Interferon-beta inhibits human glioma stem cells rather than neural stem cells via cell cycle and immune signaling

Background: Malignant Glioma is considered to be a highly heterogeneous brain tumor, which may have originated from mutated neural stem cells (NSCs) many years before it was diagnosed. Malignant Glioma is characterized by strong self-renewal potential and immature differentiation potential. The main reason is that malignant glioma holds a key cluster cells, glioma stem cells (GSCs). GSCs contribute to tumorigenesis, tumor progression, recurrence, and treatment resistance. Interferon-beta (IFN-β) is well known for its anti-proliferative efficacy in diversity cancers. IFN-β also displayed potent antitumor results in malignant glioma. IFN-β treatments administered in response to gliomas affect both GSCs and NSCs. However, the function comparison, similarities and differences of IFN-β on GSCs and NSCs are rarely reported. Methods: We used human GSCs and NSCs to detect the response difference between two cells by IFN-β. Human GSCs were constructed by ourselves and human normal NSCs line were differentiated from human embryonic stem cells. Various concentrations of IFN-β were separately used to treat human GSCs (hGSCs) and human NSCs (hNSCs). Cell morphology, growth, expression of special stemness genes, neural-related makers or proliferation related genes were staining and observed, such as SRY-box transcription factor 2 (Sox2), S100 calcium binding protein B (S100-beta) and marker of proliferation Ki-67 (Ki67). Genes alterations were carefully analyzed by RNA-seq. We used


Introduction
Malignant glioma is highly aggressive and represents the most common primary brain tumors. The overall survival period is typically less than 15 months. Even with regular therapy, including surgical resection and chemoradiation, tumor recurrence appears to be inevitable [1]. Approximately 70% of malignant glioma patients experience disease progression within one year of diagnosis [2]. Thus, the use of personalized drugs that target molecular receptors and immunotherapy have been viewed as promising new options for glioma treatment [3].
Malignant glioma is heterogeneous, with multiple epigenetic and genetic variations identified in associated tumor cells. The intrinsic, aggressive behavior of malignant glioma has also been shown to depend on the complex tumor microenvironment (TME). Malignant glioma and their TME consist of GSCs, mature neural cells (oligodendrocytes, astrocytes, microglia, and ependymal cells), and some immune cells etc [4,5].
The identification of cancer stem cells in malignant glioma are first reported in 2002 [6]. Several groups isolate and characterize stem-like cancer cells in glioma [6][7][8], which lead to the realization that GSCs are resistant to chemotherapy and radiotherapy [9,10]. Some groups try to find the genes which regulate GSCs maintenance and glioma progression [11][12][13][14][15][16][17]. Inspite of so many struggles, the prognosis of GBM has not enhanced in the past decade [17]. At the same time, the underlying mechanisms of GSCs survival after treatment remain unclear.
Interferon (IFN) factors are pleiotropic cytokines that can be categorized into 3 classes. Type I IFNs include approximately 20 members. Human IFNs induce the Janus kinase-signal transducer (JAK) and activator of transcription (STAT) cascade by binding to the IFN-α/β receptors (IFNARs), IFNAR1 and IFNAR2 [18,19]. IFN-β signaling is needed for its antiproliferative effects on a variety of cancer cells [20][21][22][23][24]. During malignant glioma treatment, at least two aspects of NSCs are related to GSCs. On the one hand, mutated NSCs are considered to be the initiation cells of glioma [17,25,26]. On the other hand, normal NSCs are considered to have the ability to move towards GSCs, and can be used as carriers for the treatment of gliomas [27,28]. The human F3 NSCs cell line has been used in multiple studies to perform NSC-based gene therapy, delivering both IFN-β and cytosine deaminase (CD)/5-fluorocytosine (5-FC) prodrugs to glioma cells [29][30][31]. Although IFN-β has been generally employed as a clinical treatment, whole-transcriptome analyses examining the effects of IFN-β stimulation in GSCs and NSCs are still rare. A systematic understanding of the genetic variations that occur following IFN-β treatments can provide additional evidence for the optimization of IFN-β-associated gene therapy for malignant glioma treatment in clinical trials.
In the present study, we used various concentrations of IFN-β to separately treat human GSCs (hGSCs) and human NSCs (hNSCs). Both morphological and genetic alterations were carefully observed and analyzed. We found that IFN-β decreased the cell and nuclear sizes of hGSCs. The number of sphere-like cells observed in hGSC populations was reduced, both during short-term treatments and under conditions of continuous stimulation. However, INF-β did not appear to have the same or similar effects on hNSCs. Transcriptome analysis was performed to identify genes with changes in expression levels in hGSCs but stable expression levels in hNSCs. Immune response and cell adhesion genes were upregulated by IFN-β treatment, whereas the expression levels of cell cycle and ribosome genes were strongly reduced, which was consistent with our observations of the cellular morphology. We observed that IFN-β preferentially restrained hGSCs rather than hNSCs. A few cyclin genes CCNA2, CCNB1/2 and CCND1, reduced in hGSCs after IFN-β treatment. Our exploration may facilitate to design new and more effective pharmacological strategies for killing hGSCs while protecting hNSCs during glioma treatment.

Materials and methods
Cell culture hGSCs line was established from surgical specimens. Surgical specimens were gotten strictly according to Ethics Committee permission. Patients with glioma provided informed consent and donated surgical specimens sample to the study. All the tumor samples were taken to the laboratory in time for continuation dealing. Briefly, after surgery, surgical specimens were collected for primary culture. The specimens were washed with 1× Hank's Balanced Salt Solution (HBSS, Gibco) at least six times. Then, the specimens were cut into small pieces, and the tissue fragments were placed into centrifuge tubes containing 1 U/mL Dispase II (Roche) in 3 mL Dulbecco's Modified Eagle Medium (DMEM)/F12 (Gibco) and kept in water-bath at 37°C for 30 minutes to allow digestion. After digestion, the specimens were centrifuged at 1,000 × g for 3 minutes. The supernatant was discarded, and the tissues were resuspended with 3 mL DMEM/F12, followed by centrifugation. Finally, the precipitate was resuspended in growth medium (DMEM/F12 supplemented with N-2, B-27, GlutaMAX, bFGF, EGF, heparin and penicillin-streptomycin) for daily culture. All the work concentration of FGF, EGF and heparin were 20 ng/Ml, N-2, GlutaMax and penicillin-streptomycin were 100X while B-27 were 50X in the culture medium.
hNSCs line was generated from human embryonic stem cells (hESCs) [32]. Brief description, StemPro Accutase (Thermo Fisher) was used to digest the hESCs for 20 min at 37°C. The cells were plated onto gelatin-coated plates for 1 hour at 37°C. Because hESCs remain suspended, whereas mouse embryonic fibroblast (MEF) cells are adherent and separate hESCs from MEFs. The nonadherent hESCs were washed and seeded on Matrigel-precoated dishes in MEF-conditioned medium. Then we changed the medium to remove the ROCK inhibitor after 24 hours. Single adherent hESCs were expanded in cell medium until they were almost confluent. Noggin (500 ng/mL, R&D) and transforming growth factor-beta (TGF-β) inhibitor (10 mmol/L, Tocris) were added to confluent cells.

Plate coating
Plates were coated with gelatin (Sigma) or Matrigel (BD) for NSCs induction. Culture plates were precoated with poly-L-ornithine (Sigma) and laminin (Thermo Fisher) for human NSCs cultures. The 24-well plates and 6-well plates were freshly coated with gelatin or Matrigel and incubated overnight at 4°C for better packing effects. The next day, the dishes were treated with 0.5 µg/mL poly-Lornithine (dissolved in water) at room temperature for 16 hours. Then, we washed the dishes with 1× phosphate-buffered saline (PBS). Finally, 5 µg/mL laminin was added to the dishes for at least 16 hours. The coated dishes were stored at −20°C, and the supernatant was discarded before use.

Cell immunofluorescent staining
hGSCs were assessed using a staining assay, similar to that described in our previous study [32]. In brief, hGSCs were grown for 3 or 7 days in optimized culture conditions and fixed in 4% paraformaldehyde (PFA) for 12 minutes. Then, the cells were permeabilized using 2.5% Triton X-100 in PBS and incubated for 15 minutes. We discarded the supernatant and blocked all non-specific reactions with 5% bovine serum albumin (Solarbio) in 1× PBS for 1.5 hours. All procedures were performed at room temperature. hGSCs were incubated in primary antibodies against Sox2 (Goat, R&D), S100-β (Mouse, Abcam), or Ki67 (Rabbit, Thermo Fisher) for 2 days at 4°C. After three washes with 0.1% Tween-20 (Sigma) in 1× PBS, the cells were exposed to secondary antibodies (Jackson Immuno Research), including Alexa Fluor 633-conjugated donkey anti-goat IgG antibody, Alexa Fluor 488-conjugated donkey anti-mouse IgG antibody, Alexa Fluor Cy3-conjugated donkey anti-goat IgG antibody, and Alexa Fluor 633-conjugated donkey anti-rabbit antibody. The antibodies were dissolved in PBS containing 2.5% bovine serum albumin and incubated with the cells at room temperature for 2 hours. Finally, the cells were washed 3 times with 0.1% Tween-20 in PBS, and 4', 6-diamidino-2-phenylindole (DAPI, Sigma) was used to stain the nuclei. An inverted fluorescence microscope (Nikon TE2000) was used to obtain images of the immunofluorescent-stained cells.

Sequencing and transcriptome analysis
We collected both hGSCs and hNSCs after 7 days of culture. After discarding the supernatant medium, 1× PBS was used to wash the cells once, and 2 ml Trizol (Invitrogen, Carlsbad, CA, USA) was added to each plate for RNA extraction. The RNA integrity was determined using an Agilent 2100 Bioanalyzer (Agilent, Palo Alto, CA, USA). The quantity was determined on a Nanodrop (Thermo Fisher Scientific, Wilmington, DE, USA). Poly-A containing mRNA molecules were purified by using ploy-T oligo-attached magnetic beads, followed by fragmentation and priming for cDNA synthesis using an Illumina TruSeqTM RNA sample preparation kit (Illumina Inc., San Diego, CA, USA), according to the manufacturer protocol. The cDNA was further converted into double-stranded DNA using the reagents supplied in the kit. The dsDNA was purified with AMpure XP beads and was then end-repaired and A-tailed, according to Illumina's protocol. After adapter ligation, PCR was applied to enrich the DNA fragments with adapter molecules on both ends and to amplify the amount of DNA in the library. The resulting molecular libraries were pooled together and sequenced on a HiSeq 2500 sequencer (Illumina Inc.). Then, the fragments per kilobase of transcript per million mapped reads (FPKM) values were analyzed as the gene expression base. Differentially expressed gene (DEG) analysis was performed using online software (Morpheus, https://software.broadinstitute.org/morpheus) to identify up-and downregulated genes. Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses were also performed using an online database (g:Profiler) [33].

Statistical Analysis
All data were collected and analyzed based on three or more replicates. The error bars represent the standard deviation of the mean. Statistical analysis was performed using GraphPad Prism version 8.0.0 for Windows, GraphPad Software, San Diego, California USA, www.graphpad.comGraphPad Prism 7.0. For multiple comparisons, Student's t-test was used to determine significant differences. Significance is indicated as *p < 0.05. **p < 0.01, and ***p < 0.001.

IFN-β inhibits the growth of hGSCs
To detect the functional role of IFN-β in hGSCs, we firstly used IFN-β, at concentrations of 0, 0.625, 1.25, 2.5, 5, and 11 ng/mL, to treat the hGSCs (Fig. 1a). hGSCs typically display two types of cell morphologies: sphere-like hGSCs and adherent hGSCs (Fig. 1b). We quantified the sizes of the sphere-like hGSC and the coverage of adherent hGSCs to examine the effects of IFN-β on cell growth. The sizes of the sphere-like cells significantly decreased, even when using the lowest treatment dose of 0.625 ng/mL IFN-β (Fig. 1c). However, no changes in the coverage of adherent cells were observed until the treatment concentration reached 5 ng/mL IFN-β (Fig. 1d). These observations suggested that adherent hGSCs have a higher tolerance against IFN-β treatment than sphere-like cells. We repeated the treatment assay using higher IFN-β concentrations, including 11, 33, and 100 ng/mL, using both hGSCs and hNSCs (Fig. 1e, h). High concentrations of IFN-β also inhibited hGSC growth, but no significant morphological changes were observed among the treated hNSCs (Fig. 1f, i). The quantitative analysis of cell coverage supported our visual observations (Fig.  1g, j). In conclusion, both low-dose (0.625 ng/mL) and high-dose (up to 33 ng/mL) IFN-β treatments were able to block cell growth in hGSCs, without affecting hNSCs.

The nuclear sizes of hGSCs was decreased by IFN-β
We next performed immunostaining to detect the expression of Ki67, S100-β, and Sox2 in hGSCs after 6 days of treatment of IFN-β at both low (11 ng/mL) and high (33 ng/mL) doses. Similar to our morphological observations, the number of Ki67-positive hGSCs decreased after IFN-β treatment (Fig. 2a, b), as did the numbers of S100-β-and Sox2-positive cells (Fig. 2a, b). The quantitative assessment of Ki67, S100-β, Sox2, and DAPI nuclear stain in hGSCs following IFN-β stimulation can be seen in Fig. 2c. Because the total cell number decreased, the ratio of Ki67 to DAPI staining did not show a difference between the control group and the IFN-β-treated group. Overlapping images showed that the relative expression level of S100-β increased significantly after IFN-β stimulation (Fig. 3a). IFN-β might represent an effective agent for controlling cell differentiation in hGSCs [34]. S100-β has long been considered a biomarker for astrocytes [35]. Recently, some studies have also identified the high expression levels of S100-β with malignant tumors [36]. The increased S100-β expression observed in hGSCs following IFN-β stimulation agrees with previous reports. The higher magnification image revealed the nuclear status (Fig. 3b). Representative images of DAPI staining and the schematic diagram of the nucleus revealed significant nuclear enlargement after IFN-β stimulation (Fig. 3c). The nucleus size was quantified using ImageJ, which revealed that the size of the nucleus increased in a dose-dependent manner following IFN-β treatment (Fig. 3d).

The cell growth and morphology of hGSCs was reduced by repetitive stimulation treatment with IFN-β
Clinically, IFN-β is administered over a long time course. Thus, we examined the effects of continuous IFN-β stimulation. Both hGSCs and hNSCs were cultured in either control medium or IFN-β medium for one generation (6 days), and then the control cells were passaged into control medium, whereas IFN-β-treated cells were separately passaged into both control medium and IFN-β medium (Fig. 4a), generating three groups of cells: control to control, IFN-β to control, and IFN-β to IFN-β. On day 2, the number of sphere-like hGSCs in the IFN-β to IFN-β group decreased compared with the numbers observed in the control to control group (Fig. 4b). The hGSCs conditions in IFN-β to control group showed more sphere-like cells than IFN-β to IFN-β group but fewer than observed in the control to control group. Although no significant effects were observed among hNSCs following short-term IFN-β stimulation (Fig. 1i), continuous IFN-β stimulation resulted in decreased cell growth (Fig. 4c). We observed these groups of cells again on day 8, which revealed very few surviving sphere-like cells in hGSCs within the IFN-β to IFN-β group (Fig. 4d). The enlargement of both the cell body and the nuclear size was also observed in this group (Fig. 4e). The hGSCs in the IFN-β to IFN-β group displayed an oligodendrocyte-like morphology, with a long and massive synapse (Fig. 4e). Quantitative analysis of hGSCs, including the coverage rate (Fig. 4f) and the single clone area (Fig. 4g), was performed on the surviving clones, which revealed significant reductions associated with IFN-β treatment.

Different gene responses in hGSCs and hNSCs after IFN-β treatment
To better understand the molecular mechanism associated with the IFN-β treatment effects observed in hGSCs and hNSCs, we performed RNA sequencing on samples treated with or without IFN-β in both hGSCs and hNSCs. Gene clustering analysis indicated the high quality of our wholetranscriptome data (Fig. 5a). The analysis of DEGs was used to identify genes that were up-and downregulated following IFN-β stimulation, resulting in 1,707 and 1,338 genes designated as hGSC + and hGSC − genes, respectively, whereas 1,553 and 1,169 genes were respectively designated as hNSC + and hNSC − genes (Fig. 5b). The integrative analysis of these four groups resulted in the identification of 995 genes characterized as hGSC − hNSC NA (genes only downregulated in hGSCs but with no change in hNSCs) and 969 genes characterized as hGSC + hNSC NA (genes only upregulated in hGSCs but with no change in hNSCs) (Fig. 5c). We then performed GO and KEGG analyses on these two groups to further explore the potential downstream mechanisms associated with the response to IFN-β. The GO biological process (BP) analysis revealed that hGSC + hNSC NA genes were primarily enriched in cytokine-mediated signaling, immune response, response to external stimulus, and cell adhesion pathways (Fig. 5d). The KEGG analysis revealed that hGSC−hNSC NA genes were enriched in the cell cycle and ribosome pathways (Fig. 5e). Our whole-transcriptome analysis results agreed with our previous morphological observations, supporting increased cell adherence and decreased cell growth. The expression patterns of the immune response and cell adhesion genes were displayed as heatmaps for both hGSCs and hNSCs (Fig. 5f). Consistently with known IFN-β downstream pathways, JAK2, STAT6, and NFKB1/2 were identified within the immune response pathways [19,37,38]. We also explored the detailed expression patterns of 22 genes in the cell cycle pathway (Fig. 5g). Typical cell cycle-related genes, such as CCNA2, CCNB1, CCNB2, and CCND1, were significantly downregulated following IFN-β stimulation in hGSCs but remained unchanged in hNSCs.

IFN-β decreased cell proliferation in hGSCs
IFN-β decreased hGSCs cell growth on both the morphological and genetic levels. Cell cycle-related genes were significantly downregulated after IFN-β stimulation ( Fig. 5g and 6). The number of sphere-like cells also reduced significantly following IFN-β treatment, and both cell body and nuclear size increased. Simultaneously, genes associated with cell adhesion were upregulated in hGSCs, which supported the observed reductions in sphere-like cells and the enlargement of the cell nucleus. We also observed multiple synapses in hGSCs under conditions of continuous stimulation with IFNβ, and this morphological change may be associated with changes in gene expression in the immune response pathway.

Discussion
Our results indicated that IFN-β affected hGSCs rather than hNSCs. hNSCs are considered to be the most likely initiation cells of malignant glioma [17,25,26]. On the other hand, hNSCs have been considered efficient vehicles for the delivery of anti-cancer agents to tumor sites during therapeutic applications [39,40]. IFN-β has been found to exert antiproliferative effects in many cancer cell types [41,42]. IFN-β can decrease cell proliferation and progression through the inhibition of the cell cycle S-phase in glioma cell lines [43]. However, the underlying molecular mechanisms responsible for the subsequent differences observed following IFN-β treatment between cancer stem cells and normal NSCs remain unclear. Microarray-based gene expression profiling performed in three glioma cell lines and primary B-cells following IFN-β treatment was reported in 2014 [44,45]. Several novel IFN-β response genes were identified in primary B-cells, including NEXN, HAPLN3, DDX60L, and IGFBP4 [45]. In this study, RNA sequencing was performed to reveal the molecular mechanisms associated with the response to IFN-β. We explored potential IFN-β response genes in glioma cells by performing DEG analysis between hGSCs and hNSCs. A total of 969 genes were identified as IFN-β-upregulated genes, whereas 995 genes were identified as IFN-β-downregulated genes in hGSCs compared with hNSCs. This large number of genes provides multiple opportunities to identify potential therapeutic targets that can be combined with IFN-β treatments for clinical applications.
A previous study indicated that IFN-β could induce spherogenicity in hGSCs [44]. In our study, we observed a significant reduction in sphere-like cells following IFN-β stimulation. The GO analysis results indicated enrichment in cell adhesion genes, providing strong support for a functional role for IFN-β in cancer stem cell morphology. NEXN encodes nexilin, which acts as a linker protein for the cytoskeleton and plays a functional role in focal adhesion junctions [46]. NEXN was previously identified as a novel IFN-β response gene in multiple sclerosis [45]. In the wholetranscriptome analysis following IFN-β treatment, NEXN was identified as a top gene in the enriched cell adhesion pathway in hGSCs after IFN-β treatment. Therefore, NEXN may also represent a novel IFN-β response gene in hGSCs.
The genes that were downregulated by IFN-β treatment were primarily enriched in cell cycle pathways, which also provides evidence to support the previously identified inhibitory role of IFN-β in cancer cells [41][42][43]. A total of 22 cell cycle-related genes were found to be downregulated in hGSCs, but no significant differences were observed in hNSCs. We suggested that these 22 genes may play irreplaceable roles in response to IFN-β treatment in hGSCs. Recently, one study attempted to combine IFN-β treatment together with a novel, effective, cyclin-dependent kinase (CDK) inhibitor (TG02), which is used clinically to treat GBM [47][48][49], resulting in synergistic functions in human glioma models [18]. In our data, CDK1 was identified in the transcriptome analysis. Many other well-known cell cycle-related genes, such as the cyclin family genes (CCNB1, CCND2, CCNB2, and CCNA2), and MCM family genes (MCM4, MCM5, and MCM6), were also identified. These results may facilitate the development of more efficient therapeutic options by guiding the selection of useful anti-tumor drugs.

Conclusions
In summary, our study established the inhibition effects of IFN-β in hGSCs rather than in hNSCs. Additional morphological details were observed following IFN-β stimulation, such as larger cell bodies, increased nuclear size, fewer sphere-like cells, and more oligodendrocyte-like synapses. The subsequently transcriptome analysis using RNA sequencing was highly consistent with our morphological observations of hGSCs. The enrichment of genes involved in the cell cycle and cell adhesion pathways supported the observed reduction in cell growth reduction and the morphological changes associated with IFN-β treatment. In addition, our exploration of the genetic modification that occurs after IFN-β treatment in both hGSCs and hNSCs. This may benefit the design of new and more effective pharmacological strategies for GBM treatment.

Availability of data and materials
All data and materials are available upon request.

Ethics approval and consent to participate
Surgical specimens were gotten strictly according to Ethics Committee permission in Shanghai Tenth People's Hospital.

Consent for publication
Not applicable. The growth of hGCSs but not hNSCs is inhibited by IFN-β. a A schematic representation showing hGCSs treated with basic media containing 10 ng/mL fibroblast growth factor (FGF), and different concentrations of IFN-β: 0, 0.625, 1.25, 2.5, 5, and 11 ng/mL. b The sphere-like hGCSs were inhibited, and more adherent hGCSs were observed after the IFN-β treatment. c IFN-β treatment resulted in smaller single sphere-like cells than control cells, and the sizes of the cell spheres are concentration-dependent within a certain range of IFN-β concentrations. d Coverage rate of adherent hGCSs decreased significantly with 5 ng/mL and 11 ng/mL IFN-β treatments. e Schematic representations of hGCSs treated with high concentrations of IFN-β. f High-concentration IFN-β also inhibits hGCS growth. g The coverage rate of hGCSs decreased after treatment with 11 ng/mL, 33 ng/mL, and 100 ng/mL IFN-β. h Schematic representation of hNSCs treated with highconcentration IFN-β. i High-concentration IFN-β does not inhibit hNSC growth. j The coverage rate of hNSCs does not decrease following treatment with 11 ng/mL, 33 ng/mL, and 100 ng/mL IFN-β. Quantitative data measured using ImageJ. Data are presented as the mean ± SD. Student's t-test. * p < 0.05, * * p < 0.01, * * * p < 0.001.

Fig. 3
Merged images showing the relative expression level of S100-β and the increased size of the cell nucleus after IFN-β stimulation. a Representative overlapping images showing staining for Ki67/DAPI, S100-β/DAPI, Sox2/DAPI, Ki67/S100-β, and Sox2/S100-β in hGCSs. Three groups, control, 11 ng/mL, and 33ng/mL IFN-β treatment groups (yellow frames indicate enlarged areas). b Representative images from (a) at higher magnification. White arrows indicate S100-β expression. Enlarged area (scale bar, 50 µm). c Representative images of DAPI staining and a schematic diagram of the nucleus. d Quantitative analysis of nuclear size, as measured using ImageJ. Scale bar, 50 µm. Data are presented as the mean ± SD. Student's t-test. * * * p < 0.001. Fig. 4 Continuous IFN-β stimulation can enhance the inhibitory effects on hGCS growth, which differs from the effects on hNSCs. a Flow chart indicating the continuous stimulation timeline and imaging time points. The strategy of passage  stimulation  waiting  re-passage  stimulation  time point photography was adopted. b Representative hGCS images after continuous stimulation, re-passage, and stimulation, day 2. Three groups: control to control, IFN-β to control, and IFN-β to IFN-β (black frames indicate enlarged areas). c Representative hNSC pictures after continuous stimulation, re-passage, and stimulation, day 2. Three groups: control to control, IFN-β to control, and IFN-β to IFN-β (black frames indicate the enlarged area). d Representative hGCS images after continuous stimulation at re-passage, day 8. e Schematic diagram of hGCSs and hNSCs after repassage and stimulation, day 2. f-g Quantitative analysis of hGCSs, including coverage rate (F) and single-clone area (G), of the surviving clones, as measured by ImageJ. Scale bar, 50 µm. Data are presented as the mean ± SD. Student's t-test. * * p < 0.01, * * * p < 0.001.

Fig. 5
Transcriptome changes in hGCSs after IFN-β treatment. a Gene clustering from control (Con-1 and Con-2) and IFN-β treatment (IFN-1 and IFN-2) samples for the hGCSs/hNSCs RNA sequencing results. b Identification of significantly upregulated and downregulated genes as hGCS + (1,707 genes), hGCS − (1,338 genes), hNSC + (1,553 genes), and hNSC − (1,169 genes). c The Venn diagram shows the overlap among the 4 gene lists identified in (B) to identify hGCS−hNSC NA (995 genes) and hGCS + hNSC NA (969 genes) lists. d Biological process (BP) items in the Gene Ontology analysis of the hGCS + hNSC NA gene list. e KEGG analysis of the hGCS − hNSC NA gene list. f Expression patterns of immune response and cell adherent genes in hGCSs and hNSCs. g Expression patterns of cell cycle genes in hGCSs and hNSCs. Fig. 6 IFN-β reduced cell proliferation in hGCSs. The ability to form spheres was inhibited by IFN-β. The cell and nuclear sizes of hGCSs increased after IFN-β treatment. On a gene level, significant upregulation was observed among immune response and cell adherence-related genes. Cell cycle and ribosome-related genes decreased.