Our study revealed at least 3 types of microvillous cells in the main olfactory epithelium of mice. Villin and espin, actin-binding proteins typically expressed in microvilli of chemosensory and mechanosensory systems [17
] were the only cell markers that labeled all microvillous cell types, confirming that they all bear microvilli. Two of the microvillous cell types show the transient receptor channel TrpM5-GFP label (TrpM5a type and TrpM5b microvillous cells). In addition, electron microscopic as well as immunohistochemical experiments revealed microvillous cells that were TrpM5-GFP-negative and expressed other cell markers not seen in the TrpM5-GFP cells. We group these cells as non-TrpM5 type microvillous cells. None of the types of microvillous cells described in this study have an axon that penetrates the basal lamina nor did we detect synapses consistent with findings from other studies.
Kaske et al. studied cells in various tissues that expressed TrpM5 and postulated that TrpM5 is an intrinsic signaling component that plays a physiological role in olfaction of odorants and pheromones [12
]. As TrpM5 is also present in cells of the taste and gastrointestinal system the authors claim that TrpM5 is a marker for chemosensory cells. Lin et al. also found solitary chemosensory cells in the respiratory epithelium that are TrpM5-positive [18
]. If TrpM5 is a marker for chemosensory cells, the TrpM5-positive cells should express other characteristics of chemosensory cells including contact with the nervous system as well as members of the chemosensory transduction pathway, and/or other markers of ORNs like OMP, PGP 9.5 etc. In the case of the solitary chemosensory cells, Lin et al. showed that TrpM5-positive cells are also labeled by antisera against PLCβ2 and PGP 9.5, known markers for chemosensory cells [18
]. Yet, none of these components double-labeled the microvillous TrpM5-GFP-positive cells in our study. Also, the antibody SUS-1 which is specific for olfactory supporting cells did not label the TrpM5-positive cells. Thus, based on the present study, we conclude that the TrpM5-GFP-positive cells – termed here TrpM5a type and TrpM5b type microvillous cells – are not among the cells traditionally described as comprising the olfactory epithelium, i.e. olfactory receptor cells, supporting cells, and basal cells. Since none of the neuronal cell markers tested showed a positive result, and the cells have neither an axon projecting to the olfactory bulb nor contact to nerve fibers of trigeminal origin, it is not even clear whether these cells are sensory cells at all. At least for chemosensory cells we question the reliability of TrpM5 as a marker. However, we do not rule out that the TrpM5-GFP-positive microvillous cells may be chemoresponsive and influence local mucosal elements through the release of diffusible mediators.
A recent study on mouse and rat olfactory epithelium described microvillous cells labeled with antibodies against IP3R3 and PLCβ2 [10
]. These microvillous cells did not degenerate after bulbectomy suggesting that they are not ORNs. The authors claimed that these cells express the transient receptor channel TrpC6 although they did not show double-label experiments for TrpC6 and either IP3R3 or PLCβ2. In the present study we did not see any cells double-labeled for TrpM5 and TrpC6. Neither did the TrpM5-GFP-positive cells double-label for IP3R3 or PLCβ2 (see also Lin et al., companion paper, this issue). Consequently, we postulate that the TrpM5-GFP-positive microvillous cells are not the ones shown in the study by Elsaesser [10
]. It is likely that the TrpC6 cells belong to the group of non-TrpM5 type microvillous cells.
In the past, several studies mentioned microvillous cells in the main olfactory epithelium of rodents, e.g. [19
]. Since the morphological descriptions varied and the proposed functions were often controversial, these studies received little attention. Okano et al. [21
] postulated that microvillous cells in the olfactory epithelium of dogs were precursors of supporting cells. Bannister [22
], based on the electron microscopic descriptions of the chicken olfactory epithelium [23
], suggested that microvillous cells may be precursors for ciliated olfactory neurons. This hypothesis was rejected by Jourdan [20
] since both cell types have very distinct features and no "intermediate" stages between microvillous and ciliated olfactory neurons have ever been detected.
Others have attributed various functions to these microvillous cells in addition to the notion of being precursor cells. One central question was: Are the microvillous cells true ORNs? Rowley and coworkers conducted experiments with HRP as a neural tracer and showed that microvillous cells could be retrogradely labeled from the rat olfactory bulb [24
]. The authors mentioned, however, that a small percentage of microvillous cells "did not effectively transport HRP" from the olfactory bulb to their somata. There is also the possibility of nonspecific labeling. A study on the OE of fish reported a third type of ORN labeled by fluorogold [25
]. Later it was shown that the fluorogold labeled cells in the OE nonspecifically [26
]. In the present study we did not find any evidence of an axon projecting towards the olfactory bulb which would have been proof that these microvillous cells are, indeed, ORNs. Although we cannot rule out completely that we may have missed cells with an axon in our ultrathin (70 nm) serial sections, we conclude that the microvillous cells described in this study are not ORNs.
Several authors reported microvillous cells without an axon, i.e. cells that were not ORNs. For instance, Carr et al. [7
] described a microvillous cell in the rat olfactory epithelium. These cells had no axon and survived ablation of the olfactory bulb but did not react with the marker for supporting cells SUS-1. Thus these cells do not belong to the olfactory system in the strict sense of basal, supporting, and olfactory receptor cells.
Occasionally the morphological features of microvillous cells reported in the past are similar to the features seen in microvillous cells of this study. Miller and coworkers described microvillous cells in the rat olfactory epithelium in a study of regeneration following exposure to toxic compounds [27
]. These cells had no synapses but finger-like processes extending between the supporting cells. Also, histochemical staining with ecto-5'-nucleotidase revealed microvillous cells in the rat olfactory epithelium that had numerous invaginations interdigitating with the cell membranes of neighboring supporting cells [28
]. We noted numerous processes of TrpM5b and non-TrpM5 type microvillous cells interdigitating with other cells, and also invaginations in non-TrpM5 microvillous cells. Asan and Drenckhahn [5
] described two types of microvillous cells in rat and mouse. Among others, the authors showed microvillous cells labeled with antibodies against villin and CK 18. In the present study we detected villin in TrpM5a, TrpM5b and non-TrpM5 microvillous cells whereas CK 18 labeled only non-TrpM5 microvillous cells. It is probable that some of the microvillous cells seen in this study may be the microvillous cells described by [7
], or [5