Therapeutic Potential of Human Nasal Inferior Turbinate-Derived Stem Cells: Microarray Analysis of Multilineage Differentiation Open Access

Mesenchymal stem cells (MSCs) are adult stem cells with the potential to differentiate into various cells lineages [1]. They can be isolated from various tissues of adults [1, 2]. Previously, we have shown that MSCs derived from nasal mucosa (inferior turbinate) (hNTSCs) have a high proliferation potential. In addition, characteristics of hNTSCs do not change regardless of cell passage number [1]. Moreover, hNTSCs show superior potential to differentiate into osteocytes and chondrocytes. These cells have better osteogenic potential regardless of accessorial osteogenic factors, including vitamin D3 and BMP-2, in comparison with MSCs isolated from bone marrow or adipose tissues [3]. These cells can differentiate easily into chondrocytes by in vitro culture without co-culture of MSCs and chondrocytes which can enhance matrix deposition [4]. MSCs are considered as important sources for cellular therapy due to their immunologically privileged properties. Additionally, the isolation of MSCs from tissues normally removed during surgery and storage of them would have significant advantage for patients with various diseases [5]. Although hNTSCs provide a novel opportunity for regenerative medicine, in otorhinolaryngologic parts, core genes during the differentiation process of hNTSCs have not been evaluated sufficiently. Thus, the objective of this study was to determine differentiation phenotypes of hNTSCs under appropriate differentiation-induction media (osteogenic, chondrogenic, and neurogenic differentiation) and perform microarray-based gene expression profiling to evaluate these differentiation phenotypes of cell populations.

hNTSCs were isolated from patients who underwent an inferior turbinate surgery. Isolation procedures were approved by the Institutional Review Board of Seoul St. Mary’s Hospital, the Catholic University of Korea (KC08TISS0341). Written informed consent was obtained from the subjects.

The inferior turbinate tissue obtained after surgery was washed 3 to 5 times with an antibiotic-antimycotic solution (Gibco, Gaithersburg, MD, USA) and 3 times with phosphate buffered saline (PBS). It was then cut into 0.5-mm³ pieces. These pieces were placed into a culture dish. The dish was covered with a sterilized glass cover slide. Alpha-MEM (WELGENE, Gyeongsan, Korea) containing 10% fetal bovine serum (FBS; WELGENE) and 1% antibiotic-antimycotic solution (Gibco) was added. It was changed every 2 days. After 15 days of culture at 37°C in a 5% CO₂ incubator, the glass cover slide was removed and tissuesfloating in the culture media were removed by washing. hNTSCs attached to the bottom of the culture dish were then detached using 1 mL of 0.05% trypsin in 1 mM EDTA. After isolation, hTMSCs were seeded at an initial cell density of 1 × 10⁴ cells/cm² and amplified in a monolayer culture media. hNTSCs were cultured for 3 passages to evaluate their multidifferentiation potential.

To investigate the expression of cell surface maker proteins in hNTSCs at passage 3, cells were detached and labeled with FITC- or PE-coupled anti-human antibodies (CD14, CD19, CD29, CD34, CD73, CD90, and HLA-DR; BD Biosciences, San Jose, CA, USA). Ten thousand labeled cells were measured using a FACS calibur flow cytometer (Becton Dickinson, Franklin lakes, NJ, USA). Results were analyzed with a CellQuest software (Becton Dickinson).

To investigate the osteogenic potential of hNTSCs, cells were cultured in a low-glucose DMEM supplemented with 10% (v/v) FBS, 0.1 μM dexamethasone, 10 mM β-glycerophosphate, 50 μM ascorbate-2-phosphate, 100 U/mL penicillin, and 100 μg/mL streptomycin. Culture media were changed twice every week for 4 weeks. Differentiation to osteogenic lineages was induced according to the procedures described by Hwang et al. [1]. A 3-dimensional culture system was used for chondrogenic differentiation with open-cell polylactic acid (BD Biosciences). After 2 × 10⁶ cells were seeded onto the scaffold, cells were cultured in chondrogenic induction media consisting of DMEM supplemented with 10% FBS, 100 nM dexamethasone, 50 μM ascorbic acid-2 phosphate (Sigma, St. Louis, MO, USA), 50 mg/mL insulin-transferrin-sodium selenite, 1 mM sodium pyruvate (Gibco), 40 μM proline, and 2 mML-glutamine added to chondrogenic supplements TGF-β1 (10 ng/mL; Sigma) and IGF-1 (10 ng/mL; Sigma). Culture media were changed twice every week for 2 weeks. To induce neurogenic differentiation, hNTSCs were seeded onto 4-well chamber slides (1 × 10⁴ cells/well; Nalgene Nunc International, Rochester, NY, USA) and cultivated in neurobasal media. Cultivated cells were resuspended in DMEM/F-12 (Gibco) with 1 mL B-27 supplement, neurogenic supplements (10 ng/mL glial-derived neurotrophin factor [Invitrogen, Carlsbad, CA, USA], 10 ng/mL neurotrophin-3 [Invitrogen], 10 ng/mL brain-derived neurotrophin factor [Invitrogen], 2 mML-glutamine), 0.5% heparin, and 1% penicillin-streptomycin (Gibco) for 2 weeks. Culture medium was refreshed every 2 days.

The differentiated hNTSCs were fixed with 4% paraformaldehyde in PBS for 20 min, and nonspecific binding sites were blocked by a 30-min incubation in PBS containing 0.5% Triton X-100 (Promega Co. Medison, WI, USA) at room temperature, then stained to analyze cytologically with Safranin O and Alizarin red for chondrocyte and osteocyte differentiation. For immunocytofluorescence staining, it is followed by antigen retrieval with proteinase K (Abcam) to expose the antigenic sites. After blocking with 5% normal goat serum, cells were each incubated with primary antibodies against NeuN (Abcam) and β-tubulin (Abcam) at room temperature overnight for double staining and washed triply with 0.01 mol/L PBS. After incubating with secondary antibodies, Alexa 488 (1:100) and Alexa 546 (1:100) at room temperature for 2 h, antigen expression was each identified. For staining nuclei, cells were mounted with mounting medium (Vectashield, Burlingame, CA, USA) conjugated with DAPI and viewed using a fluorescence-attached microscope (OlympusAX70TR62A02, Tokyo, Japan).

Total RNA was extracted from cells cultured under different induction media for chondrogenic, osteogenic, or neurogenic differentiation with an RNA isolation kit (Kontes, Vineland, NJ, USA). Cy3-labeled cDNA was synthesized using 200 ng of total RNA with a Low Input Quick Amp Labeling Kit (Agilent Technologies, Santa Clara, CA, USA). cRNAs were hybridized on a Custom Gene Expression Microarray, 4 × 44k (Agilent Technologies) at 65°C for 17 h. After that, microarrays were processed in accordance with the manufacturer’s recommended protocol. We applied quantile normalization to compensate interarray variations. Genes with little variation and those with entropy values less than the 10th percentile were excluded. The genes were analyzed for significance by selecting 100 genes with the most significant differences by the SAM method. There are genes with the same significance (delta value), so the specific number is slightly different.

Fibroblast-like cells appeared within 3 days, adhered to the culture dish, and formed into a single layer. Flow cytometric analysis was performed for cell surface markers of in vitro-cultured hNTSCs. Results revealed that these cells did not show any expression of hematopoietic cell markers (CD14, CD19, CD34, or HLA-DR). However, they showed expression of MSC markers (CD29, CD73, and CD90) (Fig. 1).

The cells exposed to osteogenic medium showed direct evidences of calcium mineralization seen by Alizarin Red staining. Cells after neurogenic induction for 14 days were analyzed with immunocytochemistry and the expression of nestin and β-tubulin was detected. Safranin O binds specifically detected the proteoglycan in chondrogenic differentiated hNTSCs (Fig. 2).

In osteogenic differentiated hNTSCs, 36 transcripts such as E2F transcription factor 1 (E2F1), activating transcription factor 5 (ATF5), and AKR1B10 were upregulated and 36 transcripts such as CA9, PPFIA4, HAS2, and COL4A4 were downregulated. In chondrogenic differentiated hNTSCs, 3 transcripts (NUDT14, CPA4, and heparin-binding epidermal growth factor-like growth factor [HBEGF]) were upregulated and 82 transcripts such as PTGS1, CLEC2D, and TET1 were downregulated. In neural differentiated hNTSCs, 61 transcripts such as insulin-like growth factor-binding protein (IGFBP)-1, nerve growth factor receptor (NGFR), FGF1, OLFML1, and EPGN were upregulated and 98 transcripts such as ACAN, RUNX2, and C21orf96 were downregulated. These up- and downregulated genes in trilineage differentiated hNTSCs are listed in Tables 1-3 and online suppl. Table 1; see www.karger.com/doi/10.1159/000516016 for all online suppl. material, (fold change in expression). Ad-ditionally, principal component analysis plot for comparing differences of cell groups according to differentiation and a heatmap for comparing genes differentially expressed in undifferentiated and trilineage differentiated hNTSCs were prepared. Results are shown in Figures 3,4.

Major over- and underrepresented gene ontology (GO) classes are shown in Tables 4-6. In neural differentiated hNTSCs, major GO classes of cytokine or chemokine were abundant. In chondrogenic differentiated hNTSCs, catalysis GO terms were overrepresented. In osteogenic differentiated hNTSCs, GO terms related to oxidoreductase were overrepresented.

Results of the present study verified specific gene expression patterns in hNTSCs according to trilineage differentiation. These results were compared to expression profiles of undifferentiated hNTSCs as a reference. Expression profiles of osteo-, chondro-, and neuro-differentiated hNTSCs were found to be different from each other. Based on expression profile analysis, many genes were shown to be up- or downregulated. Furthermore, GO analyses of regulated gene products showed over- or underexpressed GO classes. These results provide a first step for revealing core molecules associated with characteristics of hNTSCs.

Microarrays have become vital tools for detecting the expression of thousands of genes and for identifying differentially expressed genes. They enable us to explain and compare gene expression patterns of cells during or after differentiation. Based on microarray data, we can comprehend mechanisms controlling the expression of individual cells’ characteristics. Some microarray studies have been conducted to check new specific markers and investigate the origin of various diseases including cancers [6-8].

In the otolaryngologic part, partial turbinate resection is one of the most popular and efficient surgery for reducing nasal symptoms. MSCs derived from the inferior tubinate can be obtained in ease and abundance. Due to this feature, hNTSCs could serve as important resources of adult stem cells. However, characteristics and differentiation of hNTSCs using genome-wide research techniques have not been reported yet. Studies on MSC derived from different tissues have shown a phenotype analogous to each other. Nevertheless, considerable differences in the molecular phenotype among MSCs from different tissues have been reported, showing ontological and functional differences [9]. Thus, it would be important to perform genome-wide gene expression profiling and GO analysis for hNTSCs. Thus, this is the first study to obtain gene expression profiles of hNTSCs through microarray.

HBEGF is a mitogenic and chemotactic molecule related to tissue repair and other tissue-modeling phenomena. Krampera et al. [10] have found that HBEGF can increase the proliferation of bone marrow MSCs but prevent chondrogenic differentiation reversibly. However, in our study, hNTSCs showed increasing HBEGF expression during chondrogenic differentiation. Transforming growth factor beta can induce extracellular matrix proteins during chondrogenic differentiation in vitro. In chondrocyte progenitor cells stimulated by transforming growth factor beta, HBEGF was strongly expressed during induction, suggesting that HBEGF could be important in extracellular matrix manufacture as with cell function [11]. Our results showed that hTMSCs after differentiation were most similar to chondrocyte progenitor cells. This could explain why HBEGF is an important role during the chondrogenic differentiation of hNTSCs. ATF5 is a member of the ATF/CREB family of beta-ZIP transcription factors. It is strongly expressed in various neoplasms. It also controls stem cell functions [12]. During osteogenic differentiation of MSCs from human adipose tissue, ATF5 was highly expressed, reaching a peak of expression at the stage of bone mineralization. Vicari et al. [13] have suggested that ATF5 could play an interesting regulatory role during osteogenesis. These results were similar to our results showing that ATF5 was expressed highly during osteogenic differentiation.

E2F1 functions as a DNA-binding domain of pRb. Signaling via the interaction of pRb-E2F is considered to be associated with cell cycle regulation, cell fate, and differentiation of MSCs [14]. pRB is an activator of tissue-specific gene expression along diverse lineages including osteoblast differentiation [15]. In addition, the transcription start site of the ATF5 gene with promoter activity retaining has potential binding sites for several transcription factors, including EBF1, Sp1, and E2F1. In particular, mutation of the E2F1-binding site obviously impairs the activity of the ATF5 promoter, showing that this site is the principal cis-element for the transcriptional activation of human ATF5 gene [12]. These facts could support our results showing that E2F1 was expressed highly during induction for osteogenic differentiation. IGFBP coordinate and regulate biological activities of IGF, a powerful neurotrophic factor that provokes proliferation, migration, and differentiation of glial and neuronal cells and prevents these cells from going through apoptosis [16]. Some authors have proposed that MSCs might have neurotrophic properties to differentiate into neuron and repair nerve. A previous study using a nerve injury model has shown a considerable increment in the release of IGFBP-1 from transplanted MSCs and IGFBP-1 during repair of peripheral nerves has significant influence on engraftment and neural differentiation of transplanted cells. Therefore, it has been suggested that IGFBP-1 is an essential neurotrophic factor released from MSCs that can promote regeneration of damaged neural tissues [17]. NGFRs are a group of growth factor receptors that can specifically bind to neurotrophins. They are also used as cell surface markers to define MSC phenotype [18]. In particular, in several studies using monoclonal antibodies to NGFR for isolation of NGFR positive MSCs, this subset of cells has a higher neuronal differentiation property than the whole MSCs population [19, 20]. These results were similar to our results showing that IGFBP-1 and NGFR were expressed highly during induction for neurogenic differentiation.

GO is a method to comprehend massive-scale gene expression data. In the GO database, top-level ontologies are molecular function, biological process, and cellular component [21]. In our study, according to GO analysis, GO terms related to molecular functions were overrepresented in all trilineage differentiation of hNTSCs. However, GO terms related to cytokine or chemokine were overrepresented in neural differentiated hNTSCs while GO terms related to catalysis and GO terms related to oxidoreductase activity were mainly overrepresented in chondrogenic and osteogenic differentiated hNTSCs, respectively. These facts clearly showed that differently differentiated MSCs had other specific characteristics. In the future, further studies are needed to find the meaning of other major overrepresented GO terms in differentiated hNTSCs.

We performed microarray analysis during trilineage differentiation of hNTSCs in vitro. We found that over 300 genes were expressed differentially in hNTSCs during differentiation. However, certain limitations need to be noted in this study. Of over 300 genes in hNTSCs we found, we could not evaluate characteristics of the majority of genes regulated during in vitro differentiation. Additional studies are needed to determine changes of most genes detected by microarray analysis and characterize these genes more precisely. Additionally, it would be important to identify sequential gene expression during differentiation. Previously, temporal changes of cell growth and osteoblast phenotype-associated genes in human and rodent models have been reported [22, 23]. In future, we will evaluate sequential expression profiling of hNTSCs during differentiation. However, gene expression profiling and GO analysis of this study are useful for defining genes responsible for characteristics of hNTSCs and for analysis of responses of hNTSCs to signals that drive different differentiation processes. Our results may provide novel information of genes involved in the differentiation of hNTSCs.

We described genome-wide gene expression patterns of hNTSCs during trilineage differentiation. Differences in expression profiles of trilineage differentiated hNTSCs were found and 316 candidate genes were shown to be definitely up- or downregulated. Further analytical studies should be conducted to comprehensively determine functional roles of these differentially expressed genes detected by microarray analysis. In the future, gene expression data of hNTSCs could offer important understanding for cellular differentiation processes in MSCs.

This study was approved by the institutional review board committee (IRB No. KC08TISS0341), and informed consent was obtained from all patients.

The authors have no conflicts of interest to declare.

This work was supported by the Korea Health Industry Development Institute funded by the Ministry of Health and Welfare (HI14C3228) and the Bio & Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. NRF-2019M3A9H2032424, No. NRF-2019M3E5D5064110). This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) grant funded by the Ministry of Education (2018R1D1A1B07045421). This work was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education (2017R1D1A1B03027903). This research was supported by the Bio & Medical Technology Development Program of the National Research Foundation (NRF) funded by the Korean government (MSIT) (No. NRF-2018M3A9E8020856). This research was supported by the Basic Science Research Program through a National Research Foundation of Korea (NRF) grant funded by the Ministry of Education (2017R1D1A1B03034868). This work was also supported by the Institute of Clinical Medicine Research of Bucheon St. Mary’s Hospital, Research Fund, 2018, and this study was supported by a grant of the E.N.T. Fund of the Catholic University of Korea made in the program year of 2017 and 2018. The sponsors had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Sun Hwa Park: data analysis, drafting, final approval, and accountability for all aspects of the work; Do Hyun Kim: data analysis, drafting, final approval, and accountability for all aspects of the work; Mi Hyun Lim: data analysis, drafting, final approval, and accountability for all aspects of the work; Sang A Back: data analysis, drafting, final approval, and accountability for all aspects of the work; Byeong Gon Yun: data analysis, drafting, final approval, and accountability for all aspects of the work; Jung Ho Jeun: data analysis, drafting, final approval, and accountability for all aspects of the work; Jung Yeon Lim: data analysis, drafting, final approval, and accountability for all aspects of the work; Su Young Kim: data analysis, drafting, final approval, and accountability for all aspects of the work; Se Hwan Hwang: data analysis, drafting, final approval, and accountability for all aspects of the work; Sung Won Kim: data analysis, drafting, final approval, and accountability for all aspects of the work.

Se Hwan Hwang and Sung Won Kim also contributed equally to this work as corresponding authors.