Temporal changes in NCAM immunoreactivity during taste cell differentiation and cell lineage relationships in taste buds

Abstract

Neural cell adhesion molecule (NCAM) is a type III cell marker in the taste buds. In order to clarify the cell type of Mash1-expressing cells in taste buds, expression of NCAM was examined in Mash1-expressing taste cells of adult mice in comparison with gustducin- and T1r3-expressing cells, using a combination of NCAM immunohistochemistry and in situ hybridization. About 98% of Mash1-expressing cells were NCAM immunopositive (IP), suggesting that Mash1-expressing cells should be categorized as type III cells. Unexpectedly, small subsets of gustducin- and T1r3-expressing cells were also found to be NCAM-IP, contradicting previous immunohistochemical studies in rats, in which gustducin-IP cells were observed specifically in type II cells, which do not have NCAM immunoreactivity. Examinations of developing taste buds showed temporal changes in the ratio of NCAM-IP cells in gustducin- and T1r3-expressing cells; the ratio of NCAM-IP cells in these gene-expressing cells were ∼90% at 0.5 days after birth and decreased markedly during development. In contrast, the majority of Mash1-expressing cells showed constant NCAM immunoreactivity throughout development. In addition, BrdU-labeling experiments showed that the differentiation of Mash1-expressing cells precedes those of gustducin- and T1r3-expressing cells in taste buds of adult mice. These results suggest that T1r3- and gustducin-expressing cells are NCAM-IP at the beginning of cell differentiation, and that NCAM immunoreactivity in gustducin- and T1r3-expressing cells might remain from the previous developmental stage expressing Mash1.

Introduction

Taste buds are the sensory organs for taste and are maintained under continuous cell renewal, even in adults. Taste cells are constantly differentiated from local epithelia (Stone et al., 1995), and their average life span is estimated to be ∼10 days (Beidler and Smallman, 1965; Farbman, 1980). Therefore, taste buds consist of cells at various maturation stages of taste cell differentiation. Recently, expression of bHLH transcription factors, Mash1 and NeuroD, was reported in mouse taste buds (Kusakabe et al., 2002; Suzuki et al., 2002). The sequential expression of Mash1 followed by NeuroD is reported in the differentiation of olfactory neurons (Cau et al., 1997, 2002). These transcription factors might be related to a regulatory gene cascade and the differentiation of taste cells. Based on ultrastructural characteristics, taste bud cells have been classified into several cell types: basal cells, type I (dark) cells, type II (light) cells and type III (intermediate) cells (Kinnamon et al., 1985; Delay et al., 1986; Yee et al., 2001; Clapp et al., 2004). Immunohistochemical studies in rats have shown histochemical markers associated with cell types. The H blood group antigen is predominant in type I cells (Pumplin et al., 1999). Gustducin is reported to be expressed in a subset of type II cells (Boughter et al., 1997; Yang et al., 2000), but a marker gene expressed in all type II cells has not been reported. Although there has been debate regarding the criteria of type III (intermediate) cells in rodents, a cell type that is well defined in rabbits, recent immunohistochemical studies in rats distinguished type III cells in combination with ultrastructural analysis, and showed the relationship among immunohistochemical markers for type III cells (Yee et al., 2001). The majority of type III cells predominantly expresses neural cell adhesion molecule (NCAM), while serotonin-positive cells were observed in a subset of type III cells, and ubiquitin carboxyl terminal hydrolase (PGP 9.5) was expressed in a subset of both type II and type III cells. It is generally accepted that type I, II and III cells are derived from basal cells; however, the cell lineage relationships among the cell types in taste buds remain unclear. Conflicting views have been published: (i) taste buds are composed of one cell lineage, and morphological changes occur transiting from basal to dark (type I) to intermediate (type III) to light (type II) cells during maturation (Delay et al., 1986); (ii) taste buds are composed of at least two cell lineages, which belong to different cell types, an idea based on the difference in life span and morphology between dark cells and light cells (Farbman, 1980; Pumplin et al., 1997). As for regulatory gene expression, a previous report showed that NeuroD-expressing cells express gustducin but not NCAM, suggesting that NeuroD is expressed in type II cells (Suzuki et al., 2002). On the other hand, we have reported the segregated expression pattern of Mash1 and gustducin (Kusakabe et al., 2002), suggesting that Mash1-expressing cells may be of another cell type than type II cells. However, the cell type of Mash1-expressing cells remains unclear. It is expected that clarification of the relationship between the transcription factors and cell types in taste buds will provide a clue to cell lineage relationships in taste buds.

In the present report, we first examined NCAM expression in Mash1-expressing cells in adult mice to clarify the cell type, using a combination of NCAM immunohistochemistry and in situ hybridization. Almost all Mash1-expressing cells were NCAM-IP, suggesting that Mash1-expressing cells should be categorized as type III cells. Unexpectedly, NCAM immunoreactivity was also observed in small subsets of gustducin- and T1r3-expressing cells, contradicting the previous reports in rats (Yang et al., 2000). NCAM immunoreactivity was then examined in Mash1-, gustducin- and T1r3-expressing cells during development. In addition, the BrdU-labeling experiments were performed to examine the timing of the differentiation of Mash1-, gustducin- and T1r3-expressing cells in adult mice.

Materials and methods

Experimental animals

The animals used in this study as adults were 10- to 12-week-old C57BL/6N mice. These animals were initially purchased from Charles River Japan (Yokohama, Japan), and were bred at the National Food Research Institute. We followed the guidelines of our institute for the care and use of experimental animals.

NCAM immunohistochemistry after in situ hybridization

Mouse tongues were dissected, placed in embedding compound (OCT; Sakura Finetech., USA, Los Angeles, CA), and rapidly frozen in a liquid nitrogen bath. Tissues were sectioned at 5 μm, and in situ hybridization was performed as previously described (Miura et al., 2004). Antisense RNA probes were transcribed in vitro with digoxigenin-UTP using an RNA transcription kit (Roche Diagnostics GmbH, Mannheim, Germany) from a linearized plasmid containing one of the following cDNA inserts: Mash1 (10012: Genbank U68534-760: Genbank M65603), gustducin (Kim et al., 2003), T1r3 (Kitagawa et al., 2001) and Shh (Kitamura et al., 1997). Following hybridization, sections were incubated with an alkaline phosphatase anti-digoxigenin goat antibody (Roche Diagnostics GmbH) and anti-NCAM rabbit polyclonal antibody (2 μg/ml; Chemicon, Temecula, CA) in TBSB [Tris-buffred saline, TBS (100 mM Tris–HCl, pH 7.5, 150 mM NaCl) containing 1% blocking reagent (Roche Diagnostics GmbH)] at 4°C overnight. After they were washed in TBST (TBS containing 0.05% Tween 20), the signal for in situ hybridization was detected using 2-hydroxy-3-naphthoic acid-2′-phenylanilidephosphate (HNPP)/Fast Red alkaline phosphatase substrate (Roche Diagnostics GmbH). Sections were rinsed in TBS, incubated with a biotinylated anti-rabbit IgG (10 μg/ml; Vector, Burlingame, CA) in TBSB for 2 h at room temperature and then washed in TBST. To detect the signal for NCAM, sections were incubated with streptavidin–Alexa 488 (Molecular Probes, Eugene, OR) for 30 min and washed in TBST. Double color fluorescent images were merged using Photoshop (Adobe Systems, San Jose, CA). Signals in every fourth section were analyzed in order to avoid any double counting of cells. NCAM immunoreactivity adjacent to the nucleus of taste bud cells was counted as the signal indicating its expression in order to avoid counting the signals for nerve fibers as taste cell signals. Sections from one or two circumvallate papillae were analyzed to count the number of NCAM-positive cells in Mash1-, gustducin- and T1r3-expressing cells. In the cases of gustducin- and T1r3-expressing cells at 0.5 days after birth, the data were collected from six circumvallate papillae because the number of positive cells for these genes were very small: 37 gustducin- and 22 T1r3-expressing cells were found in counts of six whole circumvallate papillae.

BrdU labeling and the detection after in situ hybridization

Adult mice were injected intraperitoneally with bromodeoxyuridine (BrdU, 50 mg/kg; Roche Diagnostics GmbH). Mouse tongues were dissected and treated in the same manner as that used for in situ hybridization at 3 and 12 h and 1, 2, 3 and 4 days after BrdU injection. After hybridization with the digoxigenin-labelled cRNA probe, signals for in situ hybridization were obtained by incubation with HNPP/Fast Red alkaline phosphatase substrate (Roche Diagnostics GmbH). The sections were washed in TBS, incubated in TBSB for 15 min and washed again in PBS. The sections were then incubated with anti-BrdU monoclonal antibody (Roche Diagnostics GmbH) for 30 min at 37°C according to the instructions for use of the BrdU Labeling and Detection Kit II (Roche Diagnostics GmbH). After washing in PBS, the sections were incubated with an Alexa 488–anti mouse IgG (Molecular Probes) for 2 h at room temperature and washed in TBS. Double-color fluorescent images were merged using Photoshop (Adobe Systems). Signals in every fourth section were analyzed in order to avoid any double counting of the gene-expressing cells.

BrdU detection after NCAM immunohistochemistry

Sections were incubated with anti-NCAM rabbit polyclonal antibody (2 μg/ml; Chemicon, Temecula, CA) overnight, washed and then incubated with biotinylated anti-rabbit IgG (10 μg/ml; Vector, Burlingame, CA). After washing, they were incubated with streptavidin–Alexa 568 (Molecular Probes, Eugene, Oregon) for 30 min and washed. The sections were incubated at 65°C for 90 min, and BrdU signals were then detected with anti-BrdU monoclonal antibody (Roche Diagnostics GmbH) and Alexa 488–anti mouse IgG (Molecular Probes). Double-color fluorescent images were merged using Photoshop (Adobe Systems). Signals in every fourth section were analyzed. The number of BrdU NCAM double-positive cells per circumvallate trench wall was counted in each section at 3 and 12 h and 1, 2, 3 and 4 days after BrdU injection.

Results

NCAM immunoreactivity in Mash1-, gustducin- and T1r3-expressing cells

The cell type of Mash1-expressing cells in taste buds remains unclear, although we have previously reported that Mash1 has an expression pattern distinct from gustducin and T1r2, suggesting that Mash1-expressing cell is not a type II cell (Kusakabe et al., 2002). To clarify whether Mash1-expressing cells can be categorized as type III cells, Mash1-expressing cells were examined for immunoreactivity for NCAM, a type III cell marker in taste buds, using a combination of in situ hybridization and NCAM immunohistochemistry. The same procedures were used to test gustducin- and T1r3-expressing cells for NCAM. Almost all of the Mash1-expressing cells (98.2%, 372/379) in the taste buds were NCAM-immunopositive (IP), suggesting that Mash1-expressing cells should indeed be categorized as type III cells (Figure 1, Table 1). The majority of NCAM-IP cells in the taste buds expressed Mash1. Expression of gustducin and T1r3 was observed primarily in NCAM-immunonegative (IN) cells; however, small subsets of gustducin- and T1r3-expressing cells were unexpectedly found to be NCAM-IP: 12.9% (42/325) of gustducin- and 7.73% (30/388) of T1r3-expressing cells (Figure 1, Table 1). This result contradicted previous results in rats (Yang et al., 2000). If gustducin- and T1r3-expressing cells are derived from Mash1-expressing cells, as speculated previously (Kusakabe et al., 2002), it is expected that NCAM immunoreactivity in gustducin- and T1r3-expressing cells is a remnant from the Mash1-expressing stage, and is down-regulated in the maturation process of the taste cells. We next examined the ratio of NCAM-IP cells in Mash1-, gustducin- and T1r3-expssing cells during development. Double-color in situ analysis in circumvallate papillae was performed in the circumvallate papillae at 0.5, 2.5 and 10.5 days after birth. At 0.5 days after birth, taste bud structure was hardly observed, and the signals for gustducin and T1r3 were very rare but clearly detected. Only 37 gustducin- and 22 T1r3-expressing cells were observed through the analysis of six circumvallate papillae, suggesting that the expression of these genes may have just begun to turn on, while Mash1-expression was detected in almost every slice. The NCAM-IP cells were mainly observed in the dorsal epithelium of circumvallate papillae at 0.5 days, consistent with previous reports (Takeda et al., 1992). In contrast to the results in adulthood, at 0.5 days, the vast majority (∼90%) of gustducin- and T1r3-expressing cells were NCAM-IP (Figure 2, Table 1). The ratios of NCAM-IP cells for each type of gene-expressing cells in the circumvallate papillae are summarized in Table 1 and Figure 3. The ratios of NCAM-IP cells in gustducin- and T1r3-expressing cells decreased markedly during taste bud development. In contrast, the vast majority of Mash1-expressing cells were constantly NCAM-IP during development. However, the proportion of NCAM-IP cells in younger stage (88.9% at 0.5 days after birth) was lower than that in elder stage (98.2% at adulthood), raising the possibility that Mash1-expressing cells may be NCAM-IN at the onset of Mash1 expression.

The difference in the timing of differentiation of cells expressing Mash1, gustducin and T1r3 in adult mouse

We have previously reported that the expression of Mash1 precedes those of taste reception-related genes such as gustducin and T1r2 during development, suggesting the differential timing of cell differentiation (Kusakabe et al., 2002). To clarify whether the differentiation of Mash1-expressing cells precedes those of taste reception-related gene-expressing cells in adult mice, BrdU-labeling experiments were performed. The ratios of BrdU-positive cells in Mash1-, gustducin- and T1r3-expressing cells at various intervals after BrdU injection are shown in Figure 4. The BrdU signals in gustducin- and T1r3-expressing cells began to be obvious at 2 days after BrdU injection (Figure 5, Table 2), consistent with previous immunohistochemical results, in which the BrdU signal began to be detected at 2.5 days after BrdU injection (Cho et al., 1998). The BrdU signals gradually increased in these cells in the subsequent days. In Mash1-expressing cells, the BrdU signal increased more rapidly and was clearly observed at 1 day after BrdU injection (Figure 5, Table 2); a few signals, in <1% of Mash1-expressing cells, were observed at 3 and 12 h. At 2 days, BrdU signals in Mash1-expressing cells reached a peak level, and 11% (37/335) of Mash1-expressing cells were BrdU-positive. The ratio was kept at an almost constant level or slightly decreased in subsequent days. The results in Shh-expressing cells (originally published in Miura et al., 2004) are also shown in Figure 4 and Table 2 for the sake of comparison with Mash1-expressing cells. In Shh-expressing cells, BrdU signals were observed earlier than those in Mash-expressing cells, and the signal began to decrease 2 days after BrdU injection (Figure 5, Table 2).

The timing of differentiation of NCAM-IP cells

To compare the timing of differentiation of NCAM-IP cells to those of Mash1-, gustducin- and T1r3-expressing cells, BrdU-positive cells with NCAM immunoreactivity were examined at 3 and 12 h, and 1, 2, 3 and 4 days after BrdU-injection (Figure 6). In NCAM-IP cells, BrdU signals were clearly observed at 1 day after BrdU injection and reached peak level at 2 days, like that in Mash1-expressing cells (Figures 6 and 7). This result suggests that the differentiation timing of NCAM-IP cells is close to that of Mash1-expressing cells and earlier than those of T1r3- and gustducin-expressing cells (Figure 4). The change of the rate of BrdU-positive cells in NCAM-IP cells during 4 days after BrdU injection was similar to that in Mash1-expressing cells. The number of NCAM-IP cells was hard to be estimated because of the NCAM signals for nerve fibers in taste bus. However, the BrdU NCAM double-positive cells could be identified clearly. Then the number of BrdU NCAM double-positive cell per trench in each section was analyzed instead of the ratio of the number of BrdU-positive cells in NCAM-IP cells to the total number of NCAM-IP cells.

Discussion

In the present report, we revealed a gradual decrease in the ratio of NCAM-IP cells in gustducin- and T1r3-expressing cells during development, while the majority of gustducin- and T1r3-expressing cells were NCAM-IP at 0.5 days after birth. These observations indicate that the gustducin- and T1r3-expressing cells may be NCAM-IP at the beginning of cell differentiation and that NCAM immunoreactivity may be reduced during their maturation process. In this case, NCAM immunoreactivity appears to represent younger populations in gustducin- and T1r3-expressing cells in adult mouse taste buds. In addition, Mash1-expressing cells in taste buds were found to be NCAM-IP. These findings appear to support the idea that gustducin- and T1r3-expressing cells are derived from Mash1-expressing cells; in taste cell differentiation, NCAM immunoreactivity in gustducin- and T1r3-expressing cells may be a remnant from the Mash1-expressing cell stage, and might be downregulated during maturation process. This idea of taste cell differentiation has been originally proposed based on the segregated expression patterns between Mash1 and taste reception-related genes, and differential expression timing of these genes during development in our previous report (Kusakabe et al., 2002). In the present study, BrdU-labeling experiments provided additional evidence that, in adult mice as in developing mice, the differentiation of Mash1-expessing cells precedes those of gustducin- and T1r3-expressing cells in taste buds. BrdU signals in Mash1-expressing cells were rapidly elevated before those in gustducin- and T1r3-expressing cells. In addition, previous studies showed that the expression of NeuroD in taste buds overlapped gustducin but not NCAM (Suzuki et al., 2002). Taken together, our findings suggest that Mash1 expression precedes NeuroD, and may not overlap NeuroD in taste buds. Mash1 and NeuroD are bHLH transcription factors critical for neuronal differentiation, and the NeuroD expression following Mash1 is known to be essential for the differentiation of olfactory neurons (Lee et al., 1995; Cau et al., 1997, 2002). Sequential expression of Mash1 and NeuroD might be involved in the differentiation of the taste receptor cells expressing gustducin and T1r3. On the other hand, we did not observe an apparent decrease of the ratio of BrdU-positive cells in Mash1-expressing cells inversely related to the increase of BrdU-signal in gustducin- and T1r3-expressing cells. We would expect to observe this if all Mash1-expressing cells become gustducin- or T1r3-expressing cells. This suggests there are cell populations expressing Mash1 independently from the differentiation of gustducin- and T1r3-expressing cells. Mash1 might continue to be expressed in particular populations of Mash1-expressing cells maybe throughout cell life. Further studies, especially with in vitro culture systems or gene transfer systems to induce taste cell differentiation, are needed to elucidate the function of Mash1 and NeuroD in taste cell differentiation, and the cell lineages from Mash1-expressing cells to cells expressing taste reception–related genes. On the other hand, BrdU signals in Shh-expressing cells were observed earlier than that in Mash1-expressing cells. It is possible that Mash1-expressing cells are derived from Shh-expressing cells; overlap expression of Shh and Mash1 was observed in the basal region of taste buds to some extent (unpublished data), supporting this possibility. Further analysis of the basal cells in taste buds will provide better understanding of the progenitor cells for taste buds.

NCAM immunoreactivity in taste buds has been considered as a marker for type III cells, which is distinguished from type II cells including gustducin-IP cells (Takeda et al., 1992; Nelson and Finger, 1993; Yang et al., 2000; Yee et al., 2001). In the present report, we showed that almost all Mash1 expression in adult mouse taste buds occurred in NCAM-IP cells, suggesting that Mash1-expressing cells could be categorized as type III cells. However, NCAM immunoreactivity was observed also in gustducin- and T1r3-expressing cells, contradicting previous studies in which NCAM- and gustducin-immunoreactivities were linked to type III cells and type II cells, respectively, and did not overlap. Our results in mice seem to indicate that NCAM immunoreactivity is not so strictly linked to type III cells and may be observed also in type II cells, or that gustducin and T1r3 are expressed also in type III cells. In this study, to examine the co-expression pattern, we used the combination of immunohistochemistry for NCAM and in situ hybridization for Mash1, gustducin and T1r3, unlike previous reports, in which the immunohistochemical method was used for gustducin and NCAM. Because the appearance of mRNAs is expected to precede the protein products of these genes, it is possible that the overlap, which could not be observed immunohistochemically, was detected by our method in the present report. Alternatively, the difference in species may be more important. Immunoelectron microscopic analysis of the cell type of gustducin-expressing cells has been performed only in rats (Yang et al., 2000), and the case in mice remains unclear. Furthermore, a morphological difference between mice and rats has been reported in the cell types of taste buds regarding synaptic connection and dense-cored vesicles (Kinnamon et al., 1985; Delay et al., 1986; Pumplin et al., 1997; Clapp et al., 2004). In either case, based on our hypothetical model for taste cell differentiation, type II cells expressing gustducin or T1r3 are likely to be derived from NCAM-IP type III cells expressing Mash1. This is partially consistent with the previous idea that taste cell maturation progresses from type I (dark) → type III (intermediate) → type II (light), originally proposed by Delay (1986). However, our model is different from this scheme in the following two points:

1. Type III cells expressing Mash1 are assumed to comprise two distinct populations: one becomes type II cells expressing gustducin or T1r3 and the other remains in the type III cell state expressing Mash1. It is unclear whether these two populations share progenitor cells, while a serotonin-IP subpopulation in type III cells was reported to tend to be related to a separate lineage based on the results of mosaic mouse analyses (Stone et al., 2002).

2. Type I cells are not assumed to be progenitors of type III cells expressing Mash1, because of the rapid differentiation of Mash1-expressing cells: BrdU signals were observed in Mash1-expressing cells at 3 and 12 h after BrdU injection, although those were very rare. There is no delay in differentiation timing of Mash1-expressing cells in comparison with the case of dark cells (type I) (Delay et al., 1986). Furthermore, the sequential expression of Mash1 and NeuroD suggests molecular similarities in cell differentiation between taste cells and neurons, while the gene expression related to Delta-Notch signaling was recently reported in taste buds (Seta et al., 2003). Delta-Notch signaling is known to be involved in neuron/glia cell fate determination at an early phase of the development (Artavanis-Tsakonas et al., 1999; Wakamatsu et al., 2000; Grandbarbe et al., 2003), and type I cells are believed to be glia-like cells (Lindeman, 1996; Lawton et al., 2000). It is possible that Delta-Notch signaling might be involved in cell fate determination between type I cell and type III cells in the early phase of taste cell differentiation.

A schematic model of possible cell lineage relationships in taste buds is presented in Figure 8. In this scheme, the cells expressing gustducin or T1r3 in the mature stage are assumed to be type II cells, and NCAM-IP cells expressing Mash1 to be type III cells. However, the cell type of NCAM-IP cells expressing gustducin or T1r3 is ambiguous; these cells may be in a transition state. Further immunoelectron microscopic analysis at the onset of the differentiation of gustducin- and T1r3-expressing cells can provide important information regarding the classification of this cell type on the basis of morphological characteristics.

This work was supported by BRAIN. The authors would like to thank Mr Shindo for numerous discussions, Dr Ookura for antibodies and Mr Matsushita for providing the experimental mice.

References

Author notes

1National Food Research Institute, 2-1-12 Kannondai, Tsukuba-shi, Ibaraki 305-8642, Japan, 2Bio-oriented Technology Research Advancement Institution (BRAIN), 3-18-19 Toranomon Minato-ku, Tokyo 105-0001, Japan and 3Faculty of Dentistry, Kyushu University, 3-1-1 Maidashi Higashi-ku, Fukuoka-shi, Fukuoka 812-8582, Japan