MicroRNA Signature and Cellular Characterization of Undifferentiated and Differentiated House Ear Institute-Organ of Corti 1 (HEI-OC1) Cells

MicroRNAs (miRNAs) regulate gene expressions and control a wide variety of cellular functions. House Ear Institute-Organ of Corti 1 (HEI-OC1) cells are widely used to screen ototoxic drugs and to investigate cellular and genetic alterations in response to various conditions. HEI-OC1 cells are almost exclusively studied under permissive conditions that promote cell replication at the expense of differentiation. Many researchers suggest that permissive culture condition findings are relevant to understanding human hearing disorders. The mature human cochlea however consists of differentiated cells and lacks proliferative capacity. This study therefore aimed to compare the miRNA profiles and cellular characteristics of HEI-OC1 cells cultured under permissive (P-HEI-OC1) and non-permissive (NP-HEI-OC1) conditions. A significant increase in the level of expression of tubulin β1 class VI (Tubb1), e-cadherin (Cdh1), espin (Espn), and SRY (sex determining region Y)-box2 (Sox2) mRNAs was identified in non-permissive cells compared with permissive cells (P < 0.05, Kruskal–Wallis H test, 2-sided). miR-200 family, miR-34b/c, and miR-449a/b functionally related cluster miRNAs, rodent-specific maternally imprinted gene Sfmbt2 intron 10th cluster miRNAs (-466a/ -467a), and miR-17 family were significantly (P < 0.05, Welch’s t-test, 2-tailed) differentially expressed in non-permissive cells when compared with permissive cells. Putative target genes were significantly predominantly enriched in mitogen-activated protein kinase (MAPK), epidermal growth factor family of receptor tyrosine kinases (ErbB), and Ras signaling pathways in non-permissive cells compared with permissive cells. This distinct miRNA signature of differentiated HEI-OC1 cells could help in understanding miRNA-mediated cellular responses in the adult cochlea. Supplementary Information The online version contains supplementary material available at 10.1007/s10162-022-00850-6.


INTRODUCTION
MicroRNAs (miRNAs) are small 18-24-nucleotide noncoding single-stranded RNAs that repress or degrade messenger RNAs (mRNAs) by binding to their complementary sequences at the 3′-untranslated region (3'UTR) (Filipowicz et al. 2008;Obernosterer et al. 2006). miRNAs play critical roles in various kinds of biological processes, such as cellular development, differentiation, metabolism, proliferation, migration, and apoptosis (Pasquinelli 2012;Pelaez and Carthew 2012;Kim et al. 2009 MiRNAs play a fundamental role in the regulation of gene expression in the inner ear and associated structures (Mahmoudian-Sani et al. 2017). They are crucial for inner ear development and are involved in the morphogenesis and neurosensory processes that lead to a functional auditory organ (Rudnicki and Avraham 2012). The coordinated expression of miR-183 family members (miR-183, miR-96, and miR-182) has been demonstrated to be particularly important in the development of the sensory cells of the inner ear of mice and other vertebrates (Weston et al. 2006;Sacheli et al. 2009;Li et al. 2010;Friedman et al. 2009). Recent studies show that two single-base mutations in the seed region of miR-96 result in autosomal dominant, progressive hearing loss in both humans and mice (Solda et al. 2012;Mencia et al. 2009;Lewis et al. 2009). This mutation alters the function of miR-96 and their consequent gene expression profile in the mouse organ of Corti such as oncomodulin (Ocm), prestin (Slc26a5), and growth factor independent 1(Gfi1) which have been known to result in deafness and hair cell degeneration (Lewis et al. 2016). These findings demonstrate the importance of miRNA-mediated gene regulation in the cochlea.
House Ear Institute-Organ of Corti 1 (HEI-OC1) cells are one of the few auditory cell lines widely used for research purposes. These cells were derived from the auditory organ of the transgenic mouse Immorto-mouse™, which harbors a temperature-sensitive mutant of the SV40 large T antigen gene under the control of an interferon-gamma-inducible promoter element (Jat et al. 1991;Kalinec et al. 2003). Incubation of Immortomouse™-derived HEI-OC1 cells at permissive conditions (33 °C, 10 % CO 2 ) induces immortalizing gene expression, resulting in de-differentiation of the cells and accelerated proliferation; transferring these cells to non-permissive conditions (39 °C, 5 % CO 2 ) results in denaturation of the protein encoded by the gene, leading to decreased proliferation, cell differentiation, and cell death (Kalinec et al. 2003;Devarajan et al. 2002). HEI-OC1 cells are used as an in vitro system for screening of ototoxic drugs and to investigate drug-activated apoptotic pathways, autophagy, senescence, cell protection mechanisms, inflammatory responses, cell differentiation, genetic and epigenetic effects of pharmacological drugs, oxidative and endoplasmic reticulum stress, and other conditions (Kalinec et al. 2016a). Wang et al. (2010) employed a cell model of oxidative stress with HEI-OC1 cells incubated under permissive conditions to determine the impact of oxidative stress on relative miRNA and mRNA transcripts in auditory cells. To the best of our knowledge, no studies have reported the miRNA expression in differentiated HEI-OC1 cells.
Here, we performed experiments to compare the miRNA expression profile of HEI-OC1 cells maintained under non-permissive culture conditions with that of HEI-OC1 cells maintained under permissive conditions. The identity of putative and validated target genes of miRNAs found to be differentially expressed under non-permissive conditions was sought using gene functional analysis. Cellular characterization studies were undertaken to document differences in the morphology, protein, and gene expression of HEI-OC1 cells under permissive and non-permissive conditions.

METHODOLOGY
This study was approved by the Biosafety Committee of the University of British Columbia, Vancouver, Canada.

Cell Culture
HEI-OC1 cells (kindly provided by Dr. F. Kalinec), derived from the transgenic mouse postnatal organ of Corti, were used to investigate their miRNA expression profiles during proliferation and differentiation. HEI-OC1 cells were cultured under permissive and non-permissive culture conditions as recommended by Kalinec et al. (2016b) to promote proliferation and differentiation, respectively. All cultures were grown in T 25 flasks (Nunc™ Non-treated) in Dulbecco's Modified Eagle's Medium (DMEM), containing 10 % fetal bovine serum (FBS) without supplements and antibiotics in a humidified incubator. Cell morphology was captured with a phase-contrast Zeiss Axio Vert.A1 inverted microscope.
Permissive cultures (P-HEI-OC1 cells) were incubated at 33 °C and 10 % CO 2 as recommended (Kalinec et al. 2016b). The growth medium was replaced every 2 days. The cells were harvested for experiments once the cultures achieved 100 % confluence usually after 5-7 days of incubation.
Non-permissive cultures were obtained by initially incubating HEI-OC1 cells under permissive conditions until they reach 80-100 % confluence. They were then moved to previously described non-permissive conditions: 39 °C and 5 % CO 2 to promote cell differentiation (Kalinec et al. 2016b). The cells were maintained over 2 incubation periods: 1 week (NP 1 -) and 2 week (NP 2 -). HEI-OC1 cells under non-permissive culture conditions (NP-HEI-OC1 cells) changed cellular morphologies and started dying as previously described (Kalinec et al. 2003;Devarajan et al. 2002). To minimize the effects of toxins released by dead cells, the growth medium was fully replaced daily and the cultures were harvested after 1-week and 2-week incubation periods, respectively, for further study. Cells were washed with Dulbecco's phosphate-buffered saline (DPBS) buffer and trypsinized with 0.25 % trypsin-EDTA and incubated at 37 °C for 5 min. Trypsinization was stopped by adding 9 ml of DMEM medium, and the pooled suspension was centrifuged at 1500 rpm for 10 min to obtain the cell pellets for subsequent RNA extractions. miRNA was extracted from the cell pellets using miRNeasy kit (QIA-GEN) as per manufacturer's protocol. Extracted miRNAs were quantified in a BioTek (EPOCH) microplate spectrophotometer using Gen5 software.
Pre-amplification of the cDNA product after RT was performed using 12.5 μl TaqMan preAmp master mix (2X), 2.5 μl megaplex preAmp primers (Rodent Pools A + B) (10X) and nuclease-free water to a final volume of 25.0 ul in a BioRadT100™ thermal cycler according to the manufacturer's recommended thermal cycling conditions. TaqMan Low-Density Array (TLDA) The miRNA profiling of 768 miRNAs was performed with TLDA cards (Rodent Pools A + B Cards Set v3.0). To prepare the real-time PCR reaction mix, 9 μl of diluted pre-amplification product (1:4), 450 μl of TaqMan™ universal PCR master mix (no AmpErase™ UNG) (2X), and 441 μl of nuclease-free water were added to a final volume of 900 μl. One hundred microliters of the PCR reaction mix was loaded onto each row of the 384-well TLDA cards (A or B), centrifuged for 1-2 min at 1200 rpm, sealed carefully and run in a ViiA™ 7 Real-Time PCR System at recommended settings and cycling conditions. HEI-OC1 cells were grown twice under each set of culture conditions (P-, NP 1 -, and NP 2 -HEI-OC1 cells), and TLDA assays were repeated on cells drawn separately from the duplicated cell cultures. Relative miRNA levels were calculated using the comparative threshold cycle (Ct) method (∆∆Ct) normalized to a global mean value and at a cut off Ct level < 35.0.
The TLDA cards tested for 596 Mus musculus miRNAs, 78 Rattus norvegicus miRNAs, 76 Homo sapiens miRNAs, and 18 controls. All non-mouse species' differentially expressed miRNAs (DEMs) were searched to determine if they shared the same conserved sequences as mouse miRNAs using miRBase database (Release 22.1, http:// mirba se. org). Non-mouse DEMs that were homologous to mouse miRNAs were included and non-homologues were excluded from analysis. In addition, DEMs that are not defined as miRNAs currently by the miRBase database (dead entries) were also excluded.

Prediction of Putative and Validated Target Genes and Their Functional Enrichment Analysis
The putative and validated target genes of DEMs were obtained using miRWalk3.0 database with filters miRDB and miRTarBase, respectively, at a binding probability of 1.0 within the 3-UTR region (Sticht et al. 2018). The DAVID Bioinformatics Resources 6.8 NIAID/NIH functional annotation tool (da Huang da et al. 2009a;da Huang et al. 2009b) was used to determine if the identified target genes were statistically significantly (at a cut off adjusted P value < 0.05) associated with functional terms: Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology Biological Process (GOBP).
RNA (RNeasy ® mini kit, QIAGEN) was extracted from P-, NP 1 -, and NP 2 -HEI-OC1 cell pellets which were dissolved in 350 μl RLT buffer containing 0.01 % 14.3 M β-mercaptoethanol according to the manufacturer's protocol. The quantity and quality of extracted RNA were determined prior to cDNA preparation. cDNA synthesis was performed with SuperScript™ VILO™ cDNA synthesis kit (Invitrogen) as per manufacturer's protocol in a BioRadT100™ thermal cycler. Synthesized cDNAs were then diluted to a concentration of 5 ng/μl.
In brief, the RT-qPCR reaction mix per well consisted of 1 μl of HyPure™ molecular biology grade water, 5 μl SYBR select master mix at the manufacturer's supplied concentration, 1 μl of each forward and reverse primers (10 μM), and 2 μl of diluted cDNA (5 ng/μl). After the reaction mix was added to the wells, the plate was centrifuged for a few seconds in a Mini PCR Plate Spinner. RT-qPCR with an initial denaturing step of 95 °C for 10 min and followed by 40 amplification cycles of 15 s at 95 °C and 1 min at 60 °C duration was undertaken on a Quant Stu-dio™ 3 Real-Time PCR system (Applied Biosystems). All target gene tests were repeated a minimum of three times on each sample. Relative mRNA levels were determined using the comparative cycle threshold method at a cut-off Ct < 40.0. Gapdh was used as a reference gene for data normalization. The relative mRNA level was expressed as the mRNA copies of the gene of interest per 1000 copies of Gapdh mRNA [2 −∆Ct /1000 = 1000/2 ∆Ct = 1000/2^( avg.

mRNA-miRNA Interactions
Since the tested target genes were primarily selected from previous inner ear studies, it is worthwhile to predict their biological target miRNAs to determine the mRNA-miRNA interactions. Therefore, target genes used for the cellular characterization of P-, NP 1 -, and NP 2 -HEI-OC1 cells were then searched for their biological target miRNAs by searching for the presence of conserved sequences (8mer and 7mer) that match the seed region of each miRNA with TargetScanMouse version 7.2 (Agarwal et al. 2015). miRNAs that shared poorly conserved sequences were excluded. mRNA-miRNA interactions are illustrated using Cytoscape version 3.7.1 (Shannon et al. 2003).
The number of immunofluorescence-positive cells per 400 × magnification field was recorded to determine the protein expressions semi quantitatively. Viable (DAPI positive nuclei) cells were relatively sparse and unevenly distributed under non-permissive culture conditions. Thus, areas in each culture with high numbers of DAPIpositive nuclei were selected for counting the number of antibody positively and negatively stained cells. P-HEI-OC1 cultures contained high levels of evenly distributed viable cells making selection of ideal high-power fields for permissive and non-permissive culture conditions. a and d P-HEI-OC1 cells at 5th and 7th (> 80 % confluence) day of incubation, respectively. b and e NP 1 -HE1-OC1 cells at 9th and 14th day of incubation, respectively. c and f NP 2 -HE1-OC1 cells at 16th and 21st day of incubation, respectively. Images were captured with phase-contrast microscopy (scale bars are indicated) study straightforward. Counts from 5 non-overlapping fields were recorded in all culture conditions and the average counts determined.
For the gene expression, normalized mean Ct values and for protein expression, the proportions of antibody positively stained cells were compared across P-, NP 1 -, and NP 2 -HEI-OC1 cell cultures using non-parametric Kruskal-Wallis H test (2-sided), followed by Dunn's post hoc test at a Bonferroni-adjusted significance level of P < 0.05 for multiple tests.
Welch's t-test for miRNA expression and non-parametric Kruskal-Wallis H test for gene and protein expression were applied as the standard deviations were different for some tested miRNAs, genes, and proteins in permissive and non-permissive cultures. SPSS version 25.0 (IBM Corp., Armonk, New York) and GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA) were used for statistical analysis and to generate graphs.

Cell Culture Morphology
Culturing HEI-OC1 cells under permissive conditions facilitated proliferation, while non-permissive conditions promoted differentiation. Permissive condition cultures demonstrated small cell size and increased cell numbers ( Fig. 1a and d), in keeping with a high proliferative phase. The morphology changed from spindle-shaped to cobblestone-shaped cells when the cells were transitioned from permissive to non-permissive conditions as illustrated ( Fig. 1d and b, respectively). The number of cells decreased, individual cell size increased, and nuclear clumping and debris accumulation increased consistent with more cell death at 2 weeks' incubation under nonpermissive conditions (Fig. 1e, c, and f).
Excluding transcription and regulation of transcription, the GOBP term nervous system development was predominantly enriched in NP 1 -HEI-OC1 cells when compared with P-HEI-OC1 cells (P < 0.05) (Fig. 5b). None of the KEGG pathways and GOBP terms was significantly enriched for validated target gens of DEMs in NP 2 -HEI-OC1 cells when compared with NP 1 -HEI-OC1 cells (P > 0.05).

Fluorescence Immunocytochemistry (ICC) on HEI-OC1 Cells
Six protein markers myosin 7a, prestin, Sox2, nestin, e-cadherin, and vimentin were used to characterize HEI-OC1 cells maintained under permissive and non-permissive conditions, and the proportion of antibody-positive cells were analyzed using Kruskal-Wallis H test followed by Bonferroni-corrected Dunn's post hoc test (Table 6) and presented for all three cultures (Figs. 10-12).
NP 1 -HEI-OC1 cells contained both Sox2-positive and Sox2-negative cells. Sox2-negative cells could be considered as differentiated cells like Sox2-negative cells identified in NP 2 -HEI-OC1 cells. Between P-and NP 1 -HEI-OC1 cells, Sox2 was overexpressed or the signals were strong in NP 1 -HEI-OC1 cells, indicating its importance during hair cell differentiation. Nestin expression was strong in the nuclei of both P-and NP 1 -HEI-OC1 cells ( Fig. 11d and e, respectively), whereas nestin expression was predominantly detected in the cytoplasm of differentiated NP 2 -HEI-OC1 cells (Fig. 11f). Sox2 expression was significantly decreased in NP 2 -compared with NP 1 -HEI-OC1 cell cultures (P = 0.001, Bonferroni-corrected Dunn's test), whereas nestin expression was comparable across P-, NP 1 -, and NP 2 -HEI-OC1 cell cultures (P = 0.96, Kruskal-Wallis H test) ( Table 6). The proportion of Sox2and nestin-positive cells with respect to total number of DAPI-stained nuclei is presented (Fig. 11g and h).
Vimentin, a mesenchymal cell marker, was comparably expressed in all three HEI-OC1 cell cultures (Fig. 12a-c, respectively), and its semi quantification was not significantly different (P = 0.077, Kruskal-Wallis H test) among these cells (Fig. 12g). E-cadherin is an epithelial cell marker, involved in cell-cell adhesion. E-cadherin expression was comparatively weak in P-HEI-OC1 cell cultures (Fig. 12d). Due to increasing cell death and vulnerability to staining steps, remnants of e-cadherin protein were identified in non-permissive cells ( Fig. 12e and f, respectively), and therefore, protein semi quantification was not carried out.

DISCUSSION
Inner ear tissue differentiation and maintenance are regulated and controlled by conserved sets of cell-specific miRNAs (Friedman et al. 2009). Here, we demonstrated differences in the miRNA signature of undifferentiated and differentiated HEI-OC1 cells.
Putative target genes of the DEMs in non-permissive HEI-OC1 cells revealed that MAPK, ErbB, and Ras signaling pathways were the predominantly significantly enriched KEGG pathways in differentiated HEI-OC1 cells ( Fig. 3a and b). ErbBs are widely expressed in varying degrees in overlapping populations of sensory and non-sensory cells within the neonatal and adult inner ear (Hume et al. 2003). Hume et al. (2003) suggest that the expression of the ErbBs in supporting cells, hair cells, and non-sensory cells are potentially involved in the regulation of multiple processes including survival, synaptic maintenance, and cochlear homeostasis, in addition to a role in proliferation. In this current study, miRNA signature of non-permissive HEI-OC1 cells confirmed the functional enrichment of ErbB signaling pathway when compared with permissive HEI-OC1 cells. In addition to ErbB, Ras and MAPK signaling pathways were also enriched in non-permissive cells. Ras/MAPK pathway is essential in the regulation of cell cycle, differentiation, growth, and cell senescence, all of which are critical to normal growth and development (Tidyman andRauen 2009). Haque et al. (2016), for the first time, showed that mitogenactivated protein 3 kinase 4 (MEKK4) signaling is highly regulated during inner ear development and is critical to normal cytoarchitecture and function as deficient mice exhibit a significant reduction of hair cells and hearing loss. Meanwhile, FoxO signaling was predominantly enriched in NP 2 -HEI-OC1 cells when compared with NP 1 -HEI-OC1 cells (Fig. 3c), suggesting the activation of cellular physiological events such as apoptosis with increasing incubation period.
DEMs miR-200-3p and miR-34c-3p were significantly and consistently upregulated in non-permissive cultures. The miR-200 family has been shown to inhibit epithelial to mesenchymal transition, by maintaining the epithelial phenotype through direct targeting of transcriptional repressors of e-cadherin (Cdh1), Zeb1, and Zeb2 (Korpal and Kang 2008). Cdh1 expression was significantly elevated in both NP 1 -and NP 2 -HEI-OC1 cells compared with P-HEI-OC1 cells (P ≤ 0.001 and P = 0.01 respectively, Bonferroni-corrected Dunn's test) Fig. 7 A Box and Whisker plot to illustrate the normalized Ct values (< 40) obtained for target genes Atoh1, Tubb1, Tubb3, Tubb5, Cdh1, Espn, Myo7a, Nes and p27 Kip1 in P-, NP 1 -, NP 2 -HEI-OC1 cells. Box and Whisker plot was generated with GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA). Whiskers were made using Tukey's method (whiskers extend to 1.5 × IQR). Ct values normalized to endogenous control Gapdh were compared with Welch's t-test (2-tailed) and the FDR was corrected by Benjamini-Hochberg procedure. Normalized mean Ct values were compared using non-parametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significantly different inter-group differences in normalized mean Ct values expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*, < 0.01**, and < 0.001*** (Table 5), suggesting that upregulation of miR-200c-3p protects or maintains the epithelial characteristics in differentiated HEI-OC1 cells. Furthermore, Vim expression was significantly reduced in NP 2 -compared with P-and NP 1 -HEI-OC1 cells (P = 0.007 and P < 0.001, respectively, Bonferroni-corrected Dunn's test) which is consistent with the Cdh1 expression change and supports epithelialization in non-permissive HEI-OC1 cells. E-cadherin antibody staining in non-permissive cells also demonstrated ( Fig. 12e and f). Cell culture morphology displayed the mesenchymal to epithelial transition when HEI-OC1 cells were transitioned from permissive to non-permissive conditions (Fig. 1d to b). Increased cell death in non-permissive cells and vulnerability to multiple washing steps likely contributed to the failure to detect e-cadherin protein. The expression level of Zeb1 was significantly increased in NP 1 -compared with P-HEI-OC1 cells (P ≤ 0.001, Bonferroni-corrected Dunn's test), whereas Zeb2 expression was comparable across P-, NP 1 -, and NP 2 -HEI-OC1 cells (Table 5). Therefore, it is not possible to conclude that miR-200c-3p, exerts its effect on mesenchymal-to-epithelial transition in differentiated HEI-OC1 cells through transcriptional regulation of Zeb1 or Zeb2 (Fig. 9).
Sox2 is also targeted by miR-200c-3p (Fig. 9). Sox2, an important transcription factor, plays multiple roles, most prominently in cellular reprogramming and stem cell pluripotency. In addition, Sox2 is considered as a marker of the prosensory domain in the developing cochlea from which the cochlear and vestibular epithelia develop (Kiernan et al. 2005;Hume et al. 2007). Kempfle et al. (2016) reported that Sox2 is required in the cochlea to both expand progenitor cells and initiate their differentiation into hair cells. This is supported by our study where fluorescence signals obtained for Sox2-positive cells were very strong in NP 1 -compared with P-HEI-OC1 cells ( Fig. 11a and b). Furthermore, Sox2 expression was significantly elevated in both NP 1 -and NP 2 -HEI-OC1 cells (P = 0.002 and P = 0.001, respectively, Bonferroni-corrected Dunn's test) compared with P-HEI-OC1 cells (Table 5), whilst Sox2 protein expression was significantly (P = 0.001, Bonferroni-corrected Dunn's test) decreased in NP 2 -compared with NP 1 -HEI-OC1 cells (Table 6). Low levels of Sox2 protein expression despite high levels of Sox2 gene expression in the presence of miR-200c-3p upregulation in NP 2 -HEI-OC1 cells is consistent with miR-200c-3p's inhibition of Sox2 protein synthesis by translational repression. It is likely that the effect of miR-200c-3p upregulation is sub-maximal in NP 1 -HEI-OC1 cell  Pou4f3,Slc26a5,Sfmbt2,Sox2,Krt18,Vim,Zeb1, and Zeb2 in P-, NP 1 -, NP 2 -HEI-OC1 cells. Box and Whisker plot was generated with GraphPad Prism 8 (GraphPad Software Inc., San Diego, CA). Whiskers were made using Tukey's method (whiskers extend to 1.5 × IQR). Ct values normalized to endogenous control Gapdh were compared with Welch's t-test (2 tailed) and the FDR was corrected by Benjamini-Hochberg procedure. Normalized mean Ct values were compared using non-parametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significantly different inter-group differences in normalized mean Ct values expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*, < 0.01**, and < 0.001*** cultures which have not fully transitioned to a predominant differentiated cell culture as reflected by the presence of a mix of Sox2 antibody-positive and Sox2 antibody-negative cells.

Normalized mean Ct (± SD)
Several of the miRNAs implicated in mouse 3′UTR evolution derive from a single rapidly expanded rodentspecific miRNA cluster located in the intron of Sfmbt2, a maternally imprinted polycomb gene. These miRNAs are expressed in both embryonic stem cells and the placenta (Zheng et al. 2011). miR-297 s, miR-466 s, miR-467 s, and miR-669 s fall into the Sfmbt2 miRNA cluster in the 10 th intron of chromosome 2, based on sequence similarity (Zheng et al. 2011). miR-467a an abundant member of the Sfmbt2 cluster promotes cell proliferation, and the remaining members of this cluster are enriched in pathways that regulate cellular growth (Zheng et al. 2011). In our study, Sfmbt2 expression was found comparable between P-and NP 1 -HEI-OC1 cells, whereas it was not determined in NP 2 -HEI-OC1 cells (Fig. 8). Significantly downregulated miR-466a-3p in NP 1 -HEI-OC1 cells and miR-467a-5p in NP 2 -HEI-OC1 cells suggest the changes in cellular growth and lack of proliferation, respectively. As like Atoh1, Sfmbt2 showed conserved sequences that match the seed regions of miR-34 and miR-449 family miRNAs. Therefore, we propose that the coordinated regulation of Atoh1, Sfmbt2, and miR-34/-449 family miRNAs could play a vital role in HEI-OC1 cell proliferation and differentiation. In addition, miR-17 family miRNAs -17-5p and -20a-3p and its paralogous -106a-5p were significantly downregulated in NP 2 -HEI-OC1 cells (Tables 3 and 4). Downregulation of these miRNAs has been reported to be associated with ageing and senescence (Hackl et al. 2010).
In addition, Myo7a which is expressed in the apical stereocilia as well as the cytoplasm of the inner and outer hair cells (Hasson et al. 1995) and Pouf43 which has a central function in the development of all hair cells in human and mouse inner ear sensory epithelia (Hertzano et al. 2004) were significantly elevated in NP 1 -compared to P-HEI-OC1 cells (P < 0.001 and P = 0.001, respectively, Bonferroni-corrected Dunn's test) (Table 5). Pax2 which is one of the earliest genes in preotic tissue contributing to the inner ear development and governs the differentiation of precursor cells into various cell types (Christophorou et al. 2010) and p27 Kip1 which provides a link between developmental control of cell proliferation and the morphological development of the inner ear (Chen and Segil 1999) were significantly reduced in NP 2 -HEI-OC1 cells (P < 0.001 and P = 0.002, respectively, Bonferroni-corrected Dunn's test), suggesting the achievement of maturation when compared to NP 1 -HEI-OC1 cells (Table 5). Several miRNAs identified as potentially targeting Myo7a, Pou4f3, and Pax2 (Fig. 9) were not sought in this study. Nes expression was significantly increased in NP 1 -HEI-OC1 cells (P < 0.001, Bonferroni-corrected Dunn's test) compared with NP 2 -HEI-OC1 cells (Table 5) and its protein localized within the nucleus (Fig. 11e), suggesting the persistence of progenitor cell characteristics in this culture. Prestin (Slc26a5) and cytokeratin 18 (Krt18) expression were low in all three HEI-OC1 cell cultures (Fig. 6), in contrast to adult porcine derived inner ear cells where Krt18 and Slc26a5 gene expressions are highly expressed and positively correlated (Wijesinghe et al. 2021a, b). However, significantly elevated Krt18 in NP 2compared with P-HEI-OC1 cells (P = 0.007, Bonferronicorrected Dunn's test) supported the presence of adult inner ear cell characteristics in fully differentiated HEI-OC1 cells. Likewise, prestin plasma membrane localization which reflects a differentiated outer hair cell characteristic (Park et al. 2016) was strong in non-permissive Target protein expression in P-, NP 1 -and NP 2 -HEI-OC1 cells Proportions of antibody-stained cells (positive) were compared using non-parametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significantly different inter-group differences in protein expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*; < 0.01**; and < 0.001*** SD standard deviation, actual P value (P a ), not applicable (n/a) HEI-OC1 cells compared to permissive HEI-OC1 cells (Fig. 10d-f) reflecting a significantly higher level of prestin protein in both NP 1 -and NP 2 -HEI-OC1 cells (P = 0.009 and P = 0.007, respectively, Bonferroni-corrected Dunn's test). It is notable that miR-196a-5p and miR-322-5p that target Slc26a5 are differentially downregulated in NP 2 -HEI-OC1 cells (Table 3, Fig. 9) which is consistent with a reduction in miRNA inhibition of prestin gene function and can partly explain the relative abundance of prestin in differentiated HEI-OC1 cells. Downregulation of these miRNAs may not be measurably significant in NP 1 -HEI-OC1 cells which have not fully transitioned. Kalinec et al. (2016b) consider 2 weeks under non-permissive conditions the minimum time to achieve cultures of predominantly differentiated HEI-OC1 cells and that under that time there will still be a high level of undifferentiated cells. It is also important to note that miR-196a/b targets Slc26a5 and p27 Kip1 (Fig. 9) and their interactions in HEI-OC1 cells require further investigation. There are some limitations to this study. The number of viable cells reduced considerably under nonpermissive conditions, necessitating the adoption of in permissive and non-permissive HEI-OC1 cells. a Myosin 7a-positive apical projections, diffusely packed in P-HEI-OC1 cells. b and c Myosin 7a-positive apical projections, densely packed in NP 1and NP 2 -HEI-OC1 cells, respectively. d Relatively weak plasma membrane localization of prestin in P-HEI-OC1 cell. e Stable and strong prestin plasma membrane localization in NP 1 -HEI-OC1 cells. f Slightly unstable prestin expression in NP 2 -HEI-OC1 cells. g and h Proportion of myosin 7a-and prestin-positive cells, respectively in P-, NP 1 -, and NP 2 -HEI-OC1 cells (error bars indicate standard deviation). DAPI was used to stain the nuclei. Phase contrast microscopic images are presented with scale bar. Proportions of antibody-stained cells (positive) were compared using nonparametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significant inter-group differences in the proportion of positive cells expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*, < 0.01**, and < 0.001*** a semi-quantitative immunofluorescence approach to determine the protein expression levels under the different culture conditions. In addition, the treated surfaces of the chamber slides (Lab Tek, Permanox TC Surface) used for fluorescence staining could have induced hair cell differentiation, resulting in inconsistencies between the gene and protein expression findings (Wijesinghe et al. 2021a, b;Liu et al. 2016). COVID-19 pandemic restrictions on lab access and reagents/laboratory supplies prevented PCR prestin optimization with multiple primer sets. In future work, we aim to explore the impact of the DEM changes on protein expressions under permissive and non-permissive culture conditions. Despite these limitations, the distinct miRNA signature of differentiated HEI-OC1 cells could help in understanding miRNA-mediated cellular responses in the adult cochlea. Our findings suggest the potential mRNA-miRNA interactions that could be used in future inner ear hair cell regeneration and therapeutic studies.

Fig. 11
Immunofluorescence staining of stem/progenitor cell markers in permissive and non-permissive HEI-OC1 cells. a and b Sox2-positive nuclei identified in P-and NP 1 -HEI-OC1 cells, respectively. c Low/absence of Sox2-positive nuclei in NP 2 -HEI-OC1 cells. d and e Nestin-positive nuclei identified in P-and NP 1 -HEI-OC1 cells, respectively. f Nestin localized in the cytoplasm of differentiated NP 2 -HEI-OC1 cells. g and h Proportion of Sox2-and nestin-positive cells, respectively in P-, NP 1 -, and NP 2 -HEI-OC1 cells (error bars indicate standard deviation). DAPI was used to stain the nuclei. Phase contrast microscopic images are presented with scale bar. Proportions of antibody-stained cells (positive) were compared using non-parametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significant inter-group differences in the proportion of positive cells expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*, < 0.01**, and < 0.001*** Fig. 12 Immunofluorescence staining of EMT markers in permissive and non-permissive HEI-OC1 cells. a, b, and c Vimentin, a mesenchymal cell marker identified in P-, NP 1 -, and NP 2 -HEI-OC1 cells, respectively. d, e, and f Weak signals obtained for cell-cell adhesion marker e-cadherin in P-, NP 2 -, and NP 1 -HEI-OC1 cells, respectively. g Proportion of vimentin-positive cells in P-, NP 1 -, and NP 2 -HEI-OC1 cells, respectively (error bars indicate standard deviation). DAPI was used to stain the nuclei. Phase contrast microscopic images are presented with scale bar. Proportions of antibody-stained cells (positive) were compared using non-parametric Kruskal-Wallis H test (2-sided) followed by Dunn's post hoc test. Significant inter-group differences in the proportion of positive cells expressed at a Bonferroni adjusted for multiple comparisons significance level of P < 0.05 indicated as < 0.05*, < 0.01**, and < 0.001***