Pleckstrin homology-like domain family A, member 3, a miR-19a-3p-regulated gene, suppresses tumor growth in osteosarcoma by downregulating the Akt pathway

ABSTRACT Pleckstrin homology-like domain family A, member 3 (PHLDA3), is emerging as a critical regulator for multiple cancers. Nevertheless, the expression and role of PHLDA3 in osteosarcoma remain unknown. Herein, we purposed to elucidate the role of PHLDA3 in the progression and chemoresistance of osteosarcoma. According to the bioinformatics analysis, PHLDA3 expression was low in osteosarcoma patients, and low content was linked to poor prognosis. Additionally, activation of PHLDA3 suppressed osteosarcoma cell proliferation, migration, and chemoresistance, whereas PHLDA3 inhibition caused the opposite effects. Mechanistically, our data revealed that PHLDA3 negatively regulates the Akt/GSK3β signaling cascade in osteosarcoma. Furthermore, we found that miR-19a-3p might exert its oncogenic function by inhibiting PHLDA3 expression in osteosarcoma. These results demonstrated miR-19a-3p/ PHLDA3/ Akt/GSK3β axis has a pivotal role in osteosarcoma, and PHLDA3 is a prospective therapeutic target for treating osteosarcoma.


Introduction
Osteosarcoma (OS), a common bone cancer derived from mesenchymal cells, has a high incidence and mortality rate in adolescents [1,2]. It is characterized mainly by uncontrolled proliferation and a high metastasis rate: 75% of OS cases attack adjacent tissues [3]. Although several therapeutic strategies, such as surgery, adjuvant radiotherapy, and chemotherapy, have made significant advances in increasing the survival rate of OS patients, the 5-year rate of survival remains low [4,5]. Thus, it is urgent to elucidate the molecular mechanisms of OS to identify novel treatment targets.
In recent years, gene expression profiling chip, a high-throughput and effective technology, has been widely used in various fields of cancer research [6][7][8][9]. It can help in the discovery of disease-related genes and provide insight into the pathogenesis of cancers such as OS [10]. Ajimu Keremu et al. analyzed the GEO database and discovered NRSN2, a novel oncoprotein that promotes OS cell expansion and may be a prognostic biomarker of OS [11]. Hence, the gene expression profile will be helpful for us to identify reliable prognostic biomarkers to improve the prognosis and treatment of OS.
PHLDA3 is the p53 regulatory repressor of Akt [12]. It has a PH domain that can compete with the PH domain of Akt for membrane lipids docking and suppress Akt activity [13]. Recent investigations have linked PHLDA3 as a tumor repressor gene to diverse cancers, for instance prostate cancer [14], pancreatic neuroendocrine tumors [15], as well as breast cancer [16]. Nevertheless, the role of PHLDA3 in OS is still unclear.
Here, we try to investigate the potential role and modulatory mechanism of PHLDA3 in OS. Based on bioinformatics analysis, we hypothesized that PHLDA3 is a miR-19a-3p-regulated gene and lowly expressed in human OS tissues. Additionally, we investigate the effect of PHLDA3 on the Akt/GSK-3by blocking pathway. In summary, we aim to explore the function and its underlying mechanism of PHLDA3 in OS and provide a new treatment strategy for OS.

Identification of the differentially expressed genes
To identify the differentially expressed genes (DEGs) in OS, the gene expression profiles (GSE42352, GSE70414) were acquired from the GEO database. The DEGs were obtained from the gene expression profiles by utilizing the GEO2R online tool (http:// www.ncbi.nlm.nih.gov/geo/geo2r/). The adj. P < 0.05 and |logFC| > 1 were set as DEGs cutoff criterion.

Protein-protein interaction analysis
To explore the interaction of DEGs, we used the Search Tool for the Retrieval of Interacting Genes (STRING) database to analyze the protein interaction network (PPI), and the DEGs with a reliability score >0.15 were added to the Cytoscape software for the PPI.

Oncomine database analysis
Oncomine database (http://www.oncomine.org), a web-based microarray database, was used to analyze the transcription level of the DEGs in different cancer types. PHLDA3 gene expression in clinical cancer tissue was queried and compared that with normal tissue using Student t-test. The parameters included fold-change 2, P-value 1e-4, and gene rank top 10%.

UCSC Xena
The prognostic value of PHLDA3 expression in OS was assessed according to overall survival using the UCSC Xena browser (http://xena.ucsc. edu/), an online database including gene expression data and clinical data.
The expression efficiency was evaluated using Western blot analysis.

MTT assay and PI staining assay
Briefly, 3000 cells were inoculated into 96-well plates. At the specified point in time, 200 μl culture media containing 20 μl MTT was introduced into each well and left to grow for four hours. Afterward, we replaced the medium in every well with 200 μl DMSO. The OD values at 570 and 630 nm were read with a microplate reader (PerkinElmer).
For propidium iodide (PI) staining assay, OS cells were inoculated in 6-well plates and exposed to or without cisplatin (CDDP) for 24 h. The cells were then stained with PI (Coolaber, cp9161) and left to grow for 10 minutes (Coolaber, cp9161). Finally, the morphologic cell changes were recorded using a fluorescence microscope (Olympus).

Colony formation assay
We trypsinized the 143B/U2OS cells, and then sub-cultured 1000-3000 viable cells in 6-well plates and left the cells to grow for 14 days for adherence and colonization. After that, we removed the media cells, then stained the cells (in 0.04% crystal violet) to visualize colonies and assess proliferation capacity.

Sphere formation assay
Suspensions of cells were inoculated on ultra-low adherent 12 well plates at 1 × 103 cells per well in 1 mL serum-free DMEM-F12 medium (Gibco, MA) enriched with B27 (1:50) (Invitrogen, MA) along with 20 ng/ml bFGF and EGF. After 14 days of incubation, the number of spheres exhibiting a diameter > 100 mm in each well was recorded.

Cell migration assay
In compartments (BD, Biosciences, 8.00 mm pores), 143B/U2OS cells were inoculated in serum-free medium. We left the cells to migrate toward the serum-rich bottom chamber for 24 hours. Nonmigrating cells were scraped off the insert's apical surface, and membranes harboring migrated cells were fixed (in 4% paraformaldehyde), followed by staining (in crystal violet). The cells' average count per field of view was used to express cell count.
When the transfected 143B and U2OS cells grew to more than 90% of their original size on the 12well plate, a linear wound was made with a sterile 200-μl pipette tip. The cell fragments were rinsed thrice with PBS, before being inoculated in 2% FBS medium. The 12-well plate was photographed under a microscope. The cell migration images were taken after 16/24/36 h at a fixed position.

Annexin V-FITC staining and flow cytometry
Transfected 143B cells were trypsinized and collected in 300 μl of the provided Annexin V-FITC (AV) binding buffer. Cells were then co-inoculated in the dark with 5 μl AV and 10 μl PI for 10 minutes at RT (room temperature). Thereafter, we suspended the cells in binding buffer (400 μl) followed by gentle mixing. Lastly, the treated cells were subjected to flow cytometry (BD Biosciences).

Dual-luciferase reporter assay
The dual-luciferase enzyme reporter assay was conducted as previously documented [17]. The wildtype (WT) docking site of miR-19a-3p on PHLDA3 and the matching mutant (Mut) were cloned into the pGL3-basic vector. The PHLDA3 WT or PHLDA3 Mut was co-transfected into 143B/U2OS cells with miR-19a-3p mimics or inhibitor or their corresponding control. After 48 h, Promega Dual-Luciferases Reporter Assay kit (Promega E1980) was performed to measure luciferase activities according to the manufacturer's protocols.

Animal experiments
Animal experiments were conducted according with the National Institute of Health Guide for the Care and Use of Laboratory Animals with the approval of the Animal Research Committee of Dalian Medical University. 143B cells transfected with sh-PHLDA3 or shCtrl were injected into the flank of the nude mice. The mice were killed and tumors were removed from the mice at 18 days after inoculation.

Statistics and data analyses
The data are given as mean ± SD. The statistical assessment was done using one-way analysis of variance (ANOVA), with p < 0.05 signifying statistical significance.

Results
In the study, we aimed to elucidate the molecular mechanisms of OS to identify novel treatment targets. Based on bioinformatics analysis, we discovered that PHLDA3 expression was low in OS cells and tissues. Increased PHLDA3 in OS cells suppressed cell growth, migration, and enhanced cisplatin-triggered apoptosis. Whereas, PHLDA3 knockdown increased cell growth, migration, and dampened cell apoptosis. Furthermore, we established that PHLDA3 was a true target of miR-19a-3p and blocked the Akt/GSK-3β signaling cascade in OS cells.

PHLDA3 expression was downregulated in osteosarcoma cells and tissue samples
To determine the differentially expressed genes (DEGs) in OS, we obtained the associated expression microarray datasets from GEO databases, including GSE70414 and GSE42352. The two datasets were evaluated using the GEO2R data resource (http://www.ncbi.nlm.nih.gov/geo/geo2r/) to determine DEGs (|Log FC|> 1 and p-value< 0.05) between OS cells and bone mesenchymal stem cells (MSC). Volcano plots were used to display the DEGs from the two sets of each sample data (Figure 1a and b). A total of 551 and 1112 DEGs were obtained from GSE70414 and GSE42352 datasets, respectively. Individually, the Venn diagram revealed 16 common upregulated genes and 76 common downregulated genes ( Figure 1c). The heat maps of the 92 differentially co-expressed genes in various chips were generated using the 'pheatmap' package in R studio (Figure 1e). The differentially co-expressed genes were then imported into the STRING database, and those with a reliability score >0.15 were added to the Cytoscape software for the visual protein interaction network (Figure 1d).
We selected 23 genes from a literature search to screen for those with potential research value ( Figure 1f). Then, we extracted the expression profiles of the selected genes from GSE42352 datasets between osteoblast (OB) and high-grade OS pre-chemotherapy biopsy (Biopsies) and analyzed the influence of these genes on the clinical survival of patients with OS using UCSC Xena browser (http://xena.ucsc.edu/). We found that the changing trend of five genes (PHLDA3, FBLN5, ITGA5, LTBP2, and IGFBP4) in Biopsies and OB was consistent with OS cells and MSC outcomes (Figure 2a), and affected the prognosis of OS patients (Figure 2b). The transcriptome analysis of the five genes in different types of tumors in the Oncomine data resource (https:// www.oncomine.org/resource/login.html) revealed that the expression of PHLDA3, FBLN5, and LTBP2 was low in clinical samples of sarcoma (Figure 2c). Based on additional verification data sets, we eventually selected PHLDA3 as a candidate gene for further experimental verification.

PHLDA3 inhibits osteosarcoma cell proliferation and migration
The biological role of PHLDA3 in OS is poorly understood.
Therefore, we overexpressed PHLDA3 (OE-PHLDA3) in 143B and U2OS cells to explore its role in OS. The predicted level of PHLDA3 expression was verified using Western blot along with RT-PCR (Figure 3a). After that, we explored the influence of PHLDA3 on cell proliferation using MTT, colony formation, and spheroid formation assays. We found that increased PHLDA3 remarkably dampened the growth of OS cells, as illustrated in Figure 3b-f (p < 0.05). To verify this finding, we used lentiviruses expressing shRNA to knock down PHLDA3 expression. The knockdown efficiency was assessed using Western blot and RT-PCR (Figure 3g). The results indicated that silencing PHLDA3 significantly increased cell proliferation (Figure 3h-l, p < .05). Besides, sh-PHLDA3 or shCtrl transfected 143B cells were inoculated into nude mice to assess the functional role of PHLDA3 in vivo. As shown in Figure 3m-n, PHLDA3 silencing greatly increased the tumor weight of mice.
Subsequently, the effect of PHLAD3 on OS cell migration was assessed via wound healing along with Transwell assays. The results indicated that increased PHLDA3 suppressed OS cell migration (Figure 4a-d, p < .01), whereas PHLDA3 knockdown accelerated cell migration in 143B/ U2OS cells (Figure 4e-h, p < .01). Collectively, these results illustrate that PHLDA3 is a bona fide repressor of cell proliferation along with migration in OS.

PHLDA3 enhances cisplatin-induced apoptosis in osteosarcoma cells
To confirm the tumor-suppressive role of PHLDA3, we assessed its effect on chemoresistance in OS cells. We found that overexpression of PHLDA3 promoted cisplatin-induced apoptosis, as illustrated by the escalation in cleaved PARP (Figure 5b-c). Subsequently, these findings were corroborated by the MTT (Figure 5a Hence, these data illustrated that PHLDA3 plays an indispensable role in the chemoresistance in OS cells.

PHLDA3 suppresses cell proliferation and migration via Akt/GSK-3β signaling in osteosarcoma cells
Akt signaling is well recognized for its pivotal role in enhancing cell growth and dampening apoptosis in many cancers [18]. Previous studies indicate that PHLDA3 is an important suppressor of Akt signaling [19]. Therefore, to investigate whether PHLDA3 regulates OS cell proliferation along with migration through the Akt/GSK-3β cascade, we first assessed its effect on p-AKT and p-GSK-3β protein contents in OS cells. The results illustrated that PHLDA3 overexpression diminished the contents of p-AKT (ser 473) along with p-GSK-3β (ser 9) (Figure 6a-b), whereas PHLDA3 knockdown elevated the protein contents of p-AKT (ser 473) coupled with p-GSK-3β (ser 9) (Figure 6c-d).
To demonstrate that PHLDA3 suppressed cell growth along with migration by regulating the Akt cascade, OS cells with or without PHLDA3 silencing were inoculated with or without the PI3K/Akt signaling inhibitor LY294002. The data indicated that LY294002 remarkably diminished the p-AKT and p-GSK-3β induced by PHLDA3 knockdown (Figure 6e). Furthermore, cells   (e-h) PHLDA3 knockdown in 143B/U2OS cells, as well as its influence on cell migration were explored via Transwell assay (e-f) along with wound-healing assay (g-h). The data in (b), (d), (f), and (h) were given as means ± SD, **p < .01, ***p < .001 vs. control. The transfected 143B cells were exposed to different doses of cisplatin (CDDP) for 24 hours, and cells viability was explored via the MTT assay. (b-g) Osteosarcoma cells were inoculated with diverse levels of CDDP for 24 hours after OE-PHLDA3 transfection. Cell apoptosis was analyzed using Western blotting analysis of PARP (b-c), PI only staining (d-e), and Annexin V/PI double staining (f-g). (h) The transfected 143B cells were exposed to diverse levels of CDDP for 24 hours, and cells viability was explored via the MTT assay. (i-l) Osteosarcoma cells were inoculated with diverse levels of CDDP for 24 hours after Sh-PHLDA3 transfection. Cell apoptosis was analyzed using Western blotting analysis of PARP (i-j) and PI only staining (k-l). The data in (e), (f), and (l) were given as means ± SD, ***p < .001 vs. control. suppresses OS development by modulating the Akt/GSK-3β signaling cascade.

PHLDA3 is a direct downstream target of miR-19a-3p
To understand the modulatory mechanism of PHLDA3 in OS, we searched for potential PHLDA3 miRNAs in three databases: ENCORI, TargetScanHuman7.1, and Tarbase. The miRNAs were then verified using a PubMed literature review. The data illustrated that miR-19a-3p, a predicted miRNA of PHLDA3, was upregulated in OS [20,21] (Figure 7a). Thus, we purposed to establish the association of miR-19a-3p with PHLDA3. Subsequently, Target ScanHuman7.1 predicted that there was a docking site for miR-19a-3p in the 3′ UTR of PHLDA3 (Figure 7b). The relationship of miR-19a-3p expression with the survival outcomes in sarcoma patients was analyzed using ENCORI. The data illustrated that a low miR-19a-3p level may estimate a better survival outcome (Figure 7c, p < 0.05). Furthermore, there was a clear negative association of miR-19a-3p expression with PHLDA3 expression in sarcoma cancer samples (Figure 7d, p < 0.01), as well as a negative modulation of miR-19a-3p on PHLDA3 expression in 143B/U2OS cells (Figure 7e-f). The analysis of luciferase reporter genes revealed a direct cross-talk of miR-19a-3p with the 3'UTR of PHLDA3 (Figure 7g-k, p < 0.01). These data illustrate that PHLDA3 is a direct target of miR-19a-3p.
To confirm whether the effects of miR-19a-3p on OS cell proliferation along with migration were PHLDA3-dependent, we transfected miR-19a-3p into OS cells with or without PHLDA3 overexpression or knockdown. The MTT (Figure 8a, p < .05) and Transwell assays (Figure 8b-c, p < 0.05) results illustrated that restoration of PHLDA3 prevented the miR-19a-3p-induced increase in cell proliferation and migration. In contrast, PHLDA3 suppression prevented the miR-19a-3p inhibitor from decreasing cell proliferation (Figure 8d, p < .05) and migration (Figure 8e-f, p < 0.05).
Furthermore, to investigate whether miR-19a-3p regulated chemoresistance in OS cells via suppressing the PHLDA3/Akt pathway, we evaluated cell apoptosis using Western blotting (Figure 8g-h) and MTT (Figure 8i-j, p < 0.01). We observed that miR-19a-3p repressor increased chemosensitivity in 143B/U2OS cells, which was reversed by PHLDA3 knockdown. These data illustrated that miR-19a-3p has an oncogenic effect on OS cells by directly targeting PHLDA3.

Discussion
PHLDA3 is a novel tumor-related protein that mediates carcinogenesis of multiple cancers. However, the relevance of PHLDA3 in OS is still unknown. In this study, we explored the role of PHLDA3 in osteosarcoma and investigated the involvement of potential signaling pathway. We verified that PHLDA3 played a tumor suppressor role in osteosarcoma and suppressed cell proliferation, migration and enhanced cisplatin-induced cell apoptosis via regulating the Akt/ GSK-3β signaling pathway. Additionally, our data indicated that PHLDA3 was a true target of miR-19a-3b and miR-19a-3p promoted OS cell growth and migration by inhibiting PHLDA3 expression.
PHLDA3 was initially recognized as a p53 target gene with only a PH domain that competes with Akt. Accumulating studies have identified PHLDA3 as a vital tumor suppressor gene that is linked to tumorigenesis, disease progress along with poor prognosis in various human cancers. For example, decreased PHLDA3 expression was remarkably linked to tumor progress along with recurrence in individuals with squamous cell carcinomas, indicating a poor prognosis [22]. Loss of heterozygosity (LOH) at the PHLDA3 gene locus is linked to the progress and malignant phenotype of pancreatic neuroendocrine tumors (PanNETs) [13]. Moreover, several studies have demonstrated that PHLDA3 is differentially expressed in cisplatin-induced acute kidney injury patients [23,24]. In addition, PHLDA3 could regulate energy metabolism [19], impede somatic cell reprogramming [25], and ameliorate pressure overload-triggered cardiac remodeling [26]. Consistently, we proved that PHLDA3 inhibited OS cell proliferation, migration and enhanced cisplatin-triggered cell apoptosis. Hence, our findings also revealed that PHLDA3 is a tumor repressor gene in OS and elevated PHLDA3 may be beneficial to decrease  As a PH domain-only protein, PHLDA3 may compete with Akt, another PH-domaincontaining protein, for membrane lipid binding, thereby interfering with Akt translocation to the cellular membrane and activation. Most of the biological functions of PHLDA3 have been linked to blocking Akt signaling via its PH domain. It has been documented that PHLDA3 regulates the Akt/GSK3β signaling during somatic cell reprogramming [25]. As expected, we established that the suppressor role of PHLDA3 in OS was still exerted by suppressing the Akt-GSK3β signaling. The Akt/GSK-3β signaling is a widely-known indispensable controller in different biological processes of the multifarious types of tumor cells [27][28][29][30]. Accumulating evidence indicates that Akt/GSK-3β is linked to osteosarcoma progression [31] and anti-chemotherapeutic drugs [32]. Therefore, the role of the PHLDA3/Akt/ GSK-3β signaling in the regulation of OS is selfevident.
A number of reports have documented that the expression of PHLDA3 is low in many different types of tumors. For example, PHLDA3 is inactivated by both LOH and methylation in various neuroendocrine tumors (NETs), such as Lung NETs and PanNETs [33]. Typically, miRNAs can directly dock to the 3 UTR of the targeted mRNA, inhibiting translation or causing mRNA degradation [34,35]. Hence, we adopted bioinformatics tools to identify a docking site for miR-19a-3p in the PHLDA3 3'-UTR. Interestingly, the data confirmed that miR-19a-3p represses PHLDA3 expression. As a member of the known oncomiR-17-92 cluster [36], miR-19a-3p is frequently dysregulated and has pro-cancerous and pro-proliferative roles in many tumors, including gastric cancer [37], hepatocellular carcinoma [38], and breast cancer [39]. Overexpression of miR-19a-3p was identified in OS cells, and silencing of miR-19a-3p promoted its chemosensitivity via escalating the expression of tumor repressor PTEN [21]. Herein, we found that overexpressing miR-19a-3p increased OS cell proliferation, metastasis, and drug resistance. Thus, miR-19a-3p might have pro-tumorigenic effects in OS by inhibiting PHLDA3 expression.

Conclusion
Overall, our findings show that PHLDA3, a true target of miR-19a-3p, is less expressed in OS tissue specimens and suppresses the proliferation, malignant and chemoresistance biological behavior by affecting the Akt/GSK-3β signaling cascade. Furthermore, PHLDA3 may be considered an independent predictor of poor prognosis, as well as a prospective therapeutic target for OS patients.

Disclosure statement
No potential conflict of interest was reported by the author(s).