The Phenomenon of Compensatory Cell Proliferation in Olfactory Epithelium in Fish Caused by Prolonged Exposure to Natural Odorants

It was previously shown that activation of the processes of neurogenesis in the olfactory epithelium (OE) can be caused after intranasal administration of toxic or neurotrophic factors, after axon transection, or as a result of bulbectomy. Our study showed for the first time that a significant increase in olfactory cell renewal can also occur in animals due to periodic chemostimulation with natural odorants (amino acids and peptides) for 15 days. Using electron and laser confocal microscopy in fish (Paracottus knerii (Cottidae), Dybowski, 1874) from Lake Baikal, we showed that periodic stimulation of aquatic organisms with a water-soluble mixture of amino acids and peptides causes stress in OE, which leads to programmed death cells and compensatory intensification of their renewal. We estimated the level of reactive oxygen species, number of functionally active mitochondria, intensity of apoptosis processes, and mitosis activity of cells in the OE of fish in the control group and after periodic natural odorants exposure. This study showed that new stem cells are activated during enhanced odor stimulation and subsequent degenerative changes in the cells of the sensory apparatus. Those new activated stem cells are located in previously proliferatively inactive regions of OE that become involved in compensatory processes for the formation of new cells.

Neurogenesis in the olfactory epithelium (OE) of animals is maintained throughout their life due to the proliferative activity of regional stem cells (SC) that produce various types of cells in OE [1][2][3][4] . The regeneration ability is prospective for the OE to serve as a source of SC for detailed study and therapeutic use 4 . Model experiments indicate that various forms of sensory deprivation can significantly activate the natural rate of neurogenesis inherent for intact animals. Thus, after destructive changes in the epithelium resulted from exposure to toxic agents, axon transection or bulbectomy, cell proliferation increases [5][6][7][8][9][10][11] . Moreover, olfactory neurogenesis can also be activated by intranasal administration of neurotrophic factors 12 . Considering the prospective use of these technologies in medicine, new information on the mechanisms of olfactory neurogenesis is mainly intensively studied in mammals. In this regard, fish have been studied in less detail. However, the cytological and molecular mechanisms lying at the root of olfactory neurogenesis in fish may also be of interest for the study of neuroplasticity and for development of treatment methods for neurodegenerative diseases [13][14][15] .
It has been established that olfactory neurogenesis processes in fish can also be activated with similar methods of sensory deprivation [16][17][18][19][20][21] . It should be noted that all these methods for activating neurogenesis are mostly invasive and traumatizing, which limits their application in practical medicine.
In this publication we present a new approach for stimulating of cell proliferation in OE, which allows one to avoid these complications. In order to activate this processes, we proceeded from the two main facts: (1) amino

Results
Based on structural features, spatial arrangement, and specifics of molecular markers, in fish five types of OSNs were identified, which enable the perception of different types of odorants: ciliated OSNs, microvillous OSNs, crypt neurons, pear-shaped OSNs, and kappe neurons [31][32][33][34] . Ciliated and microvillous cells are the most numerous cells, that are sensitive to amino acids and taurocholic acid 25,35 .
The other types of OSNs are minor in terms of quantity; however, they have functional differences. Crypt neurons perceive bile acids, pheromonal signals, and amino acids [36][37][38] , while pear-shaped OSNs specialize in the reception of nucleotides, which arrive from nutritional sources into the aqueous medium 34 . The functional specialization of kappe neurons has not been determined yet.
This study was conducted on Paracottus knerii (Cottidae) (Dybowski, 1874) ( Supplementary Fig. 1), an endemic representative of the ichthyofauna of Lake Baikal. In P. knerii, the olfactory rosette is 2-2.5 mm in size and, as in other Baikal Cottoidei, has 5-6 folds ( Supplementary Fig. 2). The OE has a structure typical of that of teleosts 39,40 and other Cottidae 41,42 and consists of receptor, supporting, and basal cells. Based on transmission microscopy data, the OE mainly contains two types of receptor cells, namely, ciliated and microvillar receptor cells (Fig. 1A,B). It was established that, compared with that of the control (Fig. 1C), the OEs of the experimental fish, which were subjected to intensive odorant exposure to a mixture of amino acids and peptides, showed signs of densification and fragmentation of nuclei in epithelial cellular elements, which serve as an indicator of the final phase of PCD 43 . The nuclei of these cells looked like fragmented round formations 0.5 to 3 µm in size with condensed, intensely DAPI-stained chromatin (Fig. 1D). In 10-15% of receptor neurons, endoplasmic reticulum channels looked fragmented and, in some areas, widened, and dendrites contained mitochondria with distinct signs of partial or complete swelling. Compared with that in the control (Fig. 1E), these organelles had a clear matrix with loose electron-dense deposits and destructive changes in the cristae (Fig. 1F). We also observed individual cases of mitochondrial swelling in supporting cells (Fig. 1G). To objectively assess the state of the mitochondria, we used stained the OE for functionally active mitochondria with a fluorescent reagent (MitoTracker Orange) that selectively binds only native mitochondria with a conserved membrane potential. Compared with that in the control (Fig. 1H), the level of florescence in the mitochondria in the experimental fish was significantly lower (Fig. 1I). This fact was further confirmed by a comparative analysis of the volume occupied by the fluorescent signal in the mitochondria in OE cells of the control and experimental fish (Fig. 1J). The plot shows that the volume occupied by the mitochondria decreased by a factor of 6.2 (р u ≤ 0.05) between the experimental and control fish.
It is known that increasing the concentration of Ca 2+ in the cytoplasm increases the production of reactive oxygen species (ROS) and caspase-dependent apoptosis of neuronal cells 44 .Therefore, we investigated the intensity of ROS formation in the OE of the control and experimental fish using the CellROX deep red reagent, which forms fluorescent products after a covalent interaction with free radicals in the cell. In the control preparations, the ROS products appeared as fine deposits that fluoresced in the red spectrum and were spread evenly across the width of the OE ( Fig. 2A-C). In the experimental fish, however, these products appeared as more concentrated individual clusters as well as fine deposits (Fig. 2,D-F). Comparative statistical processing of the Z-stacks of the images with ROS products suggests that, compared with that in the control, this indicator in the experimental fish increased by a factor of 3.5 (р u ≤ 0.05) (Fig. 2G). We hypothesize that such an increase in ROS production causes oxidative stress in the cell, since we observe this process along with degenerative changes in the olfactory epithelium in comparison with the control.
To quantitatively assess PCD processes that may have been activated in the OE of the experimental fish during prolonged odorant stimulation, we used the TUNEL method 45 . This method allows cytochemical detection of DNA microfragmentation in cells during PCD. Figure 3A-C show a large OE fragment that exhibited scarce green-stained nuclei in cells with fragmented DNA in the control fish. By comparison, the experimental fish demonstrated a considerable increase in the number of cells undergoing PCD. These TUNEL-positive cells were distributed over the entire width of the epithelium, often forming more dense groups along the edge of the olfactory fold (Fig. 3D-F). The volume occupied by apoptotic nuclei was statistically analysed to show objectively that, compared with that in the control, the proportion of dead cells in the OE of the experimental fish increased by a factor of 4.5 (р u ≤ 0.05) (Fig. 3G). Since degenerative changes were observed in both receptor and supporting cells in the experimental fish, as determined by electron microscopy, the plot can be supposed to display the amplification of PCD processes in both types of cells. Selective degenerative changes in axons in the basal OE layers also provided morphological evidence of the selective death of receptor cells under oxidative damage in intracellular structures. Compared to control (Fig. 3H), the axons in these OSNs were widened to various degrees, lacked neurofilaments, and showed a markedly lower density of cytoplasm (Fig. 3I). Axon swelling with destruction of the cytoskeleton components is one of the universal signs of degenerative changes in neurons not only in OE 46,47 , but also among neurons of the central nervous system 48 .
Based on the above evidence of significant degenerative changes in OE, we hypothesized that compensatory cell repair processes were activated in the OE of the experimental fish. To confirm this, we used bromodeoxyuridine (BrdU) labelling 49 to compare the proliferative activity of cells in the OE of control and experimental fish subjected to prolonged odorant stimulation.  . Notably, the clusters composed of a small number of cells (9 or fewer) occupied 95% of the total volume of the tissue. The larger clusters (10-20 cells) of labelled nuclei accounted for only a small portion of the OE volume (5%) (Fig. 4D). The total number of label-incorporating cells in the control group was 5.06 ± 1.67 per 10 5 µm.
By comparison, the OE of the fish subjected to prolonged odorant stimulation with a heterogeneous mixture of amino acids and peptides exhibited a considerable enhancement of cell proliferation processes (Fig. 4E-G; Supplementary video 2). Compared with that in the control, there was a substantial (by a factor of 7.8; р u ≤ 0.05) increase in the number of areas populated by larger clusters of 10-20 BrdU-incorporating cells across the width of the OE. Simultaneously, the volume occupied by newly formed cells grouped into small clusters (up to 10 cells), decreased substantially (by a factor of 1.5; р u ≤ 0.05) (Fig. 4H). The large clusters (10-20 cells) of BrdU-incorporating cells were located in different OE regions, but the highest concentration was in the periphery of the olfactory fold. It is characteristically that in that areas we found high amounts TUNEL-positive cells (Fig. 3D-F). Our calculations showed that the rate of cell proliferation processes in the OE of the experimental fish was higher than that in the control group by a factor of 3.4 (р u ≤ 0.05) (Fig. 4I).

Discussion
Our research shows that, in addition to the already known methods, the increased cell proliferation in OE of fish may also be caused by prolonged periodic exposure to natural water-soluble odorants. Given the monospecific nature of OSNs, a diverse set of these molecules can stimulate a large number of OSNs. As a result, swelling and a significant decrease in the functional activity of mitochondria, increased production of AFC, and in the end -PCD activation were observed in OE cells. We believe that these structural-functional cell rearrangements underlie compensatory activation of cell renewal processes in OE. How does the chain of cytochemical events, which leads to OE cell renewal, develop?
It is known that repetitive and prolonged exposure to odors causes an adaptive decrease in sensitivity, or habituation, in the olfactory system of animals 50,51 . This adaptation occurs at different levels of organization of the www.nature.com/scientificreports www.nature.com/scientificreports/ olfactory analyzer, along the entire path of sensory information processing 52 . The primary element in olfactory adaptation is the processes that occur in OSNs 53 .
It was found that the bonding of odorant molecules with OSN receptors activates adenylate cyclase, leading to an increase in the intracellular concentration of cyclic adenosine monophosphate (cAMP) and opening of cyclic nucleotide-gated (CNG) channels, through which Na + and Са 2+ ions penetrate into the cell 54,55 . This process serves as a trigger mechanism for depolarizing the cell. Moreover, in the cell activation process, phospholipase C together with diacylglycerol and inositol triphosphate also increase the intracellular concentration of Са 2+ by means of intracellular depots 56 . A short-term increase in the Са 2+ concentration opens Са 2+ -activated chloride (CAC) channels that amplify the CNG channel signal 57 .
At present, some data exist on the mechanisms underlying the short-term adaptation (STA) of OSNs to odorants. It was found that the STA can be caused by removal of intracellular Са 2+ 58 ; by a decrease in the quantity of phosphodiesterase, which hydrolyzes cAMP and terminates its action 59 ; by modulation of the cAMP-gated channel by Ca 2+ feedback 60 ; by cyclicity of the functioning of G-linked receptors (GPCR) 53 ; and by other mechanisms. A decrease in the sensitivity of OSNs may occur in the presence of a prolonged stimulus as well as a short-term one. However, the processes that occur in OSNs exposed to prolonged odorant stimulation, either continuous or periodic, as of now remain unclear.
Electrophysiological studies conducted on cells of different types have shown that mitochondria are very sensitive to an excess of Ca 2+ ions in the cytoplasm, especially under oxidative stress [61][62][63] . Calcium ions activate phospholipase A2, which is localized to the inner mitochondrial membrane, leading to phospholipid hydrolysis and a destabilization of the membrane. As a result of these processes, the membranes become more permeable to cations, which accumulate with phosphate in the matrix, leading to the accumulation of water inside the mitochondria which causes the mitochondria to swell, and interfere with oxidative phosphorylation processes 64   www.nature.com/scientificreports www.nature.com/scientificreports/ factor), a direct cause of PCD, into the cytosol [65][66][67][68][69][70][71] . In turn, the enhancement of the processes of PCD can be a trigger for activating the processes of cell renewal in OE.
Thus, the developed method of noninvasive activation of cell reneval in OE allows one to generate an enriched pool of poorly differentiated neural cells as a source for transplantation. Cultivation followed by autological transplantation of these cells 72,73 can be used in different fields of medicine. Presently, neural SC transplantation is applied to the development of methods for treating traumatic lesions 74,75 of the nervous system, ischaemic brain damage 76 and neurodegenerative processes [77][78][79][80] . The olfactory neural SC have neurocompetency and multipotency 81,82 , and their regenerative potential is shown in experiments with spinal cord regeneration in rats 73 .
In addition, it is known that dementia and Parkinson disease, which occurs in humans usually due to aging, as a result of an injury, or because of other factors, is accompanied by a decrease in the quantity of OSNs and olfactory dysfunction, leading to a disadaptation of the person [83][84][85] . In light of the data obtained, the proposed method for enhancing cell proliferation processes may bring about a substantial renewal of receptor cell populations and increase chemosensitivity in the peripheral part of the olfactory system. Thus, this approach can be used to develop methods for increasing the number of OSNs in patients with neurodegenerative diseases in which anosmia is present.

Materials and Methods
Animals. We studied neurodegenerative changes, PCD and compensatory renewal in the OE after the prolonged periodic stimulation with odorants of the stone sculpin P. knerii (Cottidae) (Dybowski, 1874) ( Supplementary Fig. 1), an endemic representative of the Baikal ichthyofauna. P. knerii inhabits the entire shallow water zone of Lake Baikal to a depth of 200 m. The main prey items of P. knerii are amphipods, the larvae of amphibiotic insects, and zooplankton. Fish were caught in Southern Baikal by nets (depth 1.5-5.0 m) in spring (March-April) 2008-2018. The caught sculpins were kept in aquariums of the Baikal Museum with running water (4° С) from a depth of 400 m from Lake Baikal. The age of P. knerii was determined by the optical structure of the sagittal otoliths. The otoliths were extracted from the fishes' skulls and prepared for study using standard methods 86,87 . The sagittal otoliths of P. knerii clearly show alternating semi-transparent and opaque rings corresponding to annual growth. The opaque rings were counted in the frontal plane from the nucleus to the proximal edge. Two lab workers independently counted the opaque rings in the frontal plane from the nucleus to the proximal edge using a passing light microscope (24-60×). When differences in age calculation appeared, the otoliths were re-examined by both lab workers until a consensus was reached. The morphological features, sex, and age of the P. knerii are given in Table 1.
To enable adaptation, the fish from the control and experimental groups (n = 25 in each group) were kept in aquariums (13 liters) with continuously renewed water for 7 days at 4 °C. Then, we stopped renewing the water and added a heterogeneous mixture of amino acids and peptides (0.0002% bacteriological grade Tryptone, a pancreatic digest of casein, VWR Life Science, AMRESCO, USA, cat no. Am-J859-0.25) to the aquariums with the experimental fish. The percentage composition of this mixture has been described previously 88 . Based on these data, a 0.0002% Tryptone solution contains amino acids as odorants in the concentration range from 6.60 × 10 −7 to 1.02 × 10 −5 M, which corresponds to the optimal range [22][23][24][25][26][27] .
The concentration of the substances was sustained for 15 days, except during the nighttime (lasting 12 h), when the fish were kept in a constant flow of running fresh water. Periodic renewal of fresh water is a necessary condition for the maintenance of Baikal fish. The control fish were kept in separate aquaria under the same conditions but were not exposed to the mixture of substances. When the experimental period passed, we extracted the olfactory rosettes from the control and experimental fish and prepared them for transmission and laser confocal microscopy (LCM).
In the OE of both the control and experimental fish, we cytochemically determined the volume occupied by ROS products, the volume of functionally active mitochondria, and the intensity of PCD with TUNEL 45 and proliferative activity with BrdU 49 . Proliferative activity was assessed 12 h after intraperitoneal BrdU injection. The ultrastructural features of the preparations from the two groups were investigated using transmission electron microscopy. Both in control (n = 25) and in experiment (n = 25) for each technology LCM (3 methods) we used 5 fishes (10 olfactory organs), and for transmission electron microscopy we used 10 fishes (20 olfactory organs).
All www.nature.com/scientificreports www.nature.com/scientificreports/ lead citrate, sections were analyzed with a transmission electron microscope Leo 906E at an accelerating voltage of 80 kV. Microscope images were taken with a MegaView II digital camera with the MegaVision software package (Soft Imaging System GmbH, Germany).
Laser confocal microscopy. Since the olfactory organs of P. knerii are small (2-3 mm), their cytochemical features were studied entirely on the glass slide using an LSM 710 (Carl Zeiss) laser confocal microscope.
PCD was evaluated using the TUNEL method 45 with a Click-iT TUNEL Alexa Fluor 488 Imaging Assay Reagent Set (Thermo Fisher Scientific Inc., USA), following the manufacturer's recommendations. The process of material preparation included the following steps: fixation of the tissue in 4% paraformaldehyde, permeabilization in 0.25% Triton X-100, incubation in TdT (terminal deoxynucleotidyl transferase) reaction mixture with ethynyl-dUTP, click-reaction with Alexa Fluor 488 azide (ex/em 495/519 nm), and total nuclear staining with Hoechst 33342 (ex/em 350/460). Stained olfactory rosettes were placed on glass slides in a drop of ProLong Gold antifade reagent (Thermo Fisher Scientific Inc., USA). The preparations were then studied using lasers: channel 1, 405 nm, Ch1: 410-499; channel 2, 488 nm, Ch2: 499-725. Сell nuclei with fragmented DNA in the prepared slides fluoresced in the green region of the spectrum, and total nuclear DNA fluoresced in the blue region.
Proliferating cells in the OE were detected using the FITC BrdU Flow Kit (Becton, Dickinson and Company BD Biosciences, USA, Cat. No. 559619). This method is based on the incorporation of bromodeoxyuridine (BrdU) into DNA during the S-phase of the cell cycle 49 . The fish were labelled in vivo by intraperitoneal injection of BrdU in Dulbecco PBS (0.5 mg per 10 g body mass). Twelve hours after the BrdU injection, the olfactory rosettes were removed, fixed, permeabilized, and stained with FITC-conjugated anti-BrdU antibodies (to detect BrdU-positive cells) and 7-aminoactinomycin (to stain total nuclear DNA) 47 according to the instructions of the FITC BrdU Flow Kit. After embedding in ProLong Gold antifade reagent (Thermo Fisher Scientific Inc., USA), the preparations were analysed using a lens Plan-Apochromat 20×/0.8 M27 and lasers: 561, 488 nm.
To assess the proliferative activity of OE cells, we obtained 15-20 Z-stacks from each olfactory rosette. The number of BrdU-incorporating cells in the tissue was calculated using Zen 2010 (Carl Zeiss) and Imaris Bitplane 7.2.3 software packages (Bitplane AG, Switzerland). In order to evaluate the spatial distribution of BrdU-positive cells, the Z-stacks of the images of olfactory folds were divided into separate smaller (1 × 10 5 μm 3 ) fragments. To estimate the volume of tissue in a Z-stack, we estimated the total volume of nuclei in tissue, occupied by the 7-AAD signal. In each volume of Z-stacks tissue, we counted the number of nuclei that included BrdU (based on binding of the FITC-conjugated antibody to BrdU). Based on the data obtained, we calculated the number of nuclei in 1 × 10 6 μm 3 of tissue for each Z-stack. Using a similar algorithm, we assessed the volume occupied by ROS products, the number of functionally active mitochondria, and PCD.

Statistics.
We analysed the data with non-parametric statistics, calculated the median, 25-, and 75-percentiles using the Statistica 10 software package. Intergroup differences were estimated using Mann-Whitney U and Chi square tests.