LncRNA HCP5 promotes malignant cell behaviors in esophageal squamous cell carcinoma via the PI3K/AKT/mTOR signaling

ABSTRACT The role of lncRNA HCP5 in esophageal squamous cell carcinoma (ESCC) remains unknown despite its involvement in different malignancies. MTT assay, EdU assay, TUNEL assay, transwell assay, and sphere formation assay were conducted to reveal ESCC cell viability, proliferation, apoptosis, migration, invasion, and stemness characteristics. FISH and subcellular fraction assays were performed to reveal the subcellular location of HCP5 in ESCC cells. Luciferase reporter assay and RIP assay were conducted to explore the downstream axis of HCP5. Our findings revealed that HCP5 expression was at a higher level in ESCC tissues and cells compared to that in control tissues and cells. Additionally, HCP5 promoted ESCC cellular activities by promoting proliferation, migration, invasion ability and stemness characteristics of ESCC cells as well as suppressing cell apoptosis. Furthermore, we found that HCP5 bound with miR-139-5p to upregulate PDE4A via the competing endogenous RNA network in ESCC cells. Importantly, HCP5 was discovered to stimulate the PI3K/AKT/mTOR signaling by regulating the downstream target genes. Finally, rescue assays indicated that HCP5 promoted ESCC cell growth by activating the PDE4A-medaited PI3K/AKT/mTOR pathway. HCP5 promotes ESCC cellular development by modulating the miR-139-5p/PDE4A pathway and stimulating the PI3K/AKT/mTOR signaling pathway, which may be conducive for the improvement of ESCC treatment.


Introduction
As a main subtype of esophageal cancer, esophageal squamous cell carcinoma (ESCC) is one of the most prevalent cancers and is the primary cause of cancerrelated death globally characterized by high aggressiveness and dismal prognosis [1,2]. In recent decades, the applications of new technologies and neoadjuvant therapy improve the survival of patients with ESCC. Esophagectomy is commonly used for treatment of advanced ESCC, and the 5-year survival rates of ESCC patients remain unfavorable due to its postoperative relapse [3]. Thus, to explore the biomarkers and molecular regulatory mechanisms underlying ESCC is crucially important for ESCC diagnosis and treatment.
Long noncoding RNAs (lncRNAs) are longer than 200 nucleotides in length with no protein coding capacity [4]. They act as essential players in the pathological progression of ESCC [5]. For instance, FAM83H-AS1 promotes the development of triple-negative breast cancer via the upregulation of MTDH [6]. LncRNA PVT1 promotes colorectal cancer cell malignant behaviors such as proliferation and invasion via miR-214-3p downregulation [7]. LncRNA HAND2-AS1 inhibits the proliferation, migration of ESCC cells by sponging miR-21 [8]. As a tumor promoter, lncRNA histocompatibility leukocyte antigen complex P5 (HCP5) contributes to anaplastic thyroid cancer progression by regulating miR-128-3p [9]. HCP5 facilitates prostate cancer cell growth by interacting with miR-4656 to elevate CEMIP expression levels [10]. HCP5 modulates the miR-139-5p/ZEB1 axis to promote the epithelialmesenchymal transition process in colorectal cancer [11]. Through GEPIA website, we found that HCP5 exhibited a higher level in esophageal carcinoma tissues compared to that in nontumor tissues, while the underlying mechanism of HCP5 in ESCC remains unclear.
The purpose of our work was to probe into the specific function and the underlying mechanism of HCP5 in ESCC development. HCP5 was hypothesized to be a ceRNA and its effects in regulating ESCC cellular processes were explored, which may expand the knowledge in ESCC pathogenesis.

Tissue collection and cell culture
Sixty-two paired ESCC tissues and adjacent nontumor tissues were collected from 62 ESCC patients at the Harbin Medical University Cancer Hospital (Heilongjiang, China). Before surgery, no patients received anti-tumor therapies and all patients have signed the written informed consents. The study was approved by the Ethics Committee of Harbin Medical University Cancer Hospital. ESCC tissue samples were immediately frozen in liquid nitrogen and preserved at −80°C after collection. The normal esophageal epithelial cell-line HEEpiC and four ESCC cell lines (KYSE30, TE1, EC9706, EC109) were purchased from Chinese Academy of Sciences Cell Bank (Shanghai, China) and incubated in DMEM (Gibco, Rockville, MD) with 10% fetal bovine serum (FBS; Invitrogen) in humidified conditions at 37°C with 5% CO 2 . In addition, the PI3K activator, 740 Y-P (50 μg/mL), which was commercially obtained from TocrisBioscience (Ellisville, MO, USA), was further used to treat the transfected KYSE30 cell lines.

Cell transfection
sh-HCP5#1 or sh-HCP5#2 was transfected into KYSE30 and EC109 cells to knock down HCP5 with sh-NC as a negative control. PDE4A or HCP5 was overexpressed by pcDNA3.1/PDE4A or pcDNA3.1/HCP5 with empty pcDNA3.1 vector as a negative control. MiR-139-5p mimics or inhibitor was synthesized for overexpression or downregulation of miR-139-5p and NC mimics/ inhibitor was used as a negative control. All vectors were provided by Invitrogen company and then transfected into ESCC cells by Lipofectamine 2000 (Invitrogen) for 48 hours.

Quantitative real-time polymerase chain reaction (qRT-PCR)
The RNA in ESCC cells or tissues was extracted by TRIzol reagent (Invitrogen

Western blot analysis
RIPA lysis buffer was used to isolate total proteins from KYSE30 and EC109 cells. Next, a BCA Protein Assay Kit (Thermo Fisher) was used for protein quantification. Subsequently, the proteins were loaded on 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis gel and subsequently transferred onto polyvinylidene fluoride membranes (Millipore, USA). After being blocked with 5% defatted milk for 1 hour at ambient temperature, the membranes were incubated with the primary antibodies purchased from Abcam company (Shanghai, China) at 4°C overnight with GAPDH as an internal control. Subsequently, these membranes were incubated with the secondary antibodies at 37°C for 1 hour. Thereafter, the enhanced chemiluminescence reagents (Millipore, Plano, TX, USA) were used for visualization of the protein bands which were quantified by Bio-Rad ChemiDoc XRS+ System.

5-ethynyl-2ʹ-deoxyuridine (EdU) assay
An EdU Detection Kit (Ribobio, Guangzhou, China) was used for cell proliferation analysis. The cells were cultivated in 96-well plates for 48 hours, and then incubated with 50 μM EdU labeling medium for 2 hours under 37°C. After being treated with 4% paraformaldehyde and 0.5% Triton X-100, the ESCC cells were dyed with 100 μL of DAPI solution (Beyotime) for 30 minutes at room temperature and then observed under an FSX100 fluorescence microscope (Olympus, Beijing, China).

Terminal deoxynucleotidyl transferase-mediated dUTP-biotin nick end labeling (TUNEL) assay
Cell apoptosis was evaluated with an in situ Apoptosis Detection Kit (Takara, Dalian, China). The cells were fixed by paraformaldehyde, permeabilized by Triton X-100 and incubated for 1 hour with TUNEL reaction mixture (Takara). The cell nuclei were stained by DAPI (Beyotime). Positively stained cells were visualized and counted with an Olympus FSX100 fluorescence microscope.

Transwell assays
The ESCC cells were incubated in the upper chambers (Millipore) which were supplemented with RPMI 1640 medium (serum-free). The lower chambers were treated with RPMI 1640 medium with 10% FBS. After 48 hours, the ESCC cells invaded into the lower chambers were fixated with formaldehyde and then dyed using crystal violet. An Olympus FSX100 fluorescence microscope was used to count the invaded cells in five randomly chosen visual fields. The cell invasion assay was conducted in similar steps except that the upper chambers were precoated with Matrigel (Corning Inc).

Luciferase reporter assay
The HCP5-Wt/Mut or PDE4A-Wt/Mut vectors (Genechem, Shanghai, China) were constructed by subcloning the wide type or mutant-binding sequence of HCP5 or PDE4A on miR-139-5p into pmirGLO reporter plasmid. These plasmids were cotransfected with miR-139-5p mimics and its negative control NC mimics into KYSE30 and EC109 cells using Lipofectamine 2000. Two days later, the luciferase activities of these reporter plasmids were accessed using a Luciferase Reporter Assay System (Promega, USA).

RNA immunoprecipitation (RIP) assay
RIP assay was conducted with an EZ-Magna RIP Kit (Millipore). The ESCC cells were lyzed in RIPA lysis buffer. The anti-Ago2 antibody and anti-IgG antibody (negative control) were used for RIP assay. Next, RIPA buffer was added to the lysates of ESCC cells supplemented with antibody-conjugated magnetic beads as previously described [12]. Finally, qRT-PCR was conducted to assess expression of immunoprecipitated RNAs.

Fluorescence in situ hybridization (FISH) assay
The HCP5 probe was synthesized by RiboBio. The 4% paraformaldehyde was used to fix the ESCC cells for 15 minutes followed by treatment of 20 μg/mL protease K. After being hybridized in 5 ng/uL hybridization solution (RiboBio) containing HCP5 probe at 37°C overnight, cells were blocked by bovine serum albumin. The nuclei were then stained with DAPI (Sigma-Aldrich) for 8 min. The Olympus fluorescence microscope was used for capture of the images.

Subcellular fraction assay
Nuclear and cytoplasmatic fractions of KYSE30 and EC109 cells were extracted using a PARIS Kit (Ambion, USA) according to the manufacturer's instructions. Nuclear and cytosolic HCP5 expression was detected by qRT-PCR with GAPDH as a cytoplasmatic control and U6 as a nuclear control.

Statistical analysis
The results in this study were exhibited as the mean ± standard deviation. The GraphPad Prism 7 software (GraphPad Software, Inc.) was used for statistical analysis. Student's t test was used to compare the differences between two groups and one-way ANOVA was applied to evaluate difference among multiple groups. The gene expression correlation was accessed using Pearson's correlation analysis. Each assay was performed in triplicates. A value of p < 0.05 was regarded as statistically significant.

HCP5 expression is distinctly elevated in ESCC tissues and cells
Based on GEPIA (http://gepia2.cancer-pku.cn/ #index), the boxplot revealed the significant upregulation of HCP5 in 182 esophageal carcinoma tissues compared to that in 286 normal tissues (Figure 1(a)). Then, qRT-PCR analysis was performed to detect the levels of HCP5 in ESCC tissues and cell lines. As was revealed, HCP5 was expressed at a higher level in 62 ESCC tissues than in 62 corresponding adjacent nontumor tissues (Figure 1(b)). HCP5 expression was higher in ESCC cell lines (KYSE30, EC109, EC9706, and TE1) than in normal esophageal epithelial cell line HEEpiC (Figure 1(c)).

HCP5 interacts with miR-139-5p in ESCC cells
FISH and subcellular fraction assays indicated the cytoplasmic distribution of HCP5 in ESCC cells (Figure 3(a,b)). Therefore, we hypothesized that HCP5 might modulate the progression of ESCC by functioning as a ceRNA. On the starBase (http://starbase.sysu.edu.cn/), three miRNAs were found to possess binding site with HCP5 under the indicated screening condition (Degradome data ≥ 1, Pan-Cancer ≥ 6) (Figure 3(c)). After silencing HCP5 in ESCC cells, the expression of miR-139-5p was increased, while expression of other miRNAs exhibited no significant changes (Figure 3(d)). Therefore, miR-139-5p was used in the following assays. MiR-139-5p expression was increased by transfection of miR-139-5p mimics into ESCC cells (Figure 3(e)). Subsequently, the putative binding sequence of miR-139-5p on HCP5 was predicted (Figure 3(f)). Reduction in the luciferase activities of wild-type HCP5 reporters was observed while that of HCP5-Mut reporters showed no significant change in ESCC cells after miR-139-5p overexpression, indicating the direct interaction of HCP5 with miR-139-5p (Figure 3 (g)). Additionally, RIP assay using anti-Ago2 was conducted. The effectiveness of anti-Ago2 has been verified by the western blotting (Figure 3 (h)). MiR-139-5p and HCP5 enrichment in the anti-Ago2 group was detected by the RIP assay, suggesting that miR-139-5p and HCP5 coexisted in RNA-induced silencing complexes (RISCs) (Figure 3(i)). Relative enrichment of HCP5 in the complex of anti-Ago2 in cells transfected with miR-139-5p mimics was higher than that in cells transfected with NC mimics (Figure 3(j,k)). Next, the miR-139-5p expression was downregulated in ESCC tissues (Figure 3(l)). A negative expression correlation between miR-139-5p and HCP5 was found (Figure 3(m)). Additionally, miR-139-5p expression was lower in ESCC cell lines than in HEEpiC cell lines (Figure 3(n)).

PDE4A is a downstream target gene of miR-139-5p in ESCC
We further probed into the downstream targets of miR-139-5p in ESCC cells. As shown in Figure 4(a), based on RNA22, PicTar and TargetScan databases, five mRNAs (AP3M1, NUFIP2, PDE4A, SEPT11 and FBN2) were predicted as putative targets of miR-139-5p. Among the 5 mRNAs, only PDE4A showed decreased expression in KYSE30 and EC109 cells transfected with miR-139-5p mimics (Figure 4(b)). Further, qRT-PCR analysis verified the higher PDE4A expression in ESCC tissues and cells than in control tissues and cells (Figure 4(c)). According to the prediction by the starBase, the binding site between miR-139-5p and PDE4A was obtained (Figure 4(d)). Luciferase reporter assay revealed the reduction in luciferase activity of PDE4A-Wt reporters caused by miR-139-5p mimics, which indicated the binding of PDE4A and miR-139-5p (Figure 4(e)). Based on results of RIP assay, PDE4A and miR-139-5p were enriched in the complex of anti-Ago2, which indicated that PDE4A and miR-139-5p coexisted in RNA induced silencing complex (Figure 4(f)). Relative enrichment of PDE4A in the complex of anti-Ago2 in cells transfected with miR-139-5p mimics was higher than that in cells transfected with NC mimics (Figure 4(g)). For further analysis, the expression of miR-139-5p was decreased in KYSE30 and EC109 cells by transfection with miR-139-5p inhibitor (Figure 4(h)). Moreover, mRNA and protein expression levels of PDE4A were decreased by HCP5 knockdown, which were rescued by silenced miR-139-5p (Figure 4(i,j)). In addition, the expression correlation of miR-139-5p and PDE4A in ESCC tissues is not significant; expression of PDE4A was positively correlated with that of HCP5 in ESCC tissues, as shown in Figure 4 (k). Moreover, we identified that overexpression of HCP5 and PDE4A promoted cell viability, proliferation, migration, invasion, and stem characteristics of KYSE30 and EC109 cells (Figures 5 and 6).
Moreover, HCP5 depletion-induced suppression on cell stemness and its relevant protein levels was counteracted by PDE4A enhancement or 740 Y-P treatment in ESCC (Figure 8(f,g); Supplementary Figure 2(e)).

Discussion
As previously suggested, HCP5 exerts its carcinogenic effect in various cancers, such as anaplastic thyroid cancer, prostate cancer and colorectal cancer [9,10,14]. Nevertheless, whether HCP5 exerts the same impact on ESCC progression and how it works remain unclear. Our present study elucidated that HCP5 was significantly upregulated in ESCC tissues and cell lines. HCP5 promoted ESCC cellular development by regulating cell malignant behaviors such as proliferation, migration, invasion, and stemness characteristics, which suggested the tumor-promotive role of HCP5 in ESCC. By serving as competing endogenous RNA (ceRNA), lncRNAs can upregulate mRNA expression via binding with miRNAs [15]. We identified that miR-139-5p can bind to HCP5 in ESCC cells. Expression of miR-139-5p was lower in ESCC tissues and cells than that in control tissues and cells. As a tumor suppressor, miR-139-5p suppresses colon cancer cell stemness properties as well as inhibit cancer cell migration by decreasing E2-2 expression [16]. MiR-139-5p knockdown promotes development of prostate cancer by SOX5 [17]. MiR-139-5p suppresses the cell growth and migration by targeting HOXA10 in endometrial cancer [18]. MiR-139-5p has also been identified as a member of another ceRNA axis composed of circ_0052867 and RAP1B in ESCC cells [19]. The present study revealed that HCP5 bound with miR-139-5p to suppress its expression. Inhibition of miR-139-5p rescued the inhibitory effects of silenced HCP5 on ESCC cell viability, HCP5-Wt and HCP5-Mut plasmids in KYSE30 and EC109 were evaluated by luciferase reporter assay. (h) The effectiveness of anti-Ago2 has been verified by the western blotting. (i) RIP assay was performed to reveal the relative enrichment of miR-139-5p and HCP5 in Ago2-precipitated products in KYSE30 and EC109 cells. (j,k) RIP assay was performed to reveal the relative enrichment of HCP5 in Ago2-precipitated products in KYSE30 and EC109 cells transfected with NC mimics and miR-139-5p mimics. (l) MiR-139-5p expression in 62 paired ESCC tissues and corresponding nontumor tissues was detected by qRT-PCR. (m) Expression correlation of HCP5 and miR-139-5p in 62 ESCC tissues was analyzed by Pearson's correlation analysis. (n) MiR-139-5p expression in ESCC cells (KYSE30, EC9706, EC109, TE1) and normal esophageal epithelial cell line HEEpiC was detected by qRT-PCR. *p < 0.05, **p < 0.01, ***p < 0.001. cells of the Ago2 or IgG group was confirmed by RIP assay. (g) Enrichment of PDE4A in Ago2-precipitated products in KYSE30 and EC109 cells transfected with NC mimics and miR-139-5p mimics was detected by RIP assay. (h) MiR-139-5p levels in KYSE30 and EC109 cells post transfection of miR-139-5p inhibitor were testified with qRT-PCR (i,j) The PDE4A mRNA and protein levels in KYSE30 and EC109 cells after transfection of sh-NC, sh-HCP5#1, or cotransfection of sh-HCP5#1+ miR-139-5p inhibitor were measured by qRT-PCR and western blot. (k) The expression correlation between PDE4A and miR-139-5p or HCP5 in 62 ESCC tissues was analyzed via Pearson's correlation analysis. *p < 0.05, **p < 0.01, ***p < 0.001. proliferation, migration, invasion, and stemness characteristics.
In conclusion, the paper disclosed that the HCP5/miR-139-5p/PDE4A axis promotes ESCC cell viability, proliferation, migration, invasion, and stemness characteristics via stimulating the PI3K/AKT/mTOR pathway, which may contribute to the diagnostic and therapeutic methods of ESCC.