Introduction

The murine olfactory epithelium (OE) contains a population of specialized neurons, the olfactory receptor neurons (ORNs), which are renewed throughout life in terms of permanent neurogenesis [7]. The regeneration and maintenance of ORNs is assured by differentiation olfactory stem cells that reside near the basal lamina of the OE. So far, in murine OE two kinds of basal cells, globose basal cells (GBCs) and horizontal basal cells (HBCs) can be distinguished [6]. The GBCs are the major group of proliferating cells, which give rise to ORNs and sustentacular cells [9, 19, 23]. In contrast, the HBCs are a quiescent population characterized by slow replication and self-renewal [8, 21]. Their proliferation in vitro and in cell cultures is controlled by the epidermal growth factor (EGF) and the transforming growth factor α (TGFα) [8, 13, 14, 32]. Taken together it is evident that HBCs are the potential stem cells, which can give rise to all cell types of the OE under in vitro and in vivo [8, 29] conditions. Hence, they are supposed to be the major player for OE [8, 29, 31] regeneration.

Recent studies have described the cultivation of free-floating neurospheres obtained from murine OE in serum-free conditions [5, 28, 42]. One study showed that basic fibroblast growth factor (bFGF) acts as a major stimulator of cell proliferation in olfactory neurospheres (ONSs) [5]. Using air-interface cultures, it was possible to passage ONSs; nevertheless, so far their long-term expansion could not be realized [22] and a controlled differentiation of ONSs into ORNs could not be achieved. A few studies describe the successful cultivation of immortalized ORNs from transgenic mice (H-2Kb-tsA58) [3, 24] for examining the physiology of olfactory coding. These ORNs expressed several specific OE marker proteins like ACIII, Gαolf [2, 25] and could be stimulated by single odorants (acetophenone, heptaldehyde, isovaleric acid, and L-carvone) [3].

The major aim of our study was to establish primary long-term cultures of ONSs derived from the OB (which was used as control for our study) and the OE, which provide a stable source to perform comprehensive analysis of differentiation and regeneration processes in the OE. The cultures of both ONSs were dissociated from wild-type black-6 (C57Bl/6) and transgenic TIS21 mice. TIS-21 is an antiproliferative protein synthesized in the last phase of cell cycle division before differentiation into neuronal precursor cells [20]. It controls the cell cycle progression and the expression of proneuronal genes by activity modulation of transcription factors. The expression of TIS21 mRNA in the precursor cells of the neuroepithelium has been demonstrated before by Haubensak and co-workers [17], who successfully raised TIS21 knock-in mice expressing the green fluorescent protein (GFP) under the control regulatory sequences of the TIS21 gene [17]. Our aim was to use TIS21 as a stem cell marker.

The in vitro differentiation of neurosphere-forming cells gave rise to different types of cells, which were recognizable at the microscopic level. The cells were further characterized by performing immunohistochemical analysis that confirmed the presence of putative neurons and glia cells in culture. The olfactory marker expression was investigated by performing PCR analysis. We also performed patch-clamp, calcium imaging experiments to investigate the functionality of the differentiated cells and to demonstrate if the differentiated putative neurons could be regarded as young ORNs and gained the capacity of odor recognition.

Materials and methods

Animals

Wild type (C57Bl/6) and transgenic TIS21 mice (provided by W.B. Huttner Max Planck Institute for Molecular Cell Biology and Genetics, Dresden) [17] were mated to obtain a stable population of TIS21-GFP expressing mice. All animal experiments were performed in accordance with the European Union Community Council guidelines. The animals were caged with water and commercial food ad libidum.

Cultivation and differentiation of neurosphere-forming cells

For the preparation of OE postnatal neurosphere-forming cells, postnatal (P0–P4) pups were decapitated and the heads were washed for 3 min in cold phosphate-buffered saline (PBS) (Gibco, Germany). After washout, the heads were deposited in fresh ice-cold DMEM/F12. Subsequent steps were performed under a binocular microscope (Wild M5A Heerbrugg, Switzerland). After removing the jaw and tongue, the OE was dissected from the nasal septum and collected in 35-mm cell culture dishes (Sarstedt, Germany) with ice-cold DMEM/F12 (Gibco, Germany). The collection of the tissue was followed by mechanical trituration via a glass pipette. The suspension was then collected in 2-ml tubes and subsequently centrifuged for 5 min at 500g and 4 °C. After removal of the supernatant, the pellet was re-suspended in enzymatic medium containing 30 U/ml papain (Sigma, Germany) and 40 µg/ml DNaseI (Sigma, Germany), and then digested for 30–35 min in a water bath at 37 °C with gentle agitation. After digestion, the resulting single cells were harvested (5 min, 500g, 4 °C) and the remaining enzymatic medium was removed. The pellet was re-suspended in neurosphere standard medium containing DMEM/F12 (Gibco, Germany), 20 µl/ml B-27 (Gibco, Germany), 1% penicillin/streptomycin (pen/strep) (Gibco, Germany) and 20 ng/ml epidermal growth factor (hEGF) (Sigma, Germany). In order to avoid the presence of explants in the cultures, cell filters (40 µm, BD-Falcon, Diagonal, Germany) were used. The cells were then plated and cultivated at 37 °C in the presence of 5% CO2. The cultures were fed every third day with different neurosphere standard media (as indicated), to avoid spontaneous differentiation.

For the differentiation (at least after 3–7 days in culture), growth factors were removed from the culture medium and single neurospheres were plated on cover slips coated with 1:100 extracellular matrix (ECM)-Gel (Sigma, Germany). Standard differentiation medium contained: DMEM-F12 (Gibco, Germany), 20 µl/ml B + 27 (Gibco, Germany), 1% pen/strep (Gibco, Germany) and 10% fetal bovine serum (FBS) (Gibco, Germany). A second approach included differentiation medium containing 10 µM retinoic acid (RA) (Sigma, Germany).

For cell passaging, the neurospheres from both OB and OE cell cultures were collected separately in 2-ml tubes and harvested for 5 min at 500g and 4 °C. After centrifugation, the medium was removed and 20 µg/ml collagenase (Sigma, Germany) was added to neurosphere standard medium. The cells were digested with gentle agitation in a water bath at 37 °C for 10 min. After digestion, the cells were harvested for 5 min at 500g and 4 °C. The cells were re-suspended in fresh standard neurosphere medium.

For preparation of postnatal OBs the whole brain was removed for better accessibility. OB neurospheres were prepared as previously described for the OE. Furthermore, OB and OE neurospheres were co-cultivated. Free-floating neurospheres obtained from the preparation of OB were selected, placed on cover slips coated with 1:100 ECM-Gel (Sigma, Germany), and culture in standard differentiation culture medium. After 24 h of differentiation, free-floating OE neurospheres were plated on the cell cultures containing differentiated OB neurospheres. The co-cultures of OB and OE neurospheres were further differentiated (72 h) for additional analysis and characterization.

Immunohistochemistry

Postnatal mice (P0–P4) were killed either by decapitation or by anesthesia with CO2 (adult). The heads were fixed in 4% paraformaldehyde (PFA)/phosphate-buffered solution (PBS−/−) (Gibco, Germany) for 1 h and further decalcified in 0.5 M EDTA-solution for 4 days. For cryoprotection, the samples were incubated in 10, 20 and 30% sucrose solution for 30 min, respectively. The heads were then embedded in Tissue Tek freezing medium and stored at −80 °C for further use. Coronal cryosections (12 μm) were then obtained using a cryomicrotome (Leica CM3050S). To block unspecific bindings, the sections were then blocked with 10% normal goat serum (NGS) (Sigma, Germany)/PBS−/− (Gibco, Germany) containing 0.1% Triton-X 100 and subsequently incubated with the primary antibody at 4 °C overnight. After washing, the sections were incubated with the appropriate secondary antibody for 1 h at room temperature (22–24 °C). For nuclei visualization, additional DAPI-staining (Sigma, Germany) was performed for 20 min at 37 °C. Immunohistochemical analysis was then performed using a confocal microscope (LSM510 Meta; Zeiss, Jena, Germany).

For staining of cells differentiated from neurospheres, PFA-fixation was performed in the culture dishes followed by treatment with blocking solution. Free-floating immunohistochemistry was performed with detached spheres collected in a reaction cup. For analysis at the confocal microscope, the stained spheres were trapped in a vaseline-bordered compartment closed by a coverslip.

The following primary antibodies were used in this study: anti-βIII-Tubulin (mouse IgG; kindly provided by Prof. Dr. C. Maier, Faculty of Medicine, Ruhr University Bochum, RUB, Germany), anti-GFAP (Sigma, Germany), anti-nestin (Chemicon International, Hofheim, Germany), anti-O4 (kindly provided by the Department of Cell Morphology and Molecular Neurobiology, Ruhr University Bochum, RUB, Germany), and anti-4′,6-diamidino-2-phenylindole DAPI (Sigma, Germany). Secondary antibodies were as follows: goat anti mouse Alexa 546, and biotinylated, CY2-, CY3 (Dianova, Germany).

RT-PCR analysis

Free-floating neurospheres or neurosphere-differentiated cells collected by direct lysis in the culture dishes were subjected to total RNA isolation using the RNAeasy MicroKit (Qiagen, Germany) according to the manufacturer. Subsequent cDNA synthesis was performed using MMLV reverse transcriptase (Fermentas, Germany) and an oligo(dT18)-primer. For PCR in a TGradient-Cycler (Biometra, Germany) the following gene-specific primer sets were used: Gαolf_for: AGTAGTGAATGTGGYMG3YTTTCTGCC; Gαolf_rev: GACCACCAAGCCTCT GGCTACCTCT; OMP_for: AAGGTCACC ATCACGGGCAC; OMP_rev: TTTAGG TTGGCAGGCTCCAC; Nestin_for: GCC ACAGTGCCCAGTTCTA; Nestin_rev: GGTCTAAACGCCTGCTGGTC; ACIII_for: TGGCAGCACCTGGCTGAC; ACIII_rev: GGGGCAGTGTAACAGAGG A; β3-tub_for: TGCTCATCAGCAAGG TGCGTG; β3-tub_rev: GGAACGGCACC ATGTTCACAG; GFAP_for: TAGCTAC ATCGAGAAGGTCC; GFAP_rev: AAG AACTGGATCTCCTCCTC. To detect the presence of olfactory receptor transcripts, PCR analysis with degenerated primers was performed. The amplification with the primer mixes IC2_for (ATGGCXTAC/TGA C/TA/CGXTAC/TG) and IC3_rev: (A/TG/CC/AG/A/TCAXGTXG/CA/TA/GAAXGC) yielded a product of 360 bp in size.

Calcium imaging analysis

Changes of the intracellular calcium concentration upon stimulation by an odorant mixture were analyzed using the calcium-sensitive ratiometric dye Fura-2AM (Molecular Probes). Prior to stimulation, cells differentiated from neurospheres were incubated for 30 min at 37 °C with 3 µM of the dye and media were replaced by Ringer’s solution (140 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgCl, 10 mM HEPES; pH 7.3) prior to stimulation. The reactivity to odorants was assayed by application of a mixture of 100 different odorants (Henkel100, Henkel GmbH, Germany). Measurements were performed using an inverse Olympus IX71 microscope, a MT20 illumination system with a 150 W Xenon arc burner and an F-view II CCD camera (Olympus Biosystems GmbH, Germany). Acquisition and calculation of fluorescence signals obtained from excitation of the Fura-2 dye at 340 and 380 nm were done using the Olympus Cell^R Imaging Software (Olympus Biosystems GmbH, Germany). Cells were stimulated for 20 s with either the odorants mixture, or forskolin as a positive control for activation of the classical olfactory pathway. ATP and high potassium, both applied for 5 s, served as controls for general viability and the presence of neuronal cells, respectively.

Electrophysiological recordings

Patch-clamp recordings from OB and OE neurosphere-differentiated cells were performed at room temperature (20–24 °C) at time intervals that ranged from 1 to 7 days post-differentiation. All cells were measured in the whole-cell configuration in an extracellular solution containing (in mM): NaCl (120), CaCl2 (2.5), KCl (4), MgCl2 (1), HEPES (10) and glucose (10). The pH was adjusted to 7.35 with NaOH and the osmolarity was corrected to 295 mOsm with glucose.

Patch electrodes were pulled from filamented borosilicate glass (GB150EFT-10, Science products) with a horizontal pipette puller (DMZ Universal Puller, Zeitz Instruments, Munich, Germany) to yield pipette resistances ranging from 3 to 7 MΩ. The pipettes were filled with a solution containing (in mM): KCl (145), MgCl2 (2), EGTA (11), CaCl2 (1), HEPES (2). The pH was adjusted to 7.4 and osmolarity corrected to 310 mOsm with glucose. The cells were voltage clamped at −70 mV holding potential and signals were filtered from 1 to 3 kHz. Sodium currents, potassium currents, and action potentials in direct differentiated neurosphere cell types were measured. Voltage-gated sodium channels were activated by a depolarizing voltage-step protocol starting from a hyperpolarizing pre-pulse (−100 mV). Potassium currents were activated by a depolarizing voltage-step protocol starting at −100 mV in the presence of the sodium channel blocker TTX (1 µM). Currents were recorded using an L/M-EPC7 patch-clamp amplifier (List, Darmstadt, Germany). Record acquisition was controlled by the Pulse Software (HEKA Instruments), which corrected on-line for liquid junction potential leak currents, fast and slow capacitance, and series resistance. Data were then exported to IGOR (Wavemetrics, Inc.) or Origin (Originlab Corporation) software packages for subsequent analyses. Only such cells were positive considered in the analysis if consistent gigaseal and series resistances throughout the experiment were obtained.

Normal distribution of data was confirmed by the Kolmogorov–Smirnov test and statistical significance was proved by the ANOVA test. All quantitative data concerning electrophysiological recordings are shown as mean ± SEM. Data were considered to be significantly different at p < 0.05.

Results

Localization and quantification of GFP-TIS-positive basal cells in the mouse olfactory epithelium

Confocal analysis of the OE obtained from TIS21 mice at postnatal (P0–P2), young (P15) and young adult stages (P21) revealed the presence of GFP-fluorescent basal cells adjacent to the lamina propria (Fig. 1a). Morphologically, GFP-positive cells are more similar to globose basal cells (GBCs) than to horizontal basal cells (HBCs). Coronal sections of adult mouse head were also prepared and semi-quantification of GFP-positive cells basal cell numbers across the epithelium at four different positions within the OE was performed: at the nasal tectum, at the first ectoturbinate, at the second ectoturbinate, and at the nasal septum. Cell numbers were moreover compared between three different positions of the head: front, middle and back of adult (P21), young (P15), and postnatal TIS21-mice (Fig. 1b). Postnatal (P0–P2) animals showed higher number of basal cells compared to adult mice; the number of basal cells was twice as high in P15 compared to adult animals. It seems that the number of GBCs rises during the developmental before it decreases in the adult OE (Fig. 1b).

Fig. 1
figure 1

Distribution and quantification of TIS21-positive basal cells in the main olfactory epithelium. Cryosections of heads of newborn (P0–P2), P15 and P21 were prepared: a representation of the basal cells in the main olfactory epithelium of adult TIS21 mice (SUS sustentacular cells, ORN olfactory receptor neurons, BC basal cells, LP lamina propria). b Mean basal cell density in P0, P15 and adult TIS21 mice

Full size image

Cultivation and differentiation of olfactory neurospheres

Neurospheres were generated by dissociation of the OE and OB of postnatal TIS21 knock-in mice (P0–P4). After 3–10 days in culture (DIV) first globular structures, that resemble neurospheres, were observed (Fig. 2a). The beneficial effects of growth factors on neurogenesis in vitro and in vivo conditions have already been described in several studies [10, 35]. Regarding the positive influence of growth factors on cultured olfactory neurospheres, two different neurosphere culture conditions were tested in this study, namely, EGF alone, and EGF together with FGF in order to first, define their benefits for the generation of olfactory neurospheres and second, establish a stable culture that allows the long-time cultivation and differentiation of OB and OE neurospheres.

Fig. 2
figure 2

Cultivation and differentiation of olfactory neurospheres. a Neurosphere formation (3–10 days in culture) of dissociated olfactory bulb (OB) (left) and olfactory epithelium (OE) (right) from neonatal TIS21 knock-in mice (P0–P4). b Combining EGF and FGF leads to significant increase of neurosphere formation from the OB but not from the OE. c The combination of EGF and FGF leads to a significant increase of secondary postnatal OB neurospheres. In contrast, the combination of FGF and EGF had no significant impact on neurosphere formation from the OE. d The application of retinoic acid (RA) had no significant effect on neurosphere formation

Full size image

OB neurospheres appeared after 3 days in the culture. The mean number of primary OB neurospheres obtained from the dissociation of postnatal OB was 33.2 ± 6.5 (n = 15 cultures) in the presence of EGF, whereas the mean number of neurospheres cultivated with both EGF and FGF was significantly higher (62.4 ± 8.2, n = 15 cultures, p = 0.02). In contrast to the results obtained for OB, the number of OE neurospheres was not significantly affected by the addition of FGF to the cultures (24.0 ± 5.3, n = 15 cultures) when compared to the mean number of neurospheres cultivated with only EGF (15.2 ± 4.6, n = 15 cultures, p > 0.05) (Fig. 2b).

The next step was to determine the ability of passaging both types of postnatal olfactory primary neurospheres. The quantification of neurospheres was performed under EGF and also under EGF + FGF conditions (Fig. 2c). The dissociation of postnatal OB primary neurospheres gave rise to secondary neurospheres after four DIV. It was possible to repeat this process up to three times. The parallel addition of EGF and FGF significantly increased the mean number of secondary postnatal OB neurospheres (91.0 ± 13.4, n = 12 cultures, p = 0.03). An increase was also observed for cultures treated only with EGF, but the difference was not significant (54.5 ± 9.6, n = 12 cultures, p > 0.05). The same tendency was also observed for the third generation of OB neurospheres. After the third passage, the mean number of postnatal OB quaternary neurospheres (n = 8 cultures) did not show significant changes in comparison with primary neurospheres (p > 0.05) (Fig. 2c).

Moreover, the passaging of OE neurospheres was successfully performed. This process was more complicated than the passaging performed with OB neurosphere cultures. At this point of our study, only secondary neurospheres could be generated from postnatal and OE neurospheres. The mean number of postnatal OE secondary neurospheres in the presence of EGF (14.2 ± 3.1, n = 10 cultures, p > 0.05) or EGF and FGF (23.5 ± 5.6, n = 10 cultures, p > 0.05) did not display significant differences when compared to the mean number of primary neurospheres cultures (Fig. 2c). Regarding the beneficial influence of combining both growth factors for the formation and passaging of OB and OE neurospheres, the culture medium containing EGF and FGF was used as the standard culture medium for the long-term cultivation of the olfactory neurospheres and the following experiments.

Additionally, we were interested in comparing the effect of the retinoic acid (RA) on differentiated olfactory neurospheres. Therefore, we added RA to our standard differentiation medium, which already contained 10% FBS. Despite beneficial influences of RA in the neurogenesis of stem cells in vitro being described in many studies [1, 27, 41], we did not observe any significant differences between untreated and cultures treated with RA (Fig. 2d).

Immunohistochemical characterization of cells differentiated from olfactory neurospheres

First, the morphology of the neurospheres derived from murine OB or OE was compared. In a second approach, the cells differentiated from olfactory neurospheres were analyzed with regard to their identity as neurons, astrocytes, and oligodendrocytes. Therefore, we performed immunohistochemical analysis using various specific antibodies. Most neurospheres in the OE cell cultures prepared from TIS21 mice showed a clear GFP-fluorescence, which confirmed the presence of basal cells in the respective cultures (Fig. 3a).

Fig. 3
figure 3

Neurosphere differentiation results in the formation of morphologically different cells. a Neurospheres developed from the OE of TIS21 mice contain high numbers of GFP-positive cells derived from basal cells. b Subsequent to differentiation the appearance of different types of cells were observed in postnatal OB neurosphere cultures, typically shaped like oligodendrocytes (red arrow) and neurons (bipolar shaped, blue arrow and pyramidal shaped, green arrow). c The differentiation of postnatal OE neurospheres resulted in neuron-like bipolar-shaped cells (blue arrow). d The adhesion of OE neurospheres (yellow arrow) was clearly enhanced in the presence of OB neurospheres (blue arrow). Scale bars 20 μM

Full size image

Already 2 h after exchange of the neurosphere growth-medium by standard differentiation medium, most OB-derived neurospheres adhered to the ECM-coated coverslips. After overnight incubation, cells differentiated from the OB-derived neurospheres (Fig. 3b). OE-derived neurospheres poorly adhered when subjected to differentiation medium alone and the appearance of differentiation cells took at least 72 h (Fig. 3c). The a priori differentiation of OB-derived spheres in the same dish obviously facilitated the differentiation of OE neurospheres such that the probability of adhering spheres was clearly enhanced (Fig. 3d).

Upon neurosphere differentiation several cell types emerged. According to morphological aspects such as a bipolar or pyramidal shape, the cells appearing after differentiation of OB neurospheres might be neurons, astrocytes and oligodendrocytes, (Fig. 3b). In contrast, the differentiation of the OE-derived neurosphere cultures yielded only bipolar-shaped cells (Fig. 3c). Immunohistochemical characterization revealed that several cells in both, OE and OB neurosphere differentiation cultures expressed the neuronal marker β-III tubulin (Fig. 4a). In several cases, β-III-tubulin positive cells from both cultures formed huge networks, which were mostly located in the center of the differentiated spheres. In the periphery, more exactly, in the corona of the settled neurospheres, glial structures could be identified, as confirmed by GFAP-staining (Fig. 4b).

Fig. 4
figure 4

Cells differentiated from postnatal OB neurospheres are immunopositive for β-III-tubulin and GFAP. a Anti β-III-tubulin antibody staining confirmed the presence of neurons in the differentiation cultures, whereas the anti-GFAP antibody b confirmed the presence of glia cells after differentiation. Scale bars 20 μM

Full size image

Calcium imaging analyses reveal the presence of neurons in OE neurosphere differentiation cultures

To investigate the physiology and functionality of the putative neurons differentiated from both, OB and OE neurospheres, with respect to their sensory properties, calcium imaging experiments were performed. We used a mixture of hundred odorants, previously described as the Henkel 100 mixture [16]. Cells differentiated from olfactory neurospheres were loaded with the calcium-sensitive dye FURA-2AM and subjected to calcium imaging analysis (Fig. 5a). The application of the odorant mixture had no effect on cells differentiated from OB neurospheres. Application of ATP at the end of the experiment proved the general excitability of the cells (Fig. 5b). In addition, the application of high potassium currents in our cultures provoked calcium transients (Fig. 5c), which together with the immunohistochemical results indicate that the cells differentiated from OE neurospheres had neuronal characteristics.

Fig. 5
figure 5

Calcium imaging analysis of cells differentiated from olfactory neurospheres. a Transmission light and fluorescence image of Fura-2 loaded cells differentiated from an OE neurosphere. Calcium imaging recordings of cells. b Differentiated from OB neurospheres, c spontaneously differentiated from OE neurospheres, and d differentiated from OB and OE neurospheres co-cultures. Scale bar 20 μm

Full size image

Electrophysiological characterization of cells differentiated from olfactory neurospheres

Morphological characteristics indicated the presence of putative neurons, astrocytes, and oligodendrocytes in our cell cultures. Moreover, the immunohistochemical experiments confirmed the presence of neuronal and astroglial structures in the cultures differentiated from OB and OE neurospheres. To verify their identity and in order to investigate their physiology, electrophysiological recordings using the patch-clamp technique (whole-cell mode) were performed. First, the general excitability of the respective cells differentiated from both OB and OE neurospheres derived from postnatal mice was analyzed by measuring electrically evoked sodium currents, potassium currents and action potentials. For the electrophysiological recordings, the cells were directly differentiated as previously described. Recordings were performed between 24 and 72 h after differentiation. For bipolar-shaped cells from OB, neither sodium currents nor action potentials were detected. Interestingly, the OB pyramidal-shaped differentiated cells showed excitable properties. About 24% of the OB pyramidal-shaped showed sodium currents (n = 17 cells; Fig. 6a). To identify other ion channels underlying active membrane properties in cells differentiated from postnatal OB neurospheres, further voltage-clamp experiments were performed. Here, the outward currents composed of two K+ channel populations were identified, a non-inactivating delayed rectified current (I DR), and fast inactivating A-type current (I A) (n = 12 cells; Fig. 6b).

Fig. 6
figure 6

Electrophysiological characterization of olfactory neurospheres. a Pyramidal-shaped neuronal cell differentiated from a postnatal OB neurosphere with patch-pipette (−70 mV membrane holding potential). Scale bar 20 μm. b Whole-cell voltage-clamp recordings of sodium currents of cells differentiated from postnatal OB neurospheres. The current shows a peak between −30 and 10 mV (n = 17 cells). c Detection of K+ currents in cells differentiated from postnatal OB neurospheres. Whole-cell voltage-clamp recordings of total [I (total)] and delayed rectifier (I DR) potassium currents were carried out in the presence of 1 μM TTX. I DR potassium currents were isolated using a −40 mV prepulse to inactivate A-type potassium channels (IA). Current/voltage (I/V) relationship of I DR (a) and I A (b) (n = 12 cells)

Full size image

To complete the electrophysiological characterization of the cells derived from differentiated olfactory neurospheres, recordings in the current-clamp mode were performed. Here, 33% of cells recorded from postnatal OB differentiated neurospheres fired action potentials, which were completely blocked by the addition of 1 μM TTX to the bath solution of the cultures (n = 15 cells). The duration (4.2 ± 0.7 ms, n = 15 cells) as well as the amplitude (37.3 ± 3.1 pA, n = 15 cells) of the action potentials was calculated (Fig. 6c).

Finally, we were interested in the physiological properties of the bipolar-shaped cells generated from differentiated postnatal OE neurospheres. However, these cells did not show any excitable properties during the electrophysiological measurements. Neither sodium currents nor action potentials were observed.

PCR analysis and quantitative real-time PCR (RT-PCR) of olfactory neurospheres

To get an idea whether the cellular activities observed in the calcium imaging and patch-clamp experiments may involve OR pathways, the expression status of components of the OR signaling cascade was analyzed by real-time PCR experiments in neurospheres and pools of differentiated cells. For this purpose we used intron-spanning primers specific for classical components of the OR pathway. Our results revealed that both OE and OB-derived neurospheres express β-III-tubulin and nestin, indicating the presence of neuronal structures in our cultures (Fig. 7a). Furthermore, both types of olfactory-derived spheres expressed the olfactory marker protein (OMP) and the adenylate cyclase type III (ACIII), which are both characteristic for olfactory structures. The ACIII was present in all samples, neurospheres and pools of cells at any time (Fig. 7b). OE but not OB neurospheres, additionally expressed Gαolf, which is essentially involved in the olfactory signaling cascade. As a control for positive Gαolf expression, intact OB tissue was used. With the quantitative RT-PCR analysis we were furthermore able to observe that Gαolf expression disappeared after differentiation. All Gαolf expressing cells also displayed the expression of the cyclic nucleotide gated channel 2 (CNGA2), which is also an important and specific component of the olfactory signaling cascade. CNGA2 was detectable in the OB tissues and in OE neurospheres but also disappeared after differentiation (Fig. 7c).

Fig. 7
figure 7

PCR and quantitative real-time PCR analysis (qRT-PCR) of olfactory neurospheres. Olfactory neurospheres were collected after 10 days in culture. a OB and OE neurospheres are positive for OMP, β-III-tubulin, nestin and ACIII. OE neurospheres additionally were positive for Gαolf. b qRT-PCR analysis reveals mRNA for Gαolf, ACIII, CNGA2 in OE neurospheres but only a very low level in cells differentiated from neurospheres

Full size image

Discussion

In the last years, the cultivation of free-floating neurospheres has become a suitable tool for the isolation and characterization of olfactory neurospheres. The aims of this study were twofold: first, the optimization of the cultivation protocol, and second, the morphological and functional characterization of cells differentiated from olfactory neurospheres derived from the olfactory bulb and the olfactory epithelium of postnatal GFP-TIS21 mice. The presence of neuronal precursor cells in the cultures was confirmed by the immunofluorescence of the antiproliferative protein TIS21 coupled to GFP. Furthermore, this protein helped us to distinguish the OE neurospheres from the OB neurospheres in our co-culture approach. In the first part of this study, the effect of EGF alone and in combination with FGF on the number of postnatal olfactory neurospheres was characterized. The OB neurospheres emerged after 3–4 DIV. The mean number of these cell-aggregates cultivated in the presence of combined growth factors (EGF + FGF) was significantly increased when compared to cultures containing EGF alone. The OE neurospheres appeared between 7 and 10 DIV and the addition of FGF increased their mean number in comparison to EGF alone. Moreover, the number of neurospheres was positively influenced after multiple passages of OB neurospheres (until passage 3), and only one passage by the OE, when both growth factors were added to the cultures.

EGF and FGF play an important role in the proliferation of neural precursor cells isolated from different regions of the brain at postnatal and adult developmental stages [10]. Regarding the OE, several growth factors that regulate stem and progenitor cell proliferation have been identified so far. Horizontal basal cell (HBC) proliferation is stimulated by EGF and TGFα [13, 32] and globose basal cell (GBC) proliferation is stimulated by FGF2 [11, 35]. In vivo studies performed by Hsu and co-workers described the expression of FGF receptor subtypes (FGFR1 and FGFR2) in GBCs, or the FGF2 expression in GBCs and supporting cells [18]. It has been shown that EGF and FGF2 are required to generate multipotent neurospheres from the olfactory mucosa [33, 34, 39].

One possible explanation for the enhanced number of neurospheres in the presence of combined EGF and FGF could be that both growth factors interact synergistically to enhance the proliferation of the olfactory neurospheres. Tropepe and co-workers tested this hypothesis for stem cells differentiated from neurospheres during brain development [40]. They could show that the combination of high and low concentrations of EGF and FGF in the developing mouse telencephalon does not interact synergistically to promote neural stem cell proliferation and suggested that both growth factors can support this process independently [40]. Regarding this, Tropepe and co-workers suggested that distinct populations of neural stem cells exist, which are either responsive to EGF or FGF [40]. We therefore suggest a similar effect at least for the olfactory epithelium-derived stem cells, where the presence of EGF-sensitive stem cells (HBCs) and an independent population of FGF-sensitive stem cells (GBCs) might explain the increased number of neurospheres cultivated in the presence of both growth factors.

In the second part of this work, we aimed on the optimization of formerly established differentiation culture conditions to facilitate the further functional characterization of the olfactory neurosphere-differentiated cells. For this purpose growth factors were removed from the culture medium and serum (FBS 10%) and/or other factors such as retinoic acid (RA) were added. The influence of RA for neurogenesis of neural stem cells of adult hippocampus has already been shown [43]. A study published previously by Wang and colleagues showed that RA treatment increases neurogenesis in SVZ neurosphere cultures from P15 mice in a concentration-dependent manner by increasing the mean number of neurons, decreasing the mean number of astrocytes but not affecting the mean number of oligodendrocytes [41]. In our work, the treatment of the postnatal olfactory neurospheres with RA did not provoke a significant effect in the differentiation pattern of OB and OE neurospheres. In terms of postnatal rodent forebrain, retinoid-binding proteins are expressed in the OB [45] and it also known that RA receptors persist throughout adulthood in the OB [27]. However, although retinoic signaling components persist in the mammalian forebrain their role in postnatal brain development is poorly understood. A paper published in 2012 by Sharow and colleagues [37] investigated the stability of RA under different culture conditions: in media supplemented with fetal bovine serum (FBS) or in chemically defined, serum-free media. The stability of RA was determined to be greatly reduced in serum-free media as compared with serum-supplemented media and it was shown that stabilization was dependent on protein concentration. These results could be extrapolated to postnatal cultures and should, theoretically, fit with ours. However, it has also to be taken into account that factors like the isomerization state of RA and its interactions with other potential differentiation factors in growth media could influence developmental processes of the cell culture. Additionally, it is also known that embryonic cell cultures with RA are susceptible to high temperatures and light. High temperatures and exposure to light cause oxidative processes that could influence negatively on the result of the study [26, 38]. The strict control of the culture conditions and the use of the correct protein concentration are the determining factors to obtain positive results in this type of studies. These two factors should be monitored and controlled in the future for a better understanding of RA influence on postnatal stem cell cultures.

At 24 and 72 h post-differentiation, several types of cells were distinguished, and—based on their morphological characteristics—assessed to probably be neurons (pyramidal-like and bipolar-like shaped cells), astrocytes, or oligodendrocytes. Immunostainings performed with these cells were positive for neuronal (β III-tubulin and nestin) and glia cell (GFAP) markers, respectively.

The intermediate filament protein nestin has been widely used as a marker for neural progenitor stem cell in the nervous system. In rats, the expression of nestin has been detected in the basal region of the mature OE. However, its expression is restricted to the more basal parts of sustentacular cells [12]. The absence of nestin expression in cells differentiated from postnatal OE neurospheres, suggests that these cells are not proliferating like basal cells of the epithelium. In order to confirm that the potential newly build neurons were functional too, voltage- and current-clamp measurements in the whole-cell mode of the patch-clamp technique were performed around 24 and 72 h after differentiation. As we were able to show, differentiation of postnatal OB neurospheres did indeed generate functional neurons. Cells differentiated from OB neurospheres showed inward Na+ currents and outward K+ currents. The cells differentiated from postnatal OB neurospheres showed two types of outward K+ currents, namely delayed-rectifier and A-type currents. In former studies, the occurrence of this type of K+ currents during neuronal differentiation has already been documented [4]. It is also known that voltage-gated K+ currents occur ubiquitously in the development of the nervous system. It is well known that cells require voltage-gated K+ channels to proliferate [15, 30, 36, 44]. We therefore suggest that the support of cell proliferation is the principal function of the K+ currents, and thus of different subpopulations of this type of currents. Finally, the neurons identified in the postnatal OB neurosphere cultures showed action potentials that were completely blocked by addition of 1 μM of TTX into the bath solution. However, the relative long duration and low amplitude of these action potentials suggest that these cells are not completely mature and therefore premature in function. In this study, it was not possible to detect sodium and potassium currents, or action potentials in postnatal OE neurosphere-differentiated neuronal bipolar-shaped cells, indicating that these cells were not physiological active.

These findings nicely fit the results from real-time PCR analysis that revealed a lack of important factors of the olfactory signaling cascade, namely Gαolf and CNGA2 in the differentiated cells. Most probably, the differentiation cultures do not provide important factors that trigger the development of functional neurons, and/or important direct contacts of the newly generated cells to an intact stimulating environment are essentially missing.

In summary, the functionality of the pyramidal-shaped neurons generated by differentiation of postnatal OB neurospheres was confirmed. To complete the electrophysiological characterization it would be interesting to determine what kind of synapses is first established at the newly built neurons. Furthermore, in this type of pyramidal-shaped neurons basal and olfactory cell markers were expressed. Another set of experiments performed revealed that this type of neurons responds to some odorants, suggesting that these cells may be related to olfactory sensory neurons (OSNs). However, further analysis of the electrophysiological characteristics will have to be performed to unravel the identity of the cells differentiated from OE neurospheres.