m6A demethylase ALKBH5 suppression contributes to esophageal squamous cell carcinoma progression

Esophageal squamous cell carcinoma (ESCC) is a highly malignant gastrointestinal cancer with a high recurrence rate and poor prognosis. Although N6-methyladenosine (m6A), the most abundant epitranscriptomic modification of mRNAs, has been implicated in several cancers, little is known about its participation in ESCC progression. We found reduced expression of ALKBH5, an m6A demethylase, in ESCC tissue specimens with a more pronounced effect in T3-T4, N1-N3, clinical stages III–IV, and histological grade III tumors, suggesting its involvement in advanced stages of ESCC. Exogenous expression of ALKBH5 inhibited the in vitro proliferation of ESCC cells, whereas depletion of endogenous ALKBH5 markedly enhanced ESCC cell proliferation in vitro. This suggests ALKBH5 exerts anti-proliferative effects on ESCC growth. Furthermore, ALKBH5 overexpression suppressed tumor growth of Eca-109 cells in nude mice; conversely, depletion of endogenous ALKBH5 accelerated tumor growth of TE-13 cells in vivo. The growth-inhibitory effects of ALKBH5 overexpression are partly attributed to a G1-phase arrest. In addition, ALKBH5 overexpression reduced the in vitro migration and invasion of ESCC cells. Altogether, our findings demonstrate that the loss of ALKBH5 expression contributes to ESCC malignancy.


INTRODUCTION
Esophageal squamous cell carcinoma (ESCC) is a highly aggressive histological subtype of esophageal cancer reported in Asia, with China having one of the highest morbidity and mortality rates [1][2][3]. The mainstay of treatment includes surgical resection, radiotherapy, and chemotherapy [1][2][3]; however, these are associated with unsatisfactory clinical outcomes due to adverse effects and limited efficacy. A detailed understanding of ESCC immunobiology would be useful in developing efficient prognostic biomarkers and therapeutic targets that can detect the tumor at an early stage, consequently resulting in early diagnosis and treatment.

Reduced ALKBH5 expression is frequently detected in ESCC tissue
To study the involvement of ALKBH5 in ESCC progression, we first checked the expression of ALKBH5 protein in 206 paraffin-embedded, archived ESCC specimens and 31 adjacent non-cancerous tissues (NT) by immunohistochemistry (IHC). Low expression of ALKBH5 was detected in 1 NT sample (3%) ( Figure 1A-a and Table 1) and 150 ESCC specimens (73%) ( Figure 1A-b, c, 1B, and Table 1; P < 0.01). Western blotting revealed reduced ALKBH5 expression in 20 fresh ESCC specimens compared with adjacent non-cancerous tissues ( Figure 1C, 1D). Furthermore, IHC revealed ALKBH5 protein to be primarily located in the nucleus of cancer cells ( Figure 1A, 1E). Table 2 shows the relationship between ALKBH5 expression and several clinicopathologic characteristics of ESCC patients. Although no association was identified between ALKBH5 expression and age (P = 0.524) and sex (P = 1.000) using 206 ESCC samples (Table 2), ALKBH5 expression was inversely correlated with tumor size (T classification; P = 0.006), lymph node invasion (N classification, P = 0.019), clinical stage (I-II versus III-IV, P = 0.027), and histological grade (P = 0.023) in these patients ( Figure 1E, 1F and Table 2). Reduced expression of ALKBH5 was more frequently observed in T3-T4,  N1-N3, clinical stage III-IV, and histological grade III  tumors than in T1-T2, N0, clinical stage I-II, and  histological grade I-II tumors (Figure 1E, 1F and  Table 2), indicating ALKBH5 loss as a major molecular event in advanced cases of ESCC. Altogether, these results suggested the involvement of ALKBH5 in the progression of ESCC.

ALKBH5 overexpression inhibits the proliferation of ESCC cells in vitro
The T classification data ( Figure 1E, 1F and Table 2) revealed downregulated ALKBH5 in large-sized tumors, indicating an essential function of ALKBH5 in tumor growth. This observation prompted us to perform gain-of-function experiments to explore the effects of ALKBH5 on ESCC cell growth. Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) and western blotting confirmed the successful overexpression of ALKBH5 transgene in TE-13, Eca-109, and KYSE-150 cells (Figure 2A, 2B). The colony formation assay showed that ALKBH5-expressing TE-13, Eca-109, and KYSE-150 cells displayed considerably fewer and smaller colonies compared with vector-expressing cells ( Figure 2C, 2D; P < 0.01). These results suggested that re-expression of ALKBH5 markedly suppressed ESCC cell proliferation in vitro.

RNAi-induced ALKBH5 silencing promotes ESCC cell proliferation in vitro
Because ALKBH5 was downregulated in ESCC tissue specimens ( Figure 1A-1D and Table 1), we proposed that the loss of ALKBH5 expression was associated with ESCC progression. Thus, we next performed lossof-function experiments to further examine the effects of loss of ALKBH5 function on ESCC cell growth. Both shRNA-ALKBH5-1 and shRNA-ALKBH5-2 specifically knocked down the expression of endogenous ALKBH5 mRNA ( Figure 3A) and protein ( Figure  Altogether, these results demonstrated that ALKBH5 negatively regulated the in vivo tumorigenicity of ESCC cells.

ALKBH5 inhibits the G1-S phase transition of ESCC cells
Based on the growth inhibitory properties of ALKBH5, we next examined the cell-cycle distribution of vectorand ALKBH5-expressing ESCC cells. Our results suggested that growth-suppressive effects of ALKBH5 overexpression were partly ascribed to a G1phase arrest.
To further study the mechanism of ALKBH5 silencing in promoting ESCC cell growth, we analyzed the cellcycle distribution using propidium iodide (PI) staining. As shown in Figure 3E, 3F, ALKBH5 depletion in TE-13 and Eca-109 cells markedly decreased the proportion of G1/G0 phase cells, and simultaneously increased the proportion of S phase cells. Thus, ALKBH5 silencing promoted the G1-S phase transition of ESCC cells. Altogether, these findings showed that ALKBH5 altered the cell cycle distribution of ESCC cells.

ALKBH5 overexpression dramatically reduces the migration and invasion ability of ESCC cells
To study whether ALKBH5 overexpression directly suppressed the invasion and migration ability of ESCC cells, we examined the phenotypic changes in ESCC cells following ectopic expression of ALKBH5. Wound-healing assays showed that exogenous expression of ALKBH5 reduced the migration of TE-13, Eca-109, and KYSE-150 cells ( Figure 4A-4D). As indicated in Figure 4E, 4F, Transwell and Boyden assays showed increased migration and invasion abilities of ALKBH5-expressing TE-13, Eca-109, and KYSE-150 cells as compared with vector-expressing cells. Altogether, these findings suggested that ALKBH5 overexpression inhibited the migration and invasion of ESCC cells in vitro.

ALKBH5 silencing enhances ESCC cell migration and invasion in vitro
Because ALKBH5 was frequently under-expressed in N1-N3 tumors than in N0 tumors ( Figure 1E, 1F and Table 2), we next explored the effects of ALKBH5 knockdown on cell migration and invasion of ESCC cells. Wound-healing assays demonstrated that ALKBH5 silencing promoted the migration of both TE-13 and Eca-109 cells ( Figure 5A, 5B). Transwell migration and Boyden invasion assays showed that the shRNA-induced knockdown of endogenous ALKBH5 dramatically promoted the migration ( Figure 5C, 5D) and invasion ( Figure 5E, 5F) of TE-13 and Eca-109 cells. Altogether, the suppression of endogenous ALKBH5 expression enhanced the migration and invasion of ESCC cells.
Promoter hypermethylation of tumor suppressor genes is frequently reported in cancer cells [54,55]. We identified ALKBH5 as a tumor suppressor, whose expression was dramatically diminished in human ESCC clinical specimens. Further, its ability to prevent proliferation, tumorigenicity, migration, and invasion of ESCC cells confirmed its tumor-suppressive potential in ESCC cells, executed via epigenetic silencing of ALKBH5 promoter.
In conclusion, ALKBH5 functions as a tumor suppressor in the pathogenesis of ESCC. We believe ALKBH5 can be a promising therapeutic target for advanced ESCC. Further functional studies, using a combination of MeRIP-seq or miCLIP-seq and RNAseq assays, to demonstrate the involvement of m 6 Amodified mRNAs in ESCC are warranted to validate its therapeutic potential.

Clinical specimens
This study was performed using paired ESCC and adjacent non-cancerous tissue samples (n = 20) obtained from the Department of Thoracic Surgery, Nanfang Hospital, Southern Medical University, Guangzhou, China. The study was conducted as per the protocols approved by the institutional review board of the Second Affiliated Hospital of Guilin Medical University and Southern Medical University and complied with patient data safety guidelines. Informed consent was obtained from the patients. The inclusion criteria for the study were: (1) pathological diagnosis of ESCC without metastasis to distant organs; (2) no anticancer therapy before surgery; (3) matched healthy tissue samples (obtained from an area more than 5 cm from the tumor lesion margin) and absence of tumor cells in healthy tissues as confirmed by histopathological examination. The Edmondson-Steiner (E-S) grading system was used for the histological grading of tumors.
A tissue chip consisting of formalin-fixed, paraffinembedded 206 ESCC tissues and corresponding adjacent tissue punches was provided by Guilin Fanpu Biotechnology Co., Ltd. These samples were obtained between 2006 and 2010 from patients with primary ESCC and age ranging from 28 to 78 years (mean = 58.5 years) at diagnosis. Other information collected included patients' histopathologic and raw survival data.

Histological analysis and immunohistochemistry
Samples (human tumor xenografts established in nude mice, human ESCC clinical specimens, and adjacent healthy tissues) were fixed in 4% paraformaldehyde (PFA), prepared in a phosphate buffer, overnight at 4° C. The samples were embedded in paraffin, sectioned into 5 mm thick pieces, mounted on slides, dewaxed, and deparaffinized. Hematoxylin and eosin staining (H&E staining) was performed as per the standard protocols.
For immunohistochemical staining, samples were deparaffinized and rehydrated. Afterward, sections were treated under high pressure in a citrate buffer (pH 6.0) for 2 min for antigen retrieval. Bovine serum albumin (1%) and H2O2 (15 min at room temperature) were used to inhibit non-specific staining and quench endogenous peroxidase activity, respectively. Next, the sections were incubated with anti-ALKBH5 antibody (Sigma, HPA007196) overnight at 4° C. PBS served as the negative control. Subsequently, the sections were incubated with secondary antibody conjugated to HRP. The complex was visualized with DAB and counterstained with hematoxylin.

Western blotting
Total protein was isolated and the lysate was prepared. The proteins were separated using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The separated proteins were transferred to a polyvinylidene difluoride (PVDF) membrane. The blots were incubated with anti-GAPDH (Proteintech, 1:1000) or anti-ALKBH5 (Sigma, HPA007196, 1:1000) antibody, followed by incubation with HRP-labeled secondary antibodies. Enhanced chemiluminescence (ECL) (Cat. No: KGP1122, KeyGen Biotech) was used to visualize the protein bands. βactin was used as a loading control.

Colony formation assay
For the clonogenic assay, 500 cells/well were plated in 6-well plates and grown for 10 to 12 days. Afterward, colony formation assay was performed using a previously described method [59].

Cell cycle analysis
For cell cycle analysis, 2 × 10 5 cells per well were plated in 6-well plates. The cells were stained with PI and cell cycle distribution was studied using flow cytometry [59,60].

Tumor xenografts in animals
Male BALB/c nude mice (3-4 weeks old) were purchased from the Medical Laboratory Animal Center of Guangdong Province and were fed autoclaved water and laboratory rodent chow. Vector-or ALKBH5expressing Eca-109 cells (1.5 × 10 7 cells) or shSCR-or shALKBH5-expressing TE-13 cells (1.5 × 10 7 cells) were subcutaneously injected into the left or right dorsal thigh of the mice, respectively. The animals were monitored daily, and tumor volumes were measured using a caliper. Tumor volume was calculated using the following formula: volume = 1/2 (width 2 × length) 0.5 × width 2 × length. All animals were sacrificed on the 14th day after transplantation. All animal experiments were performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals and were approved by the Ethics Committee of the Southern Medical University.

Wound healing assay
The cells were grown to near confluence in 6-well plates, followed by starvation in a serum-free medium for 24 h. A wound was created on the cell monolayer using a sterile 200 µL microtip. The cells were again starved for 48 h. The cells were observed under an inverted microscope (Nikon, Japan), and images of scratch areas were captured at 0, 24, 36, and 48 h to study the migration of cells to the wound area.

Transwell migration and Boyden invasion assays
Transwell migration and Boyden invasion assays were performed using the methods described previously. For the transwell migration assay, vector-or AGING ALKBH5-expressing ESCC cells (2 × 10 5 ) or shSCRor shALKBH5-expressing ESCC cells (2 × 10 5 ) were seeded into the upper chamber of the transwell insert (Corning) with serum-free DMEM. For the Boyden invasion assay, the upper chamber was coated with Matrigel (BD Biosciences). DMEM with 10% FBS was added to the lower compartment as a chemoattractant. Cells were allowed to migrate for 17 h and 20 h in the transwell migration and Boyden invasion assays, respectively.

Statistical analysis
Data from three independent experiments are presented as mean ± standard deviation (SD). Statistical analysis was performed using SPSS 16.0. Statistical significance was assessed by Student's t-test (*P < 0.05, **P < 0.01, and # P < 0.001; NS: not significant).

AUTHOR CONTRIBUTIONS
The study was conceptualized and designed by DX, YS, KCC, XLL, and TXF. The experiments were performed and results were analyzed by TXF, YL, SJX, JWX, TYL, JXZ, XYL, YLL, SHH, JSJ, and YGT. DX, YS, XLL, TXF, and YL wrote the manuscript. All authors read and approved the final manuscript.