The schematic in Figure 1 illustrates the histology of the olfactory mucosa (OM) and the cellular composition of the olfactory epithelium (OE) and lamina propria (LP). Olfactory receptor neurons (ORNs) extend their axons from the OE to the LP where they coexist with the OECs and the supportive connective tissue. Within this tissue may also be Schwann cells ensheathing nerves innervating blood vessels. In Figure 1A it can be seen that the calponin immunoreactivity was only detected in the LP of neonatal OM. Olfactory receptor nerves were intensively inmunoreactive for the neuronal marker TUJ1, and were surrounded by cells immunoreactive for calponin (Figure 1A, B). No colabelling was detected between calponin and TUJ1, with calponin positive cells and their nuclei clearly distinguishable from the nerves (Figure 1B). Since OECs are known to typically surround ORN axons, double immunofluorescence for calponin and the OEC markers S100 and p75^NTR was performed in the neonatal tissue (Figure 1C–F). As expected, S100 and p75^NTR positive OECs were located just underneath the basal layer of the OE (Figure 1D, arrow) and surrounding the nerves (Figure 1E, F, asterisk). Calponin immunoreactivity was located close to the S100 and p75^NTR positive OECs but no colocalisation was observed with either OEC markers (Figure1C–F). Even though Schwann cells (identified by the same markers as OECs) may be present in the olfactory mucosa, we did not observe any cells immunoreactive for both p75^NTR and calponin indicating lack of calponin positive OECs or Schwann cells. This staining was specific as control reactions lacking the primary antibody did not result in any immunoreactivity. The absence or presence of proteinase K also has no effect on the pattern of immunoreactivity.

To identify the cells that express calponin in the LP from neonatal rats, we carried out double immunofluorescence with antibodies to calponin together with markers of connective tissue namely fibronectin (Fn) and alpha smooth muscle actin (α-SMA). Connective tissue cells immunoreactive for α-SMA were mainly located in the central region of the LP and associated with the olfactory nerves (Figure 2A as indicated by dashed lines). Double staining for α-SMA and calponin showed colabelling of both markers in this same region (Figure 2A). α-SMA positive staining that lacked calponin in this region was associated with blood vessels. To remove the possibility that the calponin positive cells were OECs immunolabeled with α-SMA as has been suggested by Jahed et al., 26 double immunofluorescence for α-SMA and both S100 and p75^NTR was performed. No colocalisation was observed between α-SMA and both OEC markers further supporting our evidence for a lack of expression of calponin and α-SMA in neonatal OECs (Figure 2C, E, data not shown for S100). However, not all the cells immunoreactive for calponin colabelled with α-SMA suggesting there was another cell population that expressed calponin in the LP. Calponin positive cells that lacked immunoreactivity to α-SMA were seen in areas closer to the basal part of the OE (Figure 2A, arrow). Double immunofluorescence for Fn and calponin marked in general a different area of immunoreactivity than α-SMA in the LP (compare Figure 2A and 2B at arrow). Fn colocalised with some calponin in the area directly underneath the basal layer of the olfactory epithelium (Figure 2B, arrow). This suggests that another connective tissue cell co-expresses Fn and calponin but lacks α-SMA (Figure 2F). A proportion of the Fn positive cells which lacked calponin must be OECs as they coexpress p75^NTR (Figure 2D, arrow). Similar staining patterns were obtained when proteinase K was either included or excluded from the protocol (Figure 2G). Once more, control experiments in which the primary antibody was omitted, were negative (Figure 2H).

Embryonic olfactory mucosa was immunostained in the presence and absence of proteinase K with the same markers used in neonatal OM tissue to reveal possible age-related calponin expression in OECs. Embryonic olfactory neurons and their axons were imunoreactive for TUJ1 (Figure 3A). p75^NTR positive OECs and Fn positive connective tissue were detected in the LP as seen for neonatal tissue (Figure 3A–C). No immunoreactivity for calponin could be detected in any region of the embryonic olfactory mucosa. Furthermore inclusion of proteinase K did not reveal calponin staining. The lack of calponin in the embryonic tissue suggests that this protein is expressed later on in the development in the olfactory connective tissue. Control experiments in which the primary antibody was omitted were negative (Figure 3D).

In Figure 4A, B, purified cultured OECs and Schwann cells were plated onto coverslips and labeled with calponin together with p75^NTR. Since it has been reported that calponin immunoreactivity in glia was maximized by pretreatment with proteinase K 11, we compared methods in which pretreatment was omitted. In the presence and absence of proteinase K pretreatment both OECs and Schwann cells expressed weak, nuclear, background levels of calponin immunoreactivity (Figure 4Ai, ii, Bi, ii, illustrates cells treated with proteinase K but similar results occurred for cells not treated). However, when the primary antibody was omitted this staining was reduced but could still be detected suggesting that this nuclear background staining is non-specific attachment of the secondary antibody (Figure 4Aiii, iv and 4Biii, iv). Also, note the similar nuclear background staining (arrowhead) in the fibroblasts which express strong fibrillar calponin immunoreactivity (Figure 4Ci–iii, arrow). These fibroblast-like cells were devoid of p75^NTR immunoreactivity.

To confirm the connective tissue origins of the contaminating calponin positive fibroblast-like cells observed in the OEC and Schwann cell cultures, cultures enriched for non glial cells were generated from neonatal olfactory bulb and sciatic nerve by incubating the cultures in DMEM containing 10% FBS (Figure 5A–B). Cultured cells exhibited a fibroblast-like morphology and expressed strong immunoreactivity for calponin with a characteristic fibrillar appearance. These fibroblast-like cultures were heterogeneous for the expression of α-SMA and Thy1.1 that resembles the heterogeneity we observed in the LP of the in vivo olfactory mucosa (Figure 5B, D). Double immunofluorescence revealed colocalisation of calponin with α-SMA and surface fibronectin, supporting the non-glial origin of the calponin immunoreactivity observed in the purified OEC and Schwann cell cultures (Figure 5E, F).

It has been suggested that cultures of OECs are contaminated with Schwann cells 11. A major functional difference between OECs and Schwann cells can be seen in their ability to interact with astrocytes, with OECs clearly mingling with astrocytes and Schwann cells and astrocytes forming boundaries 182021. These differences are believed to be due to a differential FGF/heparin sulphate proteoglycan signaling cascade 21. To illustrate the purity of our OEC preparation, we set up these confrontation assays and immunolabeled the cells with GFAP, p75^NTR and calponin. These assays take 10–14 days for the cells to meet and are carried out in DMEM -10% FBS. As Schwann cells take up Brdu in DMEM-10% FBS (data not shown) we believe this would be enough time for any possible contaminating Schwann cells to increase in number. Therefore, we would expect any contaminating Schwann cells in these confrontation assays to affect boundary formation. Due to the similarity in the class specific secondary antibody for calponin and p75^NTR, OECs and Schwann cells in these cultures were detected by the blue DAPI nuclei associated with unlabeled cells, and sister coverslips were immunolabeled with p75^NTR and GFAP. In Figure 6A, C confrontation assays between Schwann cells and OECs with astrocytes can be seen. OECs mingled with astrocytes (Figure 6A) while the Schwann cells formed a boundary (Figure 6C, arrows). To our surprise, calponin staining could be seen to be associated with astrocytes and not OECs or Schwann cells (Figure 6B, D, asterisk). This difference in the way Schwann cells interact with astrocytes in confrontation assays when compared to OECs gives support to the view that there is no Schwann cell contamination in OEC cultures after at least 2 weeks.

To confirm cultured astrocytes express calponin, we generated two types of astrocytes, cortical astrocytes (CAs) and astrocytes differentiated from neurospheres (NsAs). Cortical astrocytes were over 98% pure as assessed by GFAP expression and these astrocytes express calponin heterogeneously (Figure 6E–F). The calponin was fibrillar and similar to that seen for fibroblasts. Astrocytes generated from P1 neurospheres by incubation in DMEM containing 10% FBS also expressed fibrillar calponin in a heterogeneous manner (Figure 6G). Control cultures without primary antibody were negative although a weak background staining was also detected as seen for OECs and Schwann cells (Figure 6H).

Primary neural cells were purified and cultured as previously described. Neonatal OECs were purified from the olfactory bulb of 7-day-old Sprague-Dawley rat pups using the O4 antibody, anti-galactocerebroside and the fluorescent activated cell sorter (FACSVantage Becton Dickenson, 14. OECs were cultured in defined medium DMEM (DMEM-BS, 46) with 5% FBS on poly-L-lysine (PLL, 13 μg/ml; Sigma) coated 25 cm² flasks, and further supplemented with fibroblast growth factor 2 (FGF2, 500 ng/ml; Peprotech, London, UK), heregulin (hrgβ1, 50 ng/ml; R&D Systems Europe Ltd., Abingdon, UK), forskolin (10^-6M; Sigma) and astrocyte conditioned medium (ACM, 47) which has been shown to be a potent mitogen for promoting prolonged growth of p75 neurotrophin receptor (p75^NTR) expressing OECs 48.

Schwann cells were obtained from sciatic nerves of 7-day-old Sprague-Dawley rat pups and purified using a modification of the method described by Brockes and colleagues 49. The modification involved an initial flick of the flask after one day in culture to preferentially select for Schwann cells, followed by incubation with Thy1.1 antibody (1:50 supernatant; Sigma, Dorst, UK) and rabbit complement (1:6; Harlan Sera-lab Ltd, Loughborough, UK) to reduce contamination by fibroblasts. Cells were plated on PLL (13 μg/ml, Sigma) coated flasks and maintained in DMEM-10% FBS supplemented with forskolin (10^-6 M; Sigma) and hrgβ1, (20 ng/ml; R&D Systems Europe Ltd, Oxon, UK).

Astrocytes were purified from the cortex as previously described 1847. The cells were maintained in DMEM containing 10% FBS on PLL coated flasks. ACM was collected from confluent flasks of astrocytes by incubating the cells in 10 mls of serum free DMEM 47 as described in Lakatos et al., 18 for 48 hrs before collecting, spinning and filtering 47. Neurospheres were generated as described by Reynolds and Weiss 50 from P1 striatum, plated in neurobasal medium containing B27 supplements (Invitrogen, Paisley, UK). Neurosphere astrocytes were established by plating neurospheres onto PLL coated coverslips and incubating them in DMEM containing 10% FBS for a week.

Cell purity was assessed by labeling the cells with GFAP (glial fibrillary acid protein) antibody for astrocytes, and p75^NTR antibody for OECs and Schwann cells 51. All cultures were greater than 98% pure. The cells were maintained in culture at 37°C in 7% CO₂ for no longer than 1 month and were refed twice a week. For immunofluorescence the cells were passaged onto PLL coated 13 mm coverslips. To study cell surface antigen, cells were incubated with the primary antibody for 30–40 min at room temperature (RT) followed by the class specific antibody for 30 min at RT. Intracellular antibodies were incubated on cells permeabilised using 100% ice cold methanol and placed in the freezer for 15 min followed by their class specific antibody as described above. Calponin staining was carried out using the method and the same calponin antibody as described by Rizek and Kawaja 11. Cells were prefixed in 4% paraformaldehyde for 10 min followed by incubation with the cell surface marker p75^NTR and its class specific fluorochrome-conjugated antibody as described above. Cells were then fixed in 100% ice cold methanol for 10 min followed by pretreatment with 20 μg/ml of proteinase K for 7 min and lastly incubated with calponin for 1 hr and its class specific antibody for 30 min. Hank's solution (Gibco) containing 5% calf serum and Hepes buffer was used to dilute the antibodies and for washing between incubations. In some experiments preincubation with proteinase K was omitted. Primary antibodies include monoclonal mouse anti-human Calponin (1:50; IgG1, Dako, Ely, UK). To identify OECs we used monoclonal mouse anti- and polyclonal rabbit anti-p75^NTR (hybridoma, 1:1; IgG1, 51, anti-rabbit, 1:800; Abcam, Cambridge, UK) respectively; and rabbit polyclonal anti-S100β 52 (1: 200; Dako). For identifying ORNs we used monoclonal mouse anti-βIII tubulin (TUJ1, 1:100; IgG1, Sigma). Connective tissue was identified by polyclonal rabbit anti-human fibronectin (1:100, Dako); mouse monoclonal anti-alpha smooth muscle actin (α-SMA, 1:400; IgG1, Chemicon, Hampshire, UK; 35 and mouse monoclonal Thy1.1/CD90 (1:100; IgG1, Serotec). Visualization of the primary antibodies was carried out using the class specific antibodies conjugated with fluorescein (green, FITC) or rhodamine (red, TRITC) (Southern Biotechnique, 1:100). Coverslips were mounted with VectorShield (Vector laboratories, Peterborough, UK) containing 4'6-diamidine-2-phenylindole-dihydrochloride (DAPI) to visualize the nuclei. Images were examined either by confocal microscopy (Olympus FV-1000) or by fluorescence microscopy (Zeiss Axioskop) using MetaMorph image analysis from Molecular Devices.

The olfactory mucosa and olfactory bulbs were dissected from the heads of postnatal day 7 Sprague-Dawley rat pups. The tissue from the P7 rats and the heads from embryonic E17-E20 pups were immersed in 4% paraformaldehyde for 1–3 days and then cryopreserved in 30% sucrose solution until equilibration. Sections of 10–20 μm were cut using a Bright cryostat onto Vectabond (Vector Lab) coated slides. For immunohistochemistry, the slides were washed 3 times for 5 min in PBS then treated with proteinase K 20 μg/ml for 7 min, washed and incubated in a blocking solution containing 10% goat serum and 0.3% triton-X 100 (Sigma) in PBS for 1 hr at RT. Sections were then incubated in the primary antibodies described above diluted in PBS-4%BSA for 2 hr at RT. After 3 washes of 5 min with PBS the sections were incubated with the appropriate FITC/TRITC conjugated anti-mouse or anti-rabbit secondary 1:100 (Southern Biotechnique, Cambridge Bioscience, Cambridge, UK) for 1 hr. After the incubation the sections were washed 3 times for 5 min with PBS and then mounted with VectorShield containing DAPI.

Confrontation assays between OECs and astrocytes and Schwann cells and astrocytes were set up as previously described 1718. Briefly, 10-μl strips containing 10,000 cells of either OECs or Schwann cells were set up opposing a parallel 10-μl strip containing 10,000 astrocytes on a PLL-coated coverslip. Cells were allowed to attach for 1 hr before washing in DMEM-FBS to remove non-attached cells. Cultures were then maintained in DMEM-FBS, and allowed to grow towards each other over a period of 10 days, giving time for cells to make contact and interact. Cultures were then immunolabeled using antibodies to GFAP and p75^NTR or GFAP and calponin and mounted with VectorShield containing DAPI to visualize nuclei. Images were examined using fluorescence microscopy (Zeiss Axioskop) using MetaMorph image analysis from Molecular Devices. Because we can induce i) OECs to form boundaries with astrocytes or ii) Schwann cells to mingle with astrocytes by modifying the FGF/heparin sulphate proteoglycan signaling cascade; it is unlikely that the different, initial, long-term culture conditions for OECs and Schwann cells mediate the differential behavior in the confrontation assays 21. Furthermore, similar differences in the way OECs and Schwann cells form boundaries with astrocytes have been demonstrated in vivo 21.