Differential expression of microRNAs among cell populations in the regenerating adult mouse olfactory epithelium

Despite a robust capacity for adult neurogenesis in the olfactory epithelium (OE), olfactory sensory losses are common. Identification of mechanisms regulating adult OE neurogenesis is, therefore, of interest. MicroRNAs (miRNAs) are broadly important in regulating vertebrate neurodevelopment, and are required in embryonic olfactory differentiation. We report here that a panel of miRNAs is differentially expressed by either progenitor or progeny cells in the regenerating mouse OE. Progenitor cells were purified from lesioned OE based on c-Kit expression, and miRNA expression was assayed in c-Kit (+) and c-Kit (-) cell populations. 28 miRNAs were significantly downregulated by at least 4 fold in the c-Kit (+) fraction, which marks the globose basal progenitor cell population. In addition, 10 miRNAs were upregulated in these basal cells. MiR-486, the most strongly downregulated miRNA identified, was further characterized to verify results. MiR-486 expression was confirmed in the c-Kit (-) OE layers using in situ hybridization. As a functional assay, over-expression of miR-486 in purified c-Kit (+) basal cell cultures resulted in a reduction in neurogenesis, consistent with a possible negative feedback regulatory model. Our data provide new insights regarding miRNA expression and function during adult OE neurogenesis, and identify candidate miRNAs warranting further study.


Introduction
MicroRNAs (miRNAs) are major posttranscriptional regulators of gene expression [1]. The first miRNA was identified in C. elegans in 1993 [2], and the importance of mammalian miR-NAs is now widely recognized in development and disease. By base-pairing with complementary sites in their target messenger RNAs (mRNAs), miRNAs control the repression of mRNAs, primarily through mRNA destabilization [3][4][5]. With each miRNA capable of a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Bar Harbor, ME) were used for tissue harvest for histology, cell purification for RNA preparation, or cell culture. Methimazole (Sigma, St. Louis) was dissolved in phosphate buffered saline (PBS) and injected intraperitoneally at 50 μg/g body weight 10 days prior to sacrifice, to induce olfactory lesion and regeneration [23]. For euthanasia, mice were anesthetized with an over dose of ketamine-xylazine intraperitoneally, prior to decapitation and tissue harvest.

Cell isolation
10 days following methimazole lesion, adult mice were deeply anesthetized, decapitated, and nasal tissue was harvested, enzymatically dissociated into a single cell suspension, and washed. Immunomagnetic selection was performed using the EasySep Mouse APC Positive Selection kit (Stemcell Technologies, Vancouver, BC) to label cells with APC-conjugated antibody against c-Kit (eBiosciences 17-1171; RRID:AB_469430), as previously described [31].

Microarrays and bioinformatics
Total RNA from sorted c-Kit (+) and (-) fractions was extracted using Quick-RNA MicroPrep kit (Zymo Research, Irvine, CA) to yield a 260-280 nm absorbance ratio of 2.0. RNA concentration and integrity was confirmed using an Agilent Bioanalyzer. Briefly, 500ng total RNA was prepared using the Affymetrix FlashTag Biotin HSR RNA Labeling Kit (Part# 901910) according to manufacturer's protocol (3 samples per group; 2 groups per experiment). Fragmented and labeled products were hybridized to the Affymetrix GeneChip miRNA 4.0 Array. Washing and staining of arrays took place on the GeneChip Fluidics Station 450 and scanning was done on the GeneChip Scanner 3000 7G system. Expression ratios were calculated as the power-2 exponential of the log2 differences. The acceptance criteria for gene array expression changes was a minimum 2 fold change in log2 (equivalent to 4 fold) and a one-way Analysis of Variance (ANOVA) t-test with Benjamini-Hochberg multiple testing correction p-value of <0.05. All microarray raw files were submitted to the NCBI Gene Expression Omnibus (GEO) database (GSE104185).

Quantitative RT-PCR
For miRNA quantification, reverse transcription was performed using miRNA-specific Taqman primers, and cDNA was amplified using Taqman primers against the mature miRNA strands. Data was analyzed using the RQ Manager 1.2 from Applied Biosystems, CA. TaqMan assays were run in duplicate for each miRNA in each sample (n = at least 3 biological replicates/condition) and all miRNA levels were normalized to endogenous Sno RNA.

In situ hybridization
Adult mice were deeply anesthetized and perfused with cold PBS followed by 4% paraformaldehyde in PBS. Nasal tissue was isolated, post-fixed, processed in 30% sucrose with 250 mM EDTA in DEPC-PBS, cryosectioned at 10 μm, collected on Superfrost Plus slides (VWR, Radnor, PA), and stored at -80˚C. In situ hybridization was performed as previously described [32]. Briefly, slides were pretreated with post-fix followed by Proteinase K (20 μg/ml) digestion for 10 min, acetylation for 10 min, and PBS rinses. Pre-hyridization was performed for 1 hour at room temperature with hybridization buffer. 3' and 5'-double-digoxigenin labeled LNA probes for miR-486-5p (CTCGGGGCAGCTCAGTACAGGA), or a scrambled negative control (GTGTAACACGTCTATACGCCCA), were obtained from Exiqon (Woburn, MA). Probes were diluted in hybridization buffer to 0.1 μM, denatured, then hybridized overnight at 60˚C. Slides were then washed in 0.2x SSC twice for 30 min at 65˚C, then PBS for 5 min, and then processed for anti-DIG-AP. Slides were blocked in PBS with 5% nonfat milk, 4% bovine serum albumin, 5% normal goat serum, and 0.1% Tween 20 for 30 min, incubated with anti-DIG-AP (#11093274910 Roche, Sigma, St. Louis) at 1:250 for 1 hour at room temperature, washed and visualized with NBT/BCIP reaction. Slides were rinsed in PBS with 1 mM EDTA, pH 5.5, then cover slipped with Aquamount (VWR). Images were digitally acquired on a Zeiss Axiophot microscope with phase contrast optics and assembled using ImageJ software.

Cell culture assays
Globose basal cells were prepared for culture as described previously [31]. The c-Kit (+) cell fraction was isolated by immunomagnetic selection as described above. Cells were cultured on Vitronectin-coated coverslips in complete NeuroCult medium (Stemcell Technologies) overnight. Cells were transfected 24 hours later with either 10 nM miR-486 or scrambled control miRNA from Exiqon (Woburn, MA) using Lipofectamine 2000 transfection reagent (Thermo Fisher Scientific) according to manufacturer's instructions. Cells were then processed for further analyses 48 hours post transfection. EdU was added at 3μM 1 hour prior to fixation and visualized by Click chemistry reaction, per manufacturer's instructions (Invitrogen).

Immunochemistry
For immunocytochemistry or immunohistochemistry, cultures were fixed with 4% paraformaldehyde in PBS for 15 min RT and washed in PBS for 5 minutes, or cryosections were prepared from paraformaldehyde perfused mice [18]. Blocking was performed using PBS with 10% normal goat serum (Vector Labs), 4% bovine serum albumin (BSA; Sigma) and 0.1% Triton X-100 (Sigma) for 30-60 minutes at room temperature. Primary antibodies (listed in Table 1) were diluted in the same blocking buffer and applied for 2 hours at room temperature, followed by three 5 min PBS washes. Appropriate Alex488 or Alexa 594 conjugated secondary antibodies (Jackson ImmunoResearch) were applied for 60 minutes at room temperature, followed by three 5 minute PBS washes. Cells were mounted with Vectashield containing 4,6-diamidino-2-phenylindole (DAPI; Vector). For quantification of stained cultures, cells were analyzed for immunofluorescent label and DAPI nuclear staining using the EVOS FL Auto system (Thermo Fisher Scientific). Cells were further analyzed using an Olympus IX81 epifluorescent microscope. A total of n = 8 coverslips from 3 independent culture experiments were counted. For each experiment, the percentage of TUJ1-positive neurons (cells with round soma and neurite-like processes) from miR-486 or scrambled-transfected cells from each coverslip was normalized to the averaged values for the scramble control. Statistical analyses were performed using a Student's t test with Prism 7 (Graphpad).

Identification of differential miRNA expression in OE cell populations
To identify miRNAs selectively enriched in different OE cell populations, we utilized an unbiased global microarray approach (Fig 1). In previous studies Dicer, an enzyme required for mature miRNA production, was conditionally disrupted in OE progenitors using Foxg1-Cre/ Dicer fl/fl mice and resulted in a failure of olfactory neurogenesis, along with widespread defects and prenatal death [10]. We reasoned that during adult olfactory tissue regeneration, the progenitor cells active in producing new olfactory neurons are similarly subject to regulation by specific miRNAs. Using the mouse methimazole lesion-regeneration model [23], we obtained OE tissue actively regenerating after experimental injury, enriched for neurogenic progenitor  1). We purified OE progenitor cells based on their expression of the receptor tyrosine kinase c-Kit [18,30,33], as previously described [31]. MiRNAs expressed in the purified c-Kit (+) progenitor cell fraction (n = 3 biologic replicates, 3 mice per preparation) and in the c-Kit (-) fraction were then assessed using miRNA microarrays. Importantly, RNA was purified acutely upon isolating the cells from tissue, without an intervening culture step, in an effort to identify miRNAs expressed during OE regeneration in vivo.
Using this approach, we hypothesized that miRNAs active in adult OE neurogenesis would be detectable, and that miRNAs selectively enriched in the c-Kit (+) or the c-Kit (-) populations should represent candidates of particular interest. Although the c-Kit (-) fraction is rich in neurons, it also contains sustentacular cells and a small number of horizontal basal cells, Bowman's gland/duct cells, and possibly some underlying stromal cells (Fig 1). Our results indicated that 28 miRNAs are significantly downregulated by at least 4 fold in c-Kit (+) versus c-Kit (-) mouse OE cells (Fig 2A, Table 2). Additionally, we identified 10 miRNAs upregulated at least 4 fold in the c-Kit (+) basal cell fraction (Fig 2B, Table 3). Two miRNAs were selected for validation by qPCR ( Fig 2C): miR-3141, enriched in basal cells, and miR-486, enriched in the c-Kit (-) fraction. While miR-3141 expression trended towards enrichment in c-Kit (+) cells but was variable among qPCR samples, miR-486 was significantly downregulated in the basal cells, and was therefore selected for further study.

MiR-486 is selectively expressed in olfactory neurons
Our cell preparation and sorting protocol was designed to collect OE cells, rather than stromal or non-epithelial components, by relying on dispase to release the OE from the basal lamina. However, to verify that miRNAs found to be enriched in the c-Kit (-) fraction are indeed expressed in OE populations, we performed RNA in situ hybridization. Using a probe for miR-486, the most strongly downregulated miRNA identified in our arrays, we found that miR-486 expression localized selectively to the olfactory neuron layers of the OE in normal unlesioned adult mouse nasal tissue sections ( Fig 3A). As expected, at high magnification miR-486 expression was noted to be absent from the basal cell layers, where c-Kit (+) cells are localized, and was also found to be absent from the apical sustentacular cells layer (Fig 3B). No signal was detected in the lamina propria. Hybridization using a scrambled negative control probe produced an appropriate absence of signal ( Fig 3C). The position of the c-Kit (+) basal cells was verified immunohistochemically using a previously validated anti-c-Kit antibody [18] ( Fig 3D and 3E). Comparison of the staining patterns from Fig 3A and 3B versus Fig 3D and  3E indicates that in the OE miR-486 is expressed in the neuronal layers apical to the c-Kit (+) basal cells.
The phenotypes of the cell layers comprising the unlesioned adult mouse OE are identifiable using a panel of cell type-specific reagents (Fig 3F-3I). Labeling tissue sections using these markers provides further context for the assessment of the miR-486 labeling pattern. For instance, the fully mature olfactory neurons reside in multiple layers in the middle and upper regions of the OE and are labeled by antibody against Olfactory Marker Protein (OMP) ( Fig  3F) [15,34]. Immature or nascent neurons are labeled by antibody Tuj1 [22] or by expression of the transcription factor OAZ [35] (Fig 3G and 3J). OAZ is a highly specific neuron lineage marker, as it functions to regulate olfactory neuron differentiation via interactions with OE transcription factors and SMADs. The transcription factor SOX2 marks a subset of multipotent GBCs and also localizes strongly to the nuclei of sustentacular cells, situated apically in the OE (Fig 3I) [36,37]. Finally, the reserve horizontal basal cells (HBCs) are identifiable by expression of cytokeratin 5, immediately above the basal lamina (Fig 3K) [38]. OE histology and cell type-specific markers used here are summarized schematically (Fig 3L). The observed expression of miR-486 in the immature and mature olfactory neuron layers is consistent with our microarray and qPCR results, providing validation of these data sets.

Over-expression of miR-486 in olfactory basal cell-derived cultures reduces neurogenesis
In an effort to identify a function for miR-486 during olfactory neurogenesis, we next performed in vitro culture assays. Adult mouse globose basal cells, purified based on c-Kit expression as described above, were placed in short term culture [31]. Immunoselection for the OE basal cell marker was used to minimize seeding cultures with lamina propria contaminants, such as fibroblasts or mesenchymal-like cells. When cultured primarily, late embryonic globose basal cells readily produce neurons over 2-3 days [39]. Here, we used a 3-day culture system seeding adult mouse basal cells harvested from regenerating OE, since miR-486 was identified from adult OE tissue. Cultures were characterized by fixing cells after 24 hours and processing for immunocytochemistry. We used OE cell type-specific markers, as described for staining of tissue sections in Fig 3. As reported previously, adult OE basal cells form adherent undifferentiated-appearing islands when seeded on vitronectin substrate in a neural stem cell medium [31]. We found here that after 24 hours, culture islands contained many SOX2 (+) cells, consistent with expansion of OE basal progenitor cells (Fig 4A and 4B). Probing cultures for expression of the neuron lineage-specific transcription factor OAZ (Fig 4A and 4B), we found that these early basal cell islands contained 18±2% OAZ (+) cells (n = 3 cultures, mean ±SEM). The staining pattern identified here confirms that the primary short-term cultures contain OE basal progenitors, and many of the cells express the early olfactory neuronal marker OAZ. Moreover, when analyzed after an additional 48 hours, abundant Tuj1 (+) process-bearing cells are present (Fig 4C). Therefore, the cultures are suitable for analyzing OE neurogenesis in vitro.
Cultures were used to assay for the effect of over-expression of miR-486 (Fig 4C-4F). Comparing cultures transfected with either miR-486 or a scrambled control (n!3 samples per group), we quantified cell number, neurons, and incorporation of a 1 hour pulse of EdU to assess proliferation. Over-expression was validated by qPCR (Fold Change =~20,000; p<0.0001). No difference in EdU labeling was identified, comparing scramble or miR-486 treated cultures (Fig 4E). Neurons were readily identifiable based on morphology as cells bearing long neurites extending from a round soma, with expression of neuron-specific β-tubulin, recognized by the Tuj1 antibody. To quantify only convincingly neuronal cells, we did not include Tuj1 (+) cells if neurite-like processes were not evident. Of interest, all cultures produced Tuj1 (+) neurons; however, we repeatedly identified approximately a 30% reduction in the percentage of neurons arising in the miR-486-transfected cultures (p = 0.0014), normalizing to control cultures (Fig 4F). Thus, in this assay, miR-486 over-expression reduces neurogenesis. We interpret these results as consistent with a possible feedback regulatory role for miR-486 on olfactory precursor cell differentiation, since no effect on proliferation was identifiable.

Discussion
We present here a global miRNA profile focusing on the process of adult olfactory neurogenesis. Using a mouse olfactory lesion-regeneration model, we identified 10 miRNAs enriched in   the progenitor population and 28 miRNAs enriched in the non-progenitor cell pool. Furthermore, miR-486, the most highly enriched miRNA in the non-progenitor pool, was chosen for additional study. We found that miR-486 was expressed in OE neurons by histologic analysis, and miR-486 inhibited neuron production in a primary culture over-expression assay. While miR-486 has been reported as highly expressed in blood cells [40,41], our results on miR-486 expression in neuronal cells are supported by two other reports that found miR-486 expressed in motor neurons [42] and iPSC-generated dopaminergic neurons [41,43]. Our functional results are in line with a study showing that up-regulation of miR-486 following spinal cord injury promoted neurodegeneration [42]. Regulatory roles for non-coding RNAs, including miRNAs, are an active area of study in many biologic processes. Given the broad importance of miRNAs in post-transcriptional gene regulation during vertebrate neural development [44], it is not surprising that they are involved in regulating adult neurogenesis. Indeed, prior studies have identified miRNA expression in developing OE and found that broadly deleting miRNA production in OE progenitors perturbs OE development [10]. However, the prenatal lethality of that approach precluded further assessment of adult mammalian olfactory tissue maintenance or regeneration. The experiments described here were designed to begin to address whether and how miRNAs may be involved in the maintenance of adult OE. During development, the miRNA-200 family was previously found to be important in OE neuron differentiation [10]. Of interest, key targets known to be regulated by this family include Polycomb complex proteins SUZ12 and BMI1, which have been recently found to be involved in adult OE neurogenic lineages [31]. Given that many acquired human olfactory losses, such as presbyosmia (age-related olfactory decline) or post-viral olfactory disorder (a poorly understood condition that may result in permanent impairment), are associated with histopathologic evidence of neurogenic exhaustion [25,45], a better understanding of the processes controlling adult OE maintenance is of interest.
MiR-486 was identified by our microarrays as highly enriched in the c-Kit (-) non-progenitor cell fraction. Therefore, we chose this miRNA for further validation studies and confirmed these findings using qPCR and in situ hybridization. Using an in vitro functional assay, we found that miR-486 over-expression suppressed the neuronal differentiation of basal progenitor cells. Interestingly, our results are in line with a report showing that up-regulation of miR-486 following spinal cord injury promoted neurodegeneration, by suppressing NeuroD6 [42]. In the mouse OE, NeuroD has been found to function in the immediate neuronal precursor stage GBCs, which directly produce new olfactory neurons [46]. As such, it is possible that miR-486 could function analogously in the OE, via interacting with NeuroD. In addition, we have demonstrated that a short-term primary cell culture assay, using purified adult c-Kit (+) basal cells, is useful for further testing of miRNAs of interest. Among the panel of differentially expressed miRNAs identified here is miR-146a. MiR-146a was also reported as down-regulated in olfactory mucosa mesenchymal stromal cells [47]. Understanding interactions between cells of the nasal lamina propria, or mesenchyme, and overlying OE is an area of active investigation, and consideration of potential roles for miR-146a in such interactions is of interest.
Although beyond the scope of the present study, our list of differentially expressed miRNAs identified in regenerating adult OE provides potential therapeutic candidates warranting further investigation. Moreover, in terms of identifying candidates with translational potential, miRNA-mimics or anti-miRs may be attractive for therapeutic use and are currently in clinical trials [48]. In humans, the OE is accessible for the delivery of localized topical therapies; further study of modulatory effects of candidate miRNAs in culture assays and in vivo mouse models may hold translational potential.