SLC34A2 promotes neuroblastoma cell stemness via enhancement of miR‐25/Gsk3β‐mediated activation of Wnt/β‐catenin signaling

Cancer stem cells contribute to cancer progression, but the mechanisms underlying neuroblastoma stem cell development are unclear. Here, we examined the roles of the transcription factor SLC34A2 in regulating the stemness of neuroblastoma cells. We found that SLC34A2 expression was negatively correlated with the overall survival and relapse‐free survival probability of neuroblastoma patients. Additionally, SLC34A2 expression was observed to be remarkably increased in spheroids derived from neuroblastoma cells. Knockdown of SLC34A2 attenuated the expression of stemness markers and spheroid formation capacity of neuroblastoma cell‐derived spheroids, and overexpression of SLC34A2 exerted the opposite effects in neuroblastoma cells. Mechanistically, SLC34A2 was found to directly bind to the promoter of MIR25, which targets glycogen synthesis kinase 3β (Gsk3β), an antagonist of Wnt signaling. Transfection of miR‐25 inhibitor or a Gsk3β overexpression plasmid attenuated the effects of SLC34A2 overexpression on the stemness of neuroblastoma cells. Our results demonstrate that miR‐25/Gsk3β‐mediated activation of Wnt signaling is responsible for SLC34A2‐induced enhancement of neuroblastoma cell stemness.

Cancer stem cells contribute to cancer progression, but the mechanisms underlying neuroblastoma stem cell development are unclear. Here, we examined the roles of the transcription factor SLC34A2 in regulating the stemness of neuroblastoma cells. We found that SLC34A2 expression was negatively correlated with the overall survival and relapse-free survival probability of neuroblastoma patients. Additionally, SLC34A2 expression was observed to be remarkably increased in spheroids derived from neuroblastoma cells. Knockdown of SLC34A2 attenuated the expression of stemness markers and spheroid formation capacity of neuroblastoma cellderived spheroids, and overexpression of SLC34A2 exerted the opposite effects in neuroblastoma cells. Mechanistically, SLC34A2 was found to directly bind to the promoter of MIR25, which targets glycogen synthesis kinase 3b (Gsk3b), an antagonist of Wnt signaling. Transfection of miR-25 inhibitor or a Gsk3b overexpression plasmid attenuated the effects of SLC34A2 overexpression on the stemness of neuroblastoma cells. Our results demonstrate that miR-25/Gsk3b-mediated activation of Wnt signaling is responsible for SLC34A2-induced enhancement of neuroblastoma cell stemness.
Neuroblastoma is the most common primary tumor in the central nervous system; its progression is rapid and recurrence rate is high [1]. Although the progress has been obtained on the treatment technology and scheme of neuroblastoma, the 5-year survival rate of neuroblastoma patients is still < 5% [2]. Cancer stem cells (CSCs) result in tumor progression [3]. It is helpful for the prevention and treatment of neuroblastoma to elucidate the mechanisms underlying neuroblastoma stem cells progression.
Transcription factor SLC34A2 was first identified in 1999 and is widely expressed in various tissues and organs at different levels [4]. A previous study has shown that the SLC34A2 level is upregulated in lung tissues [5], and mutations of the SLC34A2 gene are associated with pulmonary alveolar microlithiasis [6] and lung cancer [7]. SLC34A2 overexpression is an independent prognostic indicator in bladder cancer [8] and breast cancer [9]. These effects indicate that SLC34A2 might be engaged in tumor progression, which was confirmed by studies showing that SLC34A2 promotes proliferation and chemoresistance in colorectal cancer through reactive oxygen species (ROS)-hypoxia-inducible factor 1-induced enhancer of zeste homolog 2 upregulation [10], that SLC34A2 facilitates the progression of human osteosarcoma cells through phosphatase and tensin homologue-phosphoinositide 3-kinase-Akt signaling [11], and that SLC34A2 enhances hepatocellular carcinoma cell proliferation and invasion [12]. Notably, recent research shows that SLC34A2 expression is enhanced in breast CSCs and SLC34A2 induces chemoresistance via the SLC34A2-B cell-specific Moloney murine leukemia virus integration site 1-multidrug resistance-associated protein 5 axis [13]. However, the roles of SLC34A2 in neuroblastoma progression are still unclear.
Wnt signaling has been confirmed to be closely correlated with CSC progression [14,15]. Glycogen synthesis kinase 3b (Gsk3b), a multi-functional serine/threonine protein kinase, could promote the phosphorylation of b-catenin so that it can be degraded by proteasomes and subsequently inactivate Wnt signaling [16]. Previous studies have shown that Gsk3b could suppress stemcell-like properties and tumor growth of osteosarcoma, and induce G0/G1 arrest and apoptosis in menstrual blood-derived endometrial stem cells through inactivating Wnt signaling [17,18]. A previous study has shown that miR-25 could promote gastric cancer stem-like cell progression via directly targeting Gsk3b [19]. Bioinformatics analysis showed that miR-25 is a potential target of SLC34A2 and SLC34A2 expression was negatively correlated with the survival rate of neuroblastoma patients. Notably, SLC34A2 expression was remarkably decreased in neuroblastoma cell spheroids relative to parental cells, while miR-25 exhibited an opposite effect. Thus, we assumed that SLC34A2 might promote the stemness of neuroblastoma cells through miR-25/ Gsk3b-mediated activation of Wnt signaling. Further ChIP and luciferase reporter assays combined with in vitro experiments confirmed our speculation.

Online analysis tools
The R2 genomics analysis and visualization platform (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi) was used to analyze the correlation between SLC34A2 expression and neuroblastoma patients' survival rate, in which Kaplan-Meier analysis by gene expression was conducted. Three

Quantitative real-time PCR
Total RNA was extracted from cells using TRIzol reagent (Thermo Fisher Scientific) following the manufacturer's recommendation. Then cDNA for mRNAs was reversely synthesized using SuperScript TM First-Strand Synthesis System for RT-PCR (Invitrogen TM , Carlsbad, CA, USA) according to the standard procedure. cDNA for miRNAs was reversely synthesized using One Step miRNA RT kit (cat. no. D1801; HaiGene, Harbin, China) and quantitative real-time PCR (qRT-PCR) was performed on the StepOnePlus PCR system with TransStart Green qPCR SuperMix (Transgen Biotech, Beijing, China). GAPDH served as an internal reference. The relative expression level of transcripts was calculated using 2 ÀDDCt method.

RNA immunoprecipitation with Ago2 assays
For the detailed procedure, refer to the previous study [20]. Cells were lysed with 25 mM Tris/HCl buffer (pH 7.5) and 100 UÁmL À1 RNase inhibitor (Sigma-Aldrich, St. Louis, MO, USA), and then incubated with protein A Sepharose beads precoated with 3 lg anti-Ago2 antibody or control rabbit IgG for 1.5 h at 4°C. The RNA-protein complexes were pulled down by protein A/G agarose beads and RNA was extracted with TRIzol, followed by detecting the miR-25 level with qRT-PCR.

Spheroid formation assay
This process was outlined in the previous study [22]. Neuroblastoma cell spheroid formation was performed under anchorage-independent conditions in methylcellulose (Sigma-Aldrich). Briefly, neuroblastoma cells with different treatment were digested with Trypsin/EDTA (Sigma-Aldrich), and then cultured in DMEM/F12 medium supplemented with B27 (20 ngÁmL À1 ) and epidermal growth factor (10 ngÁmL À1 ) in non-adherent 24-well plates at 500 cells/well. After 8 days, spheres > 50 lm were counted. This experiment was performed in triplicate and repeated at least three times independently. Additionally, for experiments on spheroids, spheroids were trypsinized, re-seeded into plates and subsequently used in the experiments.

Statistical analysis
All results are denoted as mean AE SD and analyzed using PRISM (version X; GraphPad Software, La Jolla, CA, USA). Student's t test was used to assess the differences between groups. P < 0.05 or less was considered statistically significant.

SLC34A2 expression is negatively correlated with the overall survival and relapse-free survival of neuroblastoma patients
We firstly evaluated the correlation between SLC34A2 expression and the survival of neuroblastoma patients. The R2 genomics analysis and visualization platform (https://hgserver1.amc.nl/cgi-bin/r2/main.cgi) was used based on three represented dataset including different numbers of neuroblastoma patients. As shown in Fig. 1A-F, SLC34A2 expression was negatively correlated with the overall survival ( Fig. 1A-C) and relapse-free survival (Fig. 1D-F) of neuroblastoma patients. These results indicate that SLC34A2 might be engaged in neuroblastoma progression.

SLC34A2 promotes the stemness of neuroblastoma cells
Since CSCs contribute to tumor progression, we explored whether SLC34A2 regulates the stemness of  neuroblastoma cells. First, we collected spheroids formed by neuroblastoma SH-SY5Y cells ( Fig. 2A). qRT-PCR and western blot analysis indicated that SLC34A2 expression was significantly increased in cell spheroids formed by SH-SY5Y cells (Fig. 2B,C). Notably, expression of stemness markers (ALDH1 and Nanog) displayed a higher level in spheroids relative to parental SH-SY5Y cells (Fig. 2D,E). These results suggest that SLC34A2 might be involved in regulating the stemness of SH-SY5Y cells. As expected, SLC34A2 overexpression significantly increased the stemness marker expression and the spheroids formation capacity of SH-SY5Y cells (Fig. 2F-H). Consistently, SLC34A2 knockdown in spheroids formed by SH-SY5Y cells exerted opposite effects (Fig. 2I-K). The knockdown and overexpression efficiency of SLC34A2 was confirmed by qRT-PCR (Fig. 2L,M). Our results demonstrate that SLC34A2 promotes the stemness of neuroblastoma cells.

SLC34A2 directly binds to the promoter of MIR25 and increases its expression in neuroblastoma cells and spheroids
Previous study has shown that SLC34A2 could target the MIR25 promoter to promote gastric cancer progression (Fig. 3A) [19]; we wondered whether the SLC34A2-miR-25 axis exists in neuroblastoma cells. Notably, the miR-25 level exhibited a similar effect in SH-SY5Y cells and spheroids, characterized as an increase in spheroids derived from SH-SY5Y cells (Fig. 3B). qRT-PCR showed that SLC34A2 overexpression promoted miR-25 expression in SH-SY5Y cells, and knockdown of SLC34A2 decreased miR-25 expression in spheroids derived from SH-SY5Y cells (Fig. 3C). Furthermore, luciferase reporter analysis indicated that SLC34A2 overexpression increased the luciferase activity of pGL3-miR-25 in SH-SY5Y cells, while knockdown of SLC34A2 decreased pGL3-miR-25, but had no effect on pGL3-miR-25-mut (Fig. 3D,  E). Additionally, we performed a ChIP assay in SH-SY5Y cells and spheroids with an antibody against SLC34A2. Our results indicated that MIR25 promoter sequences with SLC34A2 potential binding site were enriched in DNA pulled down by anti-SLC34A2 in SH-SY5Y cells and spheroids (Fig. 3F).

SLC34A2 attenuates the stemness of neuroblastoma cells through the miR-25-Gsk3b axis
Finally, we investigated whether SLC34A2 attenuated the stemness of SH-SY5Y cells through miR-25-Gsk3b signaling. MiR-25 was knockdown or Gsk3b was overexpressed in SH-SY-5Y cells with SLC34A2 overexpression (Fig. 5A). As expected, knockdown of miR-25 or Gsk3b overexpression attenuated the promoting effects of SLC34A2 overexpression on the expression of stemness markers (Fig. 5B,C). Additionally, the increase of spheroid formation capacity mediated by SLC34A2 overexpression was partially abrogated by miR-25 knockdown or Gsk3b overexpression in SH-SY5Y cells (Fig. 5D,E). Notably, SLC34A2 overexpression indeed activated Wnt signaling, characterized as the increase of Wnt3a and bcatenin expression; this effect was partially rescued by miR-25 knockdown or Gsk3b overexpression (Fig. 5F,G). Therefore, these results suggest that SLC34A2 facilitates the stemness of SH-SY5Y cells at least through the miR-25-Gsk3b axis.

Discussion
Previous studies have shown that ROS-SLC34A2 fusion genes could promote tumor cells proliferation in glioma and non-small cell lung cancer [23,24]; this effect was regarded as being associated with the proliferation and invasion abilities of CSCs [13]. Further studies have indicated that SLC34A2 knockdown attenuates the proliferation of lung CSCs [25], and SLC34A2 overexpression facilitates the stemness of and confers chemoresistance on breast cancer cells [13,26]. However, the roles of SLC34A2 in neuroblastoma progression remain unclear.  Since CSCs contribute to tumor progression, here we focused on the role of SLC34A2 in regulating the stemness of neuroblastoma cells. Since there are no specific markers for neuroblastoma stem cells, we collected the spheroids, which had been confirmed as being enriched in CSCs [20,27]. Notably, we found that SLC34A2 expression was significantly increased in spheroids formed by SH-SY5Y cells, which prompted us to explore the roles of SLC34A2 in the stemness of neuroblastoma cells. As expected, we revealed that SLC34A2 positively regulated the stemness of neuroblastoma cells through a spheroid formation assay. To the best of our knowledge, this is the first study showing a role for SLC34A2 in neuroblastoma cell stemness. However, we must admit that the conclusion should be validated by in vivo experiment, and future works could focus on the role of SLC34A2 in regulating other functions of neuroblastoma cells, such as cell migration and invasion.
Mechanistically, we found that SLC34A2 could directly bind to the promoter of MIR25 and thus upregulate its level, which is consistent with the previous study showing that SLC34A could regulate miR-25 activity to affect gastric cancer progression [19]. As miRNAs exert their function through their downstream targets, we searched for potential targets of miR-25; Gsk3b attracted our attention as an antagonist of Wnt signaling which promotes CSC progression. We further confirmed our prediction through RIP and luciferase reporter assays. Finally, we demonstrated that SLC34A2-mediated effects on the stemness of neuroblastoma cells were at least through the miR-25-Gsk3b axis, establishing the SLC34A2-miR-25-Gsk3b regulatory axis in neuroblastoma cells. However, we cannot exclude that there are other pathways involved in SLC34A2-mediated effects in neuroblastoma cell stemness.
In conclusion, our results indicate that SLC34A2 could directly bind to the promoter of MIR25 and thus facilitate miR-25-Gsk3b axis-mediated activation of Wnt signaling, which is responsible for the SLC34A2-mediated effects on the stemness of neuroblastoma cells. Importantly, SLC34A2 expression is negatively correlated with the overall survival of neuroblastoma patients. This work provides evidence confirming that SLC34A2 could be a novel candidate for neuroblastoma treatment or prognosis.