Characterization of different olfactory markers by IHC
The classification of different olfactory markers in function of their expression in mature and immature OSNs can be a useful guide for investigators to choose an appropriate marker. OMP is known to be a critical protein in OSN maturation process14. We chose it as the reference marker to localize mature OSNs. Tuj-1 expression decreases during OSN maturation15 and it is an usual marker for immature OSNs in adult rodents16. In our study using rats of 3 to 5-week old, we observed its localization in immature OSNs by double staining assay with OMP as reported in adult rats. In addition, Tuj-1 showed a very strong labeling signal. As a consequence, Tuj-1 was chosen as the reference marker of immature OSNs.
Nine olfactory markers were characterized in this study (Figure 6): OMP was the only marker of mature OSNs. Tuj-1, DCX, and Olig2 were the markers of immature OSNs whereas N-cadherin, LHX, and PGP9.5 were detected in both populations. GAP43 and Gβ seemed specific of the OSN nerve fascicles. Immature OSNs were localized to the lower layers of the OE, and mature OSNs were localized in the upper layers. N-cadherin was the only marker that was strongly expressed in horizontal basal cells.
GAP43, a nervous system-specific protein that plays an important role in OSN axon regeneration growth17, is expressed in both OSN cytoplasm and axon18 or only in axon19. In this study, we observed GAP43 expression preferentially in the olfactory nerve fibers of the lamina propria. Two hypotheses could explain these differences in subcellular locations: 1. The expression of GAP43 in the cytoplasm of the OSN may not be constant during lifetime but decreases rapidly with the age of the animal20. 2. A single antibody may not recognize its antigenic site in all forms of the protein due to possible antigenic site concealment in particular subcellular location of the protein. Dominant expression of Gβ1, a component of the heterotrimeric G proteins, was found in OSN axons by IHC as previously reported21. These two markers could be useful to study the olfactory nerve fiber network on a flat-mounted nasal mucosa without being hindered by OE staining.
In the lamina propria which is a thin layer of loose connective tissue under the respiratory and olfactory epithelium, we observed two distinct types of nerve fiber fascicles: OSN nerve fibers/axons and nerve fibers/axons from the peripheral nervous system (PNS). These two types of nerve fibers had a very different morphology (Supplementary Figure 1) and staining. NF (L, M, and H) and peripherin were mainly observed in the nerve fibers of the PNS. These thin regular bundles were very different from the swollen fascicles of the olfactory nerve. These PNS nerve fiber bundles have been described in the literature as trigeminal nerves19,22.
We believe that this characterization of OSNs markers developed in this study, according to their cellular and subcellular localizations (Figure 6), will facilitate the choice of appropriate markers in future studies on olfactory mucosa.
Immunostaining on flat-mounted septal mucosa
The olfactory mucosa is a pseudo-stratified epithelium located on the surface of the nasal cavity. It is therefore possible to perform immunostaining on flat-mounted olfactory mucosa.
However, observations of OE with this technique have been rarely reported in literature: 9 publications over a period of 23 years were identified and summarized in Table 19,11,12,23−28. Among these studies, 5 are from the same team, directed by Professor James Schwob in Boston, supporting that this approach is not mastered by numerous teams, possibly due to technical pitfalls.
First, considering the technique, there is no consensus. Each team used a different protocol and even in a same team, different protocols are reported and adapted to each situation. Yet, tissue fixation with Formaldehyde 4% (equivalent to PFA 4% or formalin 10%) represent a common step in almost all these protocols. It is a conventional fixative and the main component of the different fixatives used (Table 1). Post-fixation treatment, consisting in antigen retrieval and cell permeabilization, was reported in most cases. Glycosidase (enzyme), SDS (denaturant), and ethanol (denaturant) can all be used for antigen retrieval. Sonication, freeze/thaw, Ethanol or SDS may play a role in cell permeabilization like triton, but can be also responsible for damages of the tissular microstructure.
Only 3 studies out of 9 used immunofluorescence, in spite of many advantages compared to immunoenzymology. One of the main causes could be the important autofluorescence of the olfactory mucosa which is opaque. For this reason, Child et al. added a tissue clarification step to make the olfactory mucosa more transparent in order to decrease autofluorescence9.
The immunostaining technique on flat-mounted cornea, we developed previously, highlighted the fact that cells on a flat-mounted tissue are more susceptible to overfixation with a 4% concentration of PFA, compared to cells in culture29. Using a low concentration of PFA, 0.5% or 1%, can greatly improve the results of immunostaining on the flat-mounted corneal tissue and still allows an overall tissue view and an accurate subcellular marker localization30,31. In addition, overfixation by formaldehyde promotes additional autofluorescence of the tissue. By applying the same technical approach to the nasal tissue, we succeeded in developing a simple and easily reproductible immunostaining technique on flat-mounted olfactory mucosa. Four of the five OSN markers selected by IHC (Tuj-1, DCX, N-cadherin and PGP9.5) were detectable directly without any antigen retrieval process. Only the OMP marker needed antigen retrieval with SDS. This antigen retrieval had also shown its effectiveness for some difficult staining in the case of cornea29.
We listed here the 5 olfactory markers suitable for immunostaining on flat-mounted septal mucosa: OMP, Tuj-1, DCX, N-cadherin and PGP9.5. We speculate that other olfactory markers that are able to stain OSN dendrites would also be suitable for immunostaining on flat-mounted olfactory mucosa. From a technical point of view, we noted in this study that antibodies validated by IHC (frozen) were also applicable for immunostaining on flat-mounted tissue.
In this study, we provided a simple and efficient protocol and described the 5 more relevant OSN markers for immunostaining on flat-mounted septal mucosa. By combining one of the 5 OSN markers with Tubulin β IV (a marker of the respiratory epithelium), the boundary between the two epithelia was clearly defined. This precise visualization of the transition zone between olfactory and respiratory epithelium may offer many opportunities and applications in the future, especially in research on respiratory metaplasia of the OE.
Density of mature, immature and transitory OSNs
Few studies have been devoted to evaluate the density of OSNs, yet this data is important for the analysis and understanding of olfactory dysfunction.
In 1981, Hinds et al. calculated the density of OSNs using electron microscopy, based on the count of olfactory knobs on a section passing through the middle of the rat’s septal OE. They estimated the density of all OSNs between 60,000 and 120,000/mm² 32. In 1989, Meisami et al. calculated the density of total OSNs from cross-sections of OE by counting the number of non-specifically counterstained nuclei. They estimated the OSNs cell density at 94170 ± 3950/mm2 in 25-day-old rats33. In 1996, Loo et al. described a count of mature and immature OSNs based on the immunostaining technique on flat-mounted olfactory mucosa24. They counted olfactory dendrites/buttons of mature OSNs stained by OMP and immature OSNs by GAP43. The technique for labeling revelation was immunoenzymology with the tracer DAB (3, 3'-diaminobenzidine). They found 30022 ± 5311/mm² mature OSNs (OMP+) and 8443 ± 1012/mm² immature OSNs (GAP43+) on the septal OE of 7-month-old rats. Double staining was impossible using immunoenzymology, so the transitory OSNs were not detected.
In this study, we found the density of total OSNs per millimeter square was 105912 ± 13899 which is consistent with the two first studies previously cited32,33. We found 56528 ± 9294/mm² of OMP+ OSNs and 63832 ± 8174 of Tuj-1 + OSNs. In general, the density of mature and immature OSNs evaluated by this study is higher than the OSNs density reported by Loo et al. 24. One of the reasons that could explain these differences could be to the sensitivity of immunofluorescence compared to immunoenzymology (DAB tracer system). Although, it can explain the difference of mature (OMP+) OSN density (56528 ± 9294 vs 30022 ± 5311), this cannot be a valuable explanation for the huge difference in immature OSNs density (63832 ± 8174 (Tuj-1+) vs 8443 ± 1012 (GAP43+)). Another reason might be the weaken and even absence of expression of GAP43 in adult rat20. In their study, Loo et al. used 7 month-old rats to count immature OSNs stained by GAP4324. In addition, younger rats we used could have more immature OSNs than older rats.
Thanks to double staining inside dendrites/knobs with both OMP and Tuj-1, we were able to identify a population of transitory OSNs that express OMP and Tuj-1 at the same time. This transitory OSN population density was 14448 ± 5865 /mm² in 3-5 week-old rats, representing about 14% of total OSNs. We noticed that this population co-expressing OMP and Tuj-1 was not easy to detect by IHC on cross sections in the OSN soma (neuron body). Indeed, very few OSNs co-expressed these two proteins in the soma (Supplementary Figure 4). This difference in co-expression of OMP and Tuj-1 in soma and dendrites was already observed in a previous study34. On the contrary, the co-staining of these two proteins seemed much more evident in the OSNs axons (Figure 1). This contrast of co-staining in the soma, dendrites and axons indicates that the retrograde transport of proteins in OSNs may be very slow.
Due to double labeling with fluorescent tracers on flat-mounted mucosa, we were able to quantify for the first time a population of transitory OSNs that express both OMP and Tuj-1 in their dendrites. We believe that this technique of counting mature, immature and transitory OSNs would be a good tool for studying pathophysiology of OE.
Direct immunostaining on ethmoid turbinates
Once finalized, the protocol of immunostaining on flat-mounted tissue was successfully transferred to the ethmoid turbinates. The high intensity of Tuj-1 staining allowed us to clearly identify the mapping of the OE on the ethmoid turbinates despite the fact that the image was acquired using a fluorescence macroscope that were not able to filter the autofluorescence of the tissue (connective tissue and bone) below the OE. We observed that the boundary between the respiratory and olfactory epithelium was clear and unambiguous on both septal and turbinate mucosa in 3 to 5 weeks old rats.
To the best of our knowledge, the only study showing the turbinates’ OE used transgenic mice in which the OSNs expressed GFP13. This transgenic technique is not transferable to human. In this study, we have, for the first time, demonstrated OE mapping directly on rat ethmoid turbinates using the immunostaining technique. We believe that this technique would be applicable for OE mapping on human turbinates. We could transpose this technique directly to the nasal cavities of human donors to allow the mapping of the olfactory zones using the technology of fluorescence video microscopy with an optic fiber.