFP⁺ cells were restricted to the basal compartment (Fig. 1B), and the number of Lgr5-EGFP⁺ cells reached a peak. Notably, the expression of Lgr5-EGFP was absent in the olfactory sensory neuron layer. Although the number of Lgr5-EGFP⁺ cells decreased gradually over time, Lgr5-EGFP signal remained in the basal layer in adulthood (Fig. 1C). The abundance of Lgr5-EGFP⁺ cells in neonatal olfactory epithelium and their location in the basal layer are consistent with the properties observed for recognized progenitor cells in olfactory epithelium.

Two main cell populations, HBCs and GBCs, reside in the basal layer. HBCs have a flattened morphology and are located directly above the basal lamina. They express markers including Krt5, Krt14, ICAM-1, and P63 (Holbrook et al., 1995; Carter et al., 2004; Fletcher et al., 2011; Packard et al., 2011b). We therefore performed immunostaining to determine which population expresses Lgr5-EGFP in the Lgr5^{EGFP-Ires-CreERT2} knock-in mice. The results showed that Lgr5-EGFP⁺ cells did not express the HBC markers, K14 and ICAM-1 (Fig. 1D,E). Instead, Lgr5-EGFP⁺ cells were localized immediately above the ICAM-1 or K14-positive HBCs, and displayed a distinct round morphology that was more characteristic of the GBC population. Consistent with this, at P2, Lgr5-EGFP⁺ cells frequently coexpress Ascl1 or NeuroD1, markers of GBCs and their immediate progeny differentiated toward the neuronal lineage (Fig. 1F,H). Among the Lgr5-EGFP⁺ population, 39.8% of the cells are Ascl1⁺ and 54.3% of the cells are NeuroD1⁺ (Fig. 1J). In adults, the number of Ascl1⁺ cells in the Lgr5-EGFP⁺ population was significantly reduced in the olfactory epithelium of 3-month-old mice (Fig. 1G,J). Likewise, NeuroD1 is rarely observed in GBCs of adult mice (Fig. 1I,J). These data suggest that Lgr5 is expressed in GBCs in adult mice.

Lgr5⁺ cells are cycling and maintain olfactory epithelium renewal

Based on the established neural progenitor cell characteristics of GBCs, we hypothesized that some Lgr5-EGFP⁺ GBCs would be cycling progenitor cells that maintain olfactory epithelium renewal, similar to its role in the small intestine and skin epithelium (Barker et al., 2007; Jaks et al., 2008). We observed that at P2, nearly all of the Lgr5-EGFP⁺ cells (91.2% of 2520 cells, n = 3) expressed Ki67, a proliferation marker (Fig. 2A,F). Similarly, Lgr5-EGFP⁺ cells also maintained proliferation (81.9% of 1180 cells, n = 3) in adult mice (Fig. 2B,F).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Lgr5⁺ cells proliferate and maintain olfactory epithelium renewal. A, B, Coexpression of Lgr5-EGFP and the proliferation marker Ki67 in P2 (A) and adult (B) olfactory epithelium. Lgr5-EGFP⁺ Ki67⁺ cells are marked by arrowheads. C, EdU labeling demonstrates that the Lgr5⁺ GBCs are proliferating cells in neonatal pups. D, In adult olfactory epithelium, EdU⁺ cells are restricted to the basal layer and mostly are Lgr5-EGFP⁺ GBCs. Pups (P2) or adults (2 months old) were injected with EdU two times and analyzed directly 6 h after the last injection. Arrowheads mark representative Lgr5-EGFP⁺EdU⁺ cells. E, In normal olfactory epithelium, EdU⁺ cells migrated apically after a 2 week chase. F, The percentage of Ki67⁺ and EdU⁺ cells in Lgr5-EGFP⁺ population. Data are given as mean ± SD. G–I, LacZ⁺ cells exhibit morphology of olfactory sensory neurons (G, H), supportive sustentacular cells (H), and Bowman's gland and duct cells (I). Olfactory epithelium cryosections from Lgr5 ^{EGFP-Ires-CreERT2}; Rosa26-LacZ mice were stained with X-gal 1 month after tamoxifen induction. The arrowhead in H marks a sustentacular cell. Typical morphology of Bowman's gland and duct cells was indicated by arrows and arrowheads (I). J, K, Confocal images show that β-gal-positive cells coexpress OMP and CK18. Scale bars: A–E, G–K, 50 μm.

We further validated the activity of Lgr5-EGFP⁺ cell proliferation by EdU (a thymidine analog that labels cycling cells during S-phase) incorporation experiment. As expected, injection of EdU labeled 85.6% (2180 cells, n = 3) and 72.6% (2320 cells, n = 3) of Lgr5-EGFP⁺ cells in P2 (Fig. 2C,F) and adult (Fig. 2D,F) mice, respectively. We also examined the fate of Lgr5-EGFP⁺ cell progeny in adult mice with an EdU pulse and chase experiment. A substantial fraction of the EdU-labeled cells migrated apically and lost the Lgr5-EGFP signal after a 2 week chase (Fig. 2E), suggesting that progeny of Lgr5-EGFP⁺ GBCs have differentiated and migrated into the neuronal zone of the olfactory epithelium. Together, these observations confirmed that Lgr5-EGFP⁺ GBCs represented actively cycling progenitor cells in normal olfactory neuroepithelium.

Previous studies have demonstrated that GBCs as neural progenitor cells directly contribute to olfactory sensory neuron regeneration (Caggiano et al., 1994; Jang et al., 2003). Transplantation of GBC-2-positive FACS-sorted GBCs into a lesioned, recipient olfactory epithelium suggested a multilineage differentiation potential including all epithelial cell types (Chen et al., 2004). However, the regenerative capacity of GBC in unlesioned olfactory epithelium remains to be further investigated. To track the Lgr5-EGFP⁺ cell fate in vivo, we crossed the Lgr5^{EGFP-Ires-CreERT2} mouse to the Rosa26-LacZ reporter strain, which allows us to mark permanently Lgr5-EGFP⁺ cells and their progeny with the LacZ reporter transgene by tamoxifen injection.

We injected double-heterozygous mice with tamoxifen at P3–P4 and then analyzed them 3–4 weeks later. X-gal staining showed that LacZ⁺ cells formed clusters and contained the major cell types of the olfactory neuroepithelium (Fig. 2G–I). In most cases, LacZ⁺ clusters contained two to six cells and individual cells appeared separate and discrete (Fig. 2G). However, compact clusters containing >6 cells could be detected occasionally. From the apical to the basal layers, LacZ⁺ descendants of Lgr5⁺ cells presented morphological characteristics of sustentacular cells, olfactory sensory neurons, and GBCs (Fig. 2H). In addition, LacZ-labeled cells with unique morphology of Bowman's glands and ducts were also detected (Fig. 2I). However, no LacZ⁺ cells were observed in the most basal layer of the olfactory epithelium. Only a small fraction of Lgr5⁺ cells (<1%) was labeled with LacZ in our experiments. The frequency of LacZ⁺ cells dramatically decreased when tamoxifen was injected at adulthood, and we did not detect the formation of LacZ⁺ cluster. The limited labeling efficiency can be explained by the weak expression of Lgr5 in neonatal and the cell proliferation further slowdown in adult. We observed no recombination in control littermates without tamoxifen treatment as previously reported (Barker et al., 2010).

The multipotent regenerative capacity of Lgr5⁺ cells was further confirmed by immunocolocalization with olfactory epithelium cell lineage-specific markers. As evidenced by confocal imaging, LacZ⁺ cells were frequently colabeled with the olfactory marker protein OMP, suggesting that Lgr5⁺ cells gave rise to mature olfactory sensory neurons (Fig. 2J). In addition, LacZ⁺ cells also coexpressed cytokeratin 18 (CK18), a marker of sustentacular cells (Fig. 2K). The apical location and distinct morphology further supported the sustentacular cell identification. However, we did not detect LacZ-positive cells in the HBC layer (Fig. 2G–K). Together, these results suggest that a subset of Lgr5⁺ GBCs function as multipotent progenitors that can generate the major cell types of the olfactory epithelium under normal physiological conditions.

Lgr5⁺ cells regenerate olfactory neuroepithelial lineages after methimazole injury

The remarkable regenerative capacity of olfactory epithelium has been well documented. Damage of the olfactory epithelium with the antithyroid drug methimazole has been identified as a reliable olfactory lesion model (Genter et al., 1995; Leung et al., 2007). To identify the characteristics of Lgr5⁺ cells after severe lesion, we injected 3- to 4-week-old Lgr5 ^{EGFP-Ires-CreERT2} mice with methimazole and analyzed Lgr5-EGFP⁺ cells and their progeny at early stages of regeneration. In striking contrast to nonlesioned controls, robust proliferation of Lgr5-EGFP⁺ cells was activated throughout the neuroepithelium at postinjury day 2 (Fig. 3A,D). During the dynamic process of regeneration, almost all Lgr5-EGFP⁺ cells coexpressed the proliferation marker Ki67. The proliferation of Lgr5-EGFP⁺ cells peaked at days 5–7 and extended to day 14 (Fig. 3A,D,E). As evidenced by double staining, most Lgr5-EGFP⁺ cells (63.5% of 820 cells at day 2; 56.5% of 1050 cells at day 5) coexpressed the GBC marker Ascl1 (Fig. 3B,E). At day 14, the frequency of Ascl1-labeled Lgr5-EGFP⁺ cells was clearly diminished. In addition, Lgr5-EGFP⁺ cells were always restricted to the GBC layer immediately above the K14-positive HBCs without overlap (Fig. 3C). These results demonstrate that the Lgr5-EGFP⁺ GBC population is readily activated to proliferate after olfactory epithelium lesion.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Lgr5⁺ cells regenerate multiple olfactory epithelium lineages after methimazole injury. A–C, Robust proliferation of Lgr5-EGFP⁺ cells was activated in the early stage of regeneration. Two-month-old Lgr5^{EGFP-Ires-CreERT2} mice were injected with methimazole to induce olfactory epithelium damage. At day 2, Lgr5-EGFP⁺ cells were activated to proliferate and usually coexpressed Ki67. At postlesion days 5–7, the number of Lgr5-EGFP⁺ cells reached a peak, and abundant proliferation continued to day 14 (A). B, Lgr5-EGFP⁺ cells colabeled with GBC marker Ascl1 in the very early stage of regeneration, but a significant reduction of Ascl1⁺ cells was observed at postlesion day 14 (B). C, Lgr5-EGFP is expressed exclusively by the GBCs that are located above the K14⁺ HBCs. Arrowheads mark Lgr5-EGFP⁺Ki67⁺ cells in A and Lgr5-EGFP⁺Ascl1⁺ cells in B. D, Quantitative analysis of Lgr5-EGFP⁺ cells during regeneration. E, The percentage of Ki67⁺ and Ascl1⁺ cells in the Lgr5-EGFP⁺ population. The data are shown as mean ± SD from three experiments. F, Genetic lineage tracing using 1-month-old Lgr5^{EGFP-Ires-CreERT2}; Rosa26-LacZ mice. G–I, Lgr5⁺ cells regenerate multiple cell types in the olfactory epithelium. LacZ⁺ cells are present in basal layer, olfactory sensory neurons, sustentacular cells (G), Bowman's gland and ducts (H), and microvillous cells (I). J–L, Confocal images of olfactory epithelium sections double stained with β-gal and OMP (J) or CK18 (K, L). β-gal signal colocalizes with both OMP and CK18. M, No colocalization was observed between β-gal and CK14 in HBCs. N, EdU labeling (arrow) remained within Lgr5-EGFP⁺ GBCs after methimazole-induced lesion followed by a 2 week EdU chase. The arrowhead marks an Lgr5-EGFP⁺EdU⁺ cell. O, LacZ⁺ cells are retained in basal layer 6 months after tamoxifen induction. Scale bars: A–C, G, H, J–O, 50 μm; I, 20 μm.

To more directly determine the contribution of Lgr5-EGFP⁺ GBCs to olfactory epithelium lineages, we traced the Lgr5-EGFP⁺ cells after methimazole-induced lesion in Lgr5^{EGFP-Ires-CreERT2}; Rosa26-LacZ mice (Fig. 3F). At 30 d after tamoxifen injection, LacZ⁺ clusters are distributed widely in the septum and turbinate neuroepithelium (Fig. 3G). Compared with nonlesioned mice, both the number of LacZ⁺ clusters and the number of cells in each cluster were dramatically increased, suggesting the progeny of Lgr5⁺ cells had undergone extensive proliferation and some of them had divided multiple rounds during regeneration.

As judged by their position and morphology, LacZ⁺ clusters frequently contained GBCs, olfactory sensory neurons, sustentacular cells, and cells of the Bowman's glands and ducts (Fig. 3G–I). Double staining demonstrated that the LacZ-labeled cells generated OMP⁺ mature olfactory sensory neurons (Fig. 3J), CK18⁺ Bowman's glands and ducts (Fig. 3K), and sustentacular cells (Fig. 3L). In the regenerated olfactory epithelium, we did not detect LacZ⁺ cells in the basal layer that displayed HBC morphology or colocalized with K14 (Fig. 3M). These observations strongly suggest that Lgr5⁺ cells are multipotent progenitor cells that contribute to most olfactory neuroepithelial lineages (except HBCs) after extensive depletion.

We observed that Lgr5-EGFP⁺ cells retained EdU labeling for at least 2 weeks after injection (Fig. 3N). These results are consistent with a recent finding that GBCs contain BrdU or EdU label-retaining cells in neonatal or lesioned-recovering olfactory epithelium (Jang et al., 2014). Furthermore, despite the decrease in the number of LacZ⁺ clusters as the neuroepithelium turned over, sporadic LacZ⁺ cells in the GBC layer could be found 6 months after tamoxifen exposure (Fig. 3O), suggesting the character of slow cycling progenitor cells. Together, these observations confirmed that Lgr5-marked GBCs contain not only actively cycling but also relatively quiescent progenitor cells, the latter of which can only be activated by lesion.

Differentiation of Lgr5⁺ cells in vitro

Based on their easy accessibility, previous studies have explored the possibility of using olfactory epithelium progenitors as a source of neural stem cells for neural repair (Jang et al., 2008; Krolewski et al., 2011). To better characterize these Lgr5⁺ olfactory stem cells in vitro, we dissociated the neuroepithelium from Lgr5^{EGFP-Ires-CreERT2} mice at P2 and cultured the cells in NSC medium. Interestingly, the Lgr5-EGFP signal was always present in a high fraction of cells in the flattened-compact colony (43 ± 3%) or floating neurosphere (39 ± 2%; Fig. 4A,B). Confocal images revealed the Lgr5⁺ cells were also Ki67⁺, suggesting that the colonies were made up of proliferative cells (Fig. 4C). Lgr5⁺ colonies or neurospheres rarely contained cells that express the HBC marker K14 (data not shown). In addition, Lgr5⁺ colonies are Nestin negative (data not shown), an intermediate filament protein, consistent with recent finding that Nestin marked progenitors only in embryonic olfactory epithelium (Murdoch and Roskams, 2008).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Proliferation and differentiation of Lgr5⁺ olfactory epithelium cell in vitro. A–C, In NSC medium, olfactory epithelium cells formed compact adhesive colonies (A) or neurosphere (B). The majority of cells in these colonies was Lgr5-EGFP⁺ and coexpressed the proliferation marker Ki67(C). Some Lgr5-EGFP⁺Ki67⁺ cells are highlighted by arrowheads (C). D–G, Sorted Lgr5-EGFP⁺ cells generated different cell types of the olfactory neuroepithelium. After differentiation in neural medium for 5 d, progeny of sorted Lgr5-EGFP⁺ cells expressed Tuj1 (E), DCX (F), and CK18 (G), markers of differentiated olfactory epithelium cells. Scale bars: A, B, F, G, 50 μm; C, E, 100 μm.

We next performed FACS to isolate Lgr5-EGFP⁺ cells and examined their differentiation potential in vitro (Fig. 4D). By 5 d differentiation, Lgr5-EGFP⁺ cells give rise to Tuj1⁺ (definitive neuronal-fate marker; Fig. 4E) or DCX⁺ (Fig. 4F) immature olfactory sensory neurons. In addition, CK18⁺ sustentacular cells can also be detected (Fig. 4G). These results further supported the identification of Lgr5⁺ GBCs as multipotent progenitors.

Wnt signaling regulates Lgr5⁺ cell proliferation and neurogenesis

Self-renewal and differentiation of neural stem cells in the olfactory epithelium are controlled by complex mechanisms (Nicolay et al., 2006; Heron et al., 2013). However, the precise cellular and molecular mechanisms governing the stemness of GBCs have not been defined. The β-catenin-mediated canonical Wnt signaling pathway has been widely implicated in maintaining stem cells in a self-renewing state. Binding of Wnt proteins to cell-surface Wnt receptors triggers a cascade of intracellular events, which results in β-catenin accumulation in the cytoplasm and subsequent nuclear translocation (Clevers and Nusse, 2012). β-Catenin interacts with the Tcf/Lef-family transcription factors to activate Wnt target genes.

Since Lgr5 was identified originally as a Wnt target (Barker et al., 2007), we examined the expression of β-catenin in the olfactory epithelium. In adults, almost the entire olfactory epithelium stained positive for β-catenin, but a slight accumulation of β-catenin in the nucleus was only detected in the basal layer (Fig. 5A). Interestingly, the cells that accumulated nuclear β-catenin colabeled with Lgr5-EGFP, suggesting the activation of Wnt signaling in these cells.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Wnt signaling maintains Lgr5⁺ cell proliferation. A, Confocal images of β-catenin staining in the olfactory epithelium. Boxed region was magnified at the bottom. Compared with upper adjacent cells (marked by arrows), Lgr5-EGFP⁺ GBCs (marked with arrowheads) have increased nuclear accumulation of β-catenin. B, Confocal images of Lgr5-EGFP⁺ GBC colony stained with Ki67. Arrowheads mark Lgr5-EGFP⁺Ki67⁺ cells. C, Quantification of Ki67 and Lgr5-EGFP double positive cells in the presence and/or absence of Wnt3a in culture. D–F, Confocal images of the olfactory epithelium stained with antibodies for Ascl1 and β-gal in early stage of regeneration. G, Quantification of Ascl1 and β-gal double positive cells in the olfactory epithelium of mice as shown in D–F. Wnt activation significantly increased the percentage of Lgr5⁺ GBC-derived Ascl1⁺ transit-amplifying neural progenitors (*p = 0.000). H, Confocal images of olfactory epithelium stained with antibodies for OMP, Ki67, and β-gal. Deletion of APC maintained Lgr5⁺ GBC progeny in a proliferative state in the basal layer 1 month after lesion. I, Quantification of β-gal⁺ clusters in the olfactory epithelium of the three mice strains 1 month after injury. Data presented are the numbers of clusters that contains 6 or >6 β-gal⁺ cells (*p = 0.003). Scale bars: A, 25 μm; B, D–F, H, 50 μm. WT, wild-type.

To address whether Lgr5⁺ GBCs in the olfactory epithelium are regulated by Wnt signaling, we first evaluated the potential effect of Wnt3a (a canonical Wnt isoform) on Lgr5⁺ cell proliferation in vitro. As expected, Wnt3a treatment substantially increased the percentage of Lgr5-EGFP and Ki67 double positive cells (Fig. 5B,C), consistent with previous observation that Wnt3a enhances neuroblast proliferation in hippocampal progenitor cells (Lie et al., 2005). This result suggested that Wnt signaling is involved in regulating Lgr5⁺ GBC proliferation.

We next investigated the role of the Wnt/β-catenin signaling pathway in regulating the proliferation and neurogenesis of Lgr5⁺ GBCs during olfactory epithelium regeneration in vivo. We manipulated the Wnt signaling pathway by injecting tamoxifen into the Lgr5^{EGFP-Ires-CreERT2}; Rosa26-LacZ; APC^flox/flox mice and the Lgr5^{EGFP-Ires-CreERT2}; Rosa26-LacZ; β-catenin^flox/flox mice after methimazole-induced neuroepithelium depletion. This strategy leads to APC or β-catenin deletion in newborn Lgr5⁺ cells (or β-gal-labeled cells) and causes Wnt pathway constitutive activation or blocking, respectively (Clevers and Nusse, 2012). We observed that in contrast to wild-type littermates, upregulation of the Wnt/β-catenin signaling pathway by APC deletion apparently elevated the number of Lgr5⁺ GBC-derived Ascl1⁺/β-gal⁺ transit-amplifying neural progenitors on postinjury day 14 (74.2% vs 21.9%; Fig. 5D,F,G), suggesting an increased neurogenesis in the early stage of regeneration. Interestingly, downregulation of Wnt/β-catenin signaling activity by β-catenin deletion caused a dramatic reduction in the number of Ascl1⁺/β-gal⁺ cells (Fig. 5E). Of note, when the tracing period was extended to 1 month, β-gal⁺ clusters with neuroblast morphology could still be detected in the basal layer and many of these cells coexpressed the proliferation marker Ki67 (Fig. 5H). Furthermore, APC deletion clearly increased the number of β-gal⁺ clusters (Fig. 5I) as well as the number of cells in each cluster (data not shown) on postinjury day 30. Together, these data suggest that Wnt activation stimulates Lgr5⁺ GBC proliferation and the proliferation of Lgr5⁺ GBC-derived transit-amplifying neural progenitor cells, leading to augmented neurogenesis in the early stage of olfactory epithelium regeneration.

Sustained Wnt activation delays olfactory sensory neuron maturation

We next sought to determine whether Wnt activation affects the differentiation of Lgr5⁺ GBC progeny in the latter stage of regeneration. Using the same strategy as described in the above experiment, we manipulated Wnt signaling in Lgr5⁺ GBCs following methimazole-induced olfactory epithelium injury and examined its effects on the differentiation of Lgr5⁺ GBC progeny into olfactory sensory neurons and sustentacular cells at postinjury day 30. We observed that in wild-type control mice, 69.9% of Lgr5⁺ GBC-derived β-gal⁺ cells coexpressed the mature olfactory sensory neuron marker OMP, which mainly is expressed in the upper half of the olfactory epithelium (Fig. 6A,D). Upregulation of Wnt signaling caused by APC deletion resulted in a dramatic reduction in the number of β-gal and OMP double positive mature olfactory sensory neurons (20.6% in β-gal⁺ cells; Fig. 6C,D), accompanied by a significant increase of immature olfactory sensory neurons identified by weak or lack of OMP expression (Fig. 6C; 42.4% OMP weak and 35.7% OMP-negative cells). Consistent with this, in APC-deleted olfactory epithelium, the majority of Lgr5⁺ GBC-derived β-gal⁺ cells coexpressed the immature olfactory sensory neuron marker GAP43 (Fig. 6E–H), suggesting that sustained Wnt activation delayed terminal maturation.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Sustained Wnt activation delays olfactory sensory neuron maturation. A–C, Confocal images of olfactory epithelium stained with antibodies specific for OMP and β-gal. Compared with wild-type control (A), downregulation of Wnt activity by β-catenin deletion obviously reduced the number and size of β-gal⁺ clusters (B). In APC-deleted olfactory epithelium, β-gal⁺ clusters were located in the immature or basal layer, and the majority of these cells was OMP negative or had weak OMP expression as highlighted by arrowheads in C. D, Quantitative analysis of OMP⁺ olfactory sensory neurons in β-gal⁺ clusters (*p = 0.001). Due to the low frequency of labeling, we did not calculate the β-gal⁺ cells after β-catenin deletion. E–G, Confocal images of the olfactory epithelium stained with GAP43 and β-gal. β-gal⁺/GAP43⁺ cells were highlighted by arrowheads in G. H, Quantification of GAP43⁺ immature neuron in β-gal⁺ population (*p = 0.003). I–K, Confocal images of the olfactory epithelium stained with antibodies specific for CK18 and β-gal. Compared with wild-type control mice (I), Wnt activation significantly reduced the number of β-gal⁺ sustentacular cells located on the surface layer (K). L, Quantification of CK18⁺ sustentacular cells that express β-gal (*p = 0.008). M, Confocal images of sorted Lgr5-EGFP⁺ GBCs in the presence or absence of Wnt3a. N, Quantification of the percentage of Tuj1-positive and EGFP-positive cells in culture (*p = 0.000; **p = 0.002). Scale bars: A–C, E–G, I–K, 25 μm; M, 50 μm.

However, downregulation of Wnt signaling caused by β-catenin deletion significantly reduced the number of β-gal⁺ clusters and none of the clusters observed contained >6 β-gal⁺ cells (Fig. 6B,F,J), suggesting that neuroregeneration was almost extinguished by downregulation of Wnt activity. These data are consistent with recent observations that downregulation of Wnt signaling inhibits neuronal production in the embryonic cortex (Munji et al., 2011). On the other hand, APC deletion significantly decreased sustentacular differentiation, with only 1.9 ± 2.3% of β-gal-labeled cells colabeled with CK18 (Fig. 6K,L), compared with15.9 ± 6.7% of β-gal⁺ cells differentiated into CK18⁺ sustentacular cells in wide-type mice (Fig. 6I,L). Together, these observations suggest that the Wnt/β-catenin signaling pathway plays an important role in regulating Lgr5⁺ GBC progeny terminal differentiation.

We also examined the effect of Wnt activation on Lgr5⁺ GBC differentiation in vitro. In differentiation media, sorted Lgr5⁺ cells readily differentiated into Tuj1⁺ neurons (Fig. 6M). However, Wnt3a treatment significantly decreased the number of Tuj1⁺ cells while maintaining Lgr5⁺ cells in an undifferentiated state (Fig. 6M,N), further suggesting that Wnt activation prevents Lgr5⁺ cells from differentiating into olfactory sensory neurons. Together, these results support the notion that Wnt signaling regulates olfactory neuroepithelium regeneration not only by controlling the proliferation/self-renewal of Lgr5⁺ GBCs, but also by affecting their subsequent terminal differentiation.