METTL13 Mediates the Translation of Snail in Head and Neck Squamous Cell Carcinoma

Background: Accumulating evidence supports that Methyltransferase like 13 (METTL13), which is implicated in protein binding and synthesis, act as oncogenic molecule. This study aimed to reveal the potential mechanism of METTL13 in head and neck squamous cell carcinoma (HNSCC) via mediating enhancing translation eciency of Snail. Methods: The expression of METTL13, Snail and ki67 were detected in patient HNSCC and paired para-carcinoma tissues. Knockdown and ectopic expression experiments were used in HNSCC cells to investigate their effects on tumor cell migration, proliferation and apoptosis. RNA-sequencing, polysome proling and SUnSET assays were developed to obtain the further mechanism of the interaction among METTL13, Snail and translation. Results: METTL13 was signi ﬁ cantly upregulated in HNSCC at both mRNA and protein level. Increased METTL13 was negatively associated with clinical prognosis. And METTL13 markedly affected HNSCC cellular phenotypes in vivo and vitro. Further mechanism study revealed that METTL13 could regulate EMT signaling pathway by mediating enhancing translation eciency of Snail, the key transcription factor in EMT, hence regulating the progression of EMT. Conclusion: Altogether, our study had revealed the clinical signi ﬁ cance of METTL13 in HNSCC, and providing a potential therapeutic strategy. effect METTL13 decrease on proliferation Cal27 and experiments were triplicate. Cal27


Background
Human head and neck squamous cell carcinoma (HNSCC) is one of the most commonly diagnosed malignancies worldwide with high mortality and morbidity, making it a growing health burden [1,2]. And this highly malignant disease is characterized by poor prognosis due to its high local recurrence and lymph node metastasis [3]. As for the metastasis of malignant tumor, epithelial-mesenchymal transition (EMT) was considered to be a crucial part [4,5]. It was activated in tumor progression [6,7] and affected cancer stem cell cellular behaviors [8,9]. During EMT, a small subset of cancer cells acquire motility with decreased adhesive ability and obtain invasive properties. Several EMT-induced transcription factors, such as Slug, Twist, Snail, have been considered to be involved in the progress [10]. Speci cally, Snail, a major transcription factor governing EMT, equip cancer cells with malignant phenotype like tumor recurrence and metastasis. Furthermore, although many of the signaling pathways could impinge on the transcriptional regulation of the Snail gene [11], post-transcriptional mechanisms affecting Snail are also emerging as crucial regulators of its activity [12].
Especially, aberrant regulation of mRNA translation to protein often account for the progression of cancer [13]. For instance, the subunits of the eukaryotic translation initiation factor 4F (eIF4F) complex and eukaryotic elongation factor 2 (eEF2), were frequently found to be dysregulated in multiple cancer types [14,15]. Increased activities of translation regulators enable the cancer cells to rapidly produce selective proteins in response to cancer development. For example, eukaryotic translation elongation factor 1 alpha 2 (eEF1A2) can mediate the expression of EMT factor matrix metallopeptidase 9 (MMP-9) in pancreatic cancer [16]. In addition, chemical modi cation of translation regulators by phosphorylation, acetylation, methylation and ubiquitination can result in changes of translation e ciency. For instance, phosphorylation of eIF2a leads to the gene-speci c translation, which further support cell survival in human cancers [17,18].
Recently, studies reported that human methyltransferase like 13 (METTL13) can catalyze dimethylation of eukaryotic elongation factor 1A (eEF1A) lysine 55 (eEF1AK55me2) to promote the protein synthesis in Ras-driven cancers [19,20]. METTL13 was characterized with oncogenicity, as it was involved in various diseases, such as gastric cancer [21], lung cancer [22,23] and breast cancer [24]. METTL13 also functions as a potential inhibitor of apoptosis, for it was found that miR-16 promoted apoptosis of tumor cells by silencing protein synthesis through posttranscriptional regulation [25]. Collectively, these ndings suggest that METTL13 could be a crucial gene of tumorigenesis and development.
In the present study, we examined the expression of METTL13 in paired HNSCC and para-carcinoma samples, and found METTL13 was signi cantly increased in HNSCC. Further study con rmed it was related to overall survival. And METTL13 knockdown signi cantly inhibits the proliferation and induced apoptosis in vitro. Conversely, the overexpression of METTL13 can promote malignant phenotypes of tumor. By RNA-seq, we gured out that METTL13 regulated the EMT pathway. In terms of mechanism, we managed to demonstrate that METTL13 could enhance translation e ciency of Snail, the key factor in EMT, hence regulating the progression of EMT. Therefore, our data revealed the function and underlying mechanism of METTL13 in HNSCC, providing both a prognostic marker and drug candidate for this malignancy. The total RNA of cells was obtained according to the instructions provided by the TRIZOL reagent (Invitrogen, 15596026). After that, 1 µg mRNA was utilized to reverse to cDNA by the standard procedure of the PrimeScript RT master mix kit (TaKaRa, RR036A). Next, TB Green™ Premix Ex Taq II kit (TaKaRa, RR820B) was used for real-time PCR according to the instructions and operated under the Bio-Rad CXF96 real-time system (Bio-Rad, USA). β-actin was functioned as internal control to calculate the relative quantity. The primers used for realtime PCR were listed in Table 2. 2.4. Total protein extraction and western blotting analysis.

Materials And
The total protein of HNSCC cell lines was harvested by with ice-cold radioimmu-noprecipitation assay (RIPA) lysis solution containing 1% PMSF protease inhibitor. And Pierce™ BCA Protein Assay Kit (Thermo Fisher Scienti c, 23227) had been used to de ne protein concentration in cell lysates. Before the western blot assay, the protein was dissolved in loading buffer and denatured at high temperature for ten minutes.

Colony formation assay
Two-hundred Cal27 or Cal33 cells were added into 6-well plates and were cultivated for about 2 weeks. Cells were xed for 30 minutes after 14 days with 4 percent paraformaldehyde and stained with 0.1 percent violet crystal. Each group had been repeated in triplicate.
Cell proliferation ability was assessed via CCK8 assay (Dojindo, CCK8-500). According to the standard procedure, 100 µl cell suspension containing 2 × 10 3 cells was seeded into each well of the 96-well plate and cultured in a cell incubator. At equal time intervals (four times every 24 hours), 10 µl reaction agent was added to each well and cultured for 1 hour at 37℃, and then taken out for absorbance value detection at 450 nm (In nite 200 PRO). 2.7. Wound healing and Cell migration assay.
As for the wound healing experiment, cells were transplanted and grown until they developed a monolayer of 90 percent fusion. Cells were then removed with the tip of an aseptic straw and treated in FBS-free medium according to the instructions. The scratched visual eld was randomly selected under the microscope, and the migration distance of the cells to the scratched area was measured.
Transwell system was utilized for migration analysis. The cells were cultured in pure DMEM in the cell culture insert (Merck millipore, MCHT06H48) with a nal concentration of 1 × 10 4 /ml. And DMEM containing 10% FBS coated in the bottle of 24-well plate was served as a chemical inducer. The cells were xed with 4% paraformaldehyde at 4 °C for 30 minutes after 18 hours of incubation at 37 °C and 5% CO 2 .
Cells were stained with crystal violet after rinsing with PBS and incubated for 45 minutes at room temperature. The cells on the cell culture insert's upper side were erased with a cotton swab, and washed thoroughly with PBS. and washed thoroughly with PBS. The average number of cells was calculated according to the random visual eld. Each sample were counted for three chambers.

Flow cytometric analysis of Annexin V apoptosis and cell cycle assay.
Cell apoptosis rate was detected using ow cytometry as per the Annexin V apoptosis detection kit (Dojindo, AD10) instructions. HNSCC cell lines were transplanted into 6-well plates (1 × 10 5 /ml per well) for apoptosis analysis and cultivated at 37 °C for 48 h in complete medium. Cells were then harvested and stained with the kit after the treatment with non-EDTA trypsin. HNSCC cells used for tracking the cell cycle were added into 6-well plates (5 × 10 5 /ml per well) and grown overnight in complete medium at 37 °C. Cells were collected after digestion with non-EDTA trypsin, and xed with 75% ethanol overnight.
And cells were stained with Kit (Multiscience, 70-CCS012) for 15 min the next day. Both of the two experiments were performed on a ow cytometer with BD FACSDiva Software 6.0 (BD Biosciences, New York, NY, USA). All the experiments were repeated at least three times.
2.9. Flow cytometric analysis of Aldehyde dehydrogenase activity assay.
Aldehyde dehydrogenase activity assay was conducted by the ALDEFLUOR uorescence reagent system (Stem Cell Technology, 01700). In short, we labeled each sample with a test tube and a control tube. The cells were lysed and collected into the test tube and activated ALDEFLUOR reagent was added to each cell suspension. Half of the cell uid in the test tube was added to the control tube, and 5 µl diethylaminobenzaldehyde (DEAB) were immediately added to the control tube. Incubated at 37 °C for 45 minutes and centrifuged. Then the supernatant was removed and re-suspended in the determination buffer. Finally, the analysis was performed on the ow cytometry.

2.11.RNA-sequencing
Cal27 and Cal33 cells transfected with shCtrl, shRNA-1 and shRNA-2 were sent for RNA-sequencing. Brie y, the total RNA was Total RNA was extracted using TRIZOL as mentioned previously. RNA quality and purity were checked by Nano-300.
Following the Illumina protocol, the construction of cDNA library was achieved from RNA samples for Illumina paired-end (PE) sequencing. The library preparations were subsequently sequenced on a platform of Illumina HiSeq 2000 (Illumina, San Diego, CA) and paired-end reads (2100 bp) were produced at Novogene Bioinformatics Technology Co., Ltd (Beijing, China).
As for translation analysis, the cells were incubated with puromycin (Beyotime, ST551-50 mg) of 1 µM nal concentration at 37℃ for 30 min. After that the cells were collected and lysed in RIPA buffer containing with 1 mM PMSF and protease inhibitor. The whole cell lysate of 5-10 µg was detected by western blotting with anti-puromycin primary antibody (Sigma-Aldrich, MABE343).

Polysome Pro ling and fractionation
Polysome pro ling assay was conducted as described previously [27]. In short, Cal27 and Cal33 cells with control or METTL13-decreased were cultured in 15 cm dishes (about 10 × 10 6 cells in each dish). Then cells were loaded with actinomycin at 100 µg / ml, then incubated for 2 minutes at 37 °C. Then cells were collected and cell extract was loaded on a sucrose gradient of 10-50% and prepared with a BioComp gradient station. The gradient was rotated at 36000 rpm and 4 °C for 2 hours in the TH-641 rotor (Sorvall). The gradient (260 nm) was analyzed and the components were collected by the BioComp gradient station.
According to the method we used before [28], the stem cell culture medium containing 3 × 10 3 cells/ml will be put into the 6-well ultralow attachment plates. Incubate for 7-10 days in a 37 °C incubator containing 5% CO 2 . The stem cell culture medium was replaced every two days. After spheres were formed, they were digested into single sphere by 0.25% Trypsin-EDTA. Photographs and counts were taken under a microscope.

Animal studies
As for the in vivo subcutaneous transplanted model studies, four-week-old BALB/C female nude mice obtained from Center of Experimental Animal of Sun Yat-sen University were divided into three groups (n = 6 per group) at random. HNSCC cell lines that transfected with METTL13-de cient or control plasmid were subcutaneously injected into mice to explore the effect in vivo. After four weeks, all the mice were sacri ced and we resected xenograft tumors intactly. The volume of transplanted tumor volume was determined as follows: length × width 2 /2. All the experimental studies were performed in the Sun Yat-sen University Laboratory Animal Centre. All procedures of animal experiments were approved by Sun Yat-sen University Animal Care and Use Committee.

Statistical analysis
The experiments had been replicated at least three times for in vitro research, each experiment was performed in duplicate or triplicate. The sample size necessary for vivo studies to discern difference in statistical signi cance is at least 6 animals in each group.
Data was analyzed by unpaired Student t-test to assess statistical signi cance. Data was shown as mean ± SD. All the mathematical analysis of the data was done by GraphPad Prism 8.0.

Results
METTL13 expression was upregulated in HNSCC and negatively correlated with clinical survival.
To investigate the expression pattern of METTL13 in HNSCC, we rstly analyzed its mRNA expression level by qPCR in 67 HNSCC and paired para-carcinoma tissues obtained from Hospital of Stomatology of Sun Yat-sen University. Notably, signi cant upregulation of METTL13 was observed (p < 0.001, Fig. 1A). Consistently, TCGA analysis data also showed that METTL13 was upregulated in the HNSCC samples ( Fig. 1B). Next, we examined the METTL13 expression at protein level by immunohistochemistry (IHC). Higher expression of METTL13 was observed in human HNSCC samples than paired normal tissues (Fig. 1C). To further explore the potential prognostic value in HNSCC, a total of 44 patients were segmented into METTL13 low and high expression groups based on the qPCR results. The data showed METTL13 signi cantly negatively correlated with overall survival of HNSCC patients (p < 0.05, Fig. 1D).
Collectively, we showed that the expression of METTL13 was upregulated in HNSCC, and it might play a key role in HNSCC development and progression.

Knockdown of METTL13 inhibited HNSCC cell malignant phenotypes in vitro.
Aberrant cellular activities often occur in cancer cells, such as continuous proliferative, evading growth inhibition, resisting cell apoptosis and active invasion and metastasis [29]. To explore the functional role of METTL13 in HNSCC cells, we rst selected eight HNSCC cell lines and compared their METTL13 expression by the Western blotting assay (Fig. 2A). The results revealed high METTL13 protein level in two HNSCC cancer cell lines (Cal27 and Cal33) and low METTL13 protein level in SCC1 cell line ( Fig. 2A). Hence, we chose Cal27, Cal33 and SCC1 cell lines for further functional study. We then generated stable METTL13 knockdown Cal27 and Cal33 cell lines using shRNA targeting METTL13 and con rmed the successful knockdown of METTL13 gene at mRNA and protein levels (Fig. 2B, C). Next, we performed clone formation assay and CCK-8 assay to explore the effect of METTL13 gene knockdown on the proliferation in HNSCC cell lines. The results showed that the depletion of METTL13 signi cantly not only inhibited the clone formation ability, but also inhibited cell growth rate in these two cell lines (Fig. 2D, E). Next, we utilized a ow cytometer to study the role of METTL13 in cell apoptosis. Our results demonstrated a higher apoptosis rate after ablation of METTL13 in HNSCC cells (Fig. 2F). Subsequently, wound healing assay results revealed that silencing of METTL13 signi cantly suppressed cell migration and prolonged the repair time (Fig. 2G). Transwell cell migration assay further proved knockdown of METTL13 attenuated the invasion ability of HNSCC cells (Fig. 2H). Taken together, these results demonstrated that METTL13 is critical for proliferation, apoptosis and invasion of HNSCC cell lines.

Mettl13 Downregulation Negatively Correlates With Hnscc Csc-like Properties
We also investigated whether METTL13 played a role in HNSCC cancer stem cells (CSCs)-like properties. First, we determined HNSCC cellular self-renewal capacity by conducting the sphere-forming assay. Downregulation of METTL13 greatly reduced the size and number of tumor spheres when compared with those formed by control cells in both Cal27 and Cal33 cell lines (Fig. 3A, B), indicating that their ability to self-renew was impaired. We then further assessed the CSC-like properties of METTL13 using assay to detect aldehyde dehydrogenase (ALDH) activity, which was regarded as a reliable marker for CSCs [30]. As showed in Fig. 3C-D, ALDH activity was signi cantly reduced in cells treated with sh-METTL13. Therefore, METTL13 played an essential part in maintaining HNSCC cells self-renewal ability.
Overexpression of METTL13 could promote HNSCC tumorigenicity.
In further explore the function of METTL13, a stable METTL13 overexpression SCC1 cell line was generated. And the up-regulation e ciency was veri ed by western blotting and qPCR assays (Fig. 4A, B).
The CCK-8 assay revealed overexpression of METTL13 remarkably increased the capacity of HNSCC cells to proliferate (Fig. 4C). Consistently, lower rate of apoptotic cells was captured in METTL13 over group by the ow cytometry (Fig. 4D). Next, we found that extra acquisition of METTL13 could accelerated cell migration of SCC1 in the transwell assay (Fig. 4E). Furthermore, METTL13 over group tended to form larger size of tumor spheres after having cultured for seven days (Fig. 4F). On the other hand, overexpression of METTL13 endowed SCC1 cells with higher ALDH activities (Fig. 4G). To sum up, these data show that METTL13 plays a potential oncogenic role by increasing or inducing the malignant proliferation, migration and CSC-like properties in HNSCC cells.

METTL13 regulated EMT signaling pathway by enhancing translation e ciency of Snail to facilitate HNSCC progress
We have demonstrated that METTL13 expressly promoted HNSCC proliferation and metastasis, but the speci c mechanism was unclear. Previous studies have reported that METTL13-mediated methylation of eukaryotic elongation factor 1A (eEF1A) promote oncogenesis via increased translation elongation and protein synthesis in cancer cells [19]. And indeed, our polysome pro le assay results revealed that decrease of METTL13 suppressed total protein translation in both Cal27 and Cal33 cell lines (Fig. 5A). In addition, the surface sensing of translation (SUnSET) monitoring method con rmed protein synthesis was inhibited after METTL13 depletion (Fig. 5B).
Previous study has indicated that METTL13 is involved in RAS-driven cancer development, we wonder whether depletion of METTL13 in HNSCC is also related to RAS signaling pathway [19]. Therefore, RNA sequencing was performed using the METTL13 control and knockdown groups in both Cal27 and Cal33 cell lines. Interestingly, results from Gene Set Enrichment Analysis (GSEA) indicated that EMT signaling pathway was signi cantly negatively correlated with the downregulation of METTL13 in HNSCC cell lines (Fig. 5C), suggesting that downregulation of protein synthesis might prone to EMT-related factors. To examine whether this is the case, a series of key proteins associated with EMT pathway were examined by western blotting (Fig. 5D, data not shown). Notably, as a key transcription factor in EMT, Snail was signi cantly downregulated in METTL13-de cient cells (Fig. 5D). Therefore, we hypothesized that METTL13 could regulate EMT signaling pathway by regulating the translation activity of Snail mRNA. To test that, we performed polysome pro le assay using sucrose gradient centrifugation to separate the ribosome and polysome segments. Based on these results, we used qPCR to detect the distribution of METTL13 affected mRNAs in the polysome segment. And the results showed Snail mRNA was signi cantly down-regulated in the METTL13 de cient groups in the polysome levels (Fig. 5E), indicating that the METTL13 is required for translation of Snail, which consistent with our previous conjecture. Altogether our results support that Snail is involved in METTL13-regulated EMT in HNSCC, and METTL13 regulated EMT signaling pathway by enhancing translation e ciency of Snail.

Decreased METTL13 expression weakened malignant phenotypes in vivo.
Xenograft studies were then conducted to investigate whether METTL13 could recapitulate the in vitro malignant phenotypes in vivo. Control and two METTL13 shRNA Cal33 cell lines were subcutaneous injected into nude mice. Mice were sacri ced after four weeks. In accordance with our in vitro results, de ciency of METTL13 in Cal33 cell dramatically suppressed tumor growth when compared with the control in vivo (Fig. 6A). And the tumor volume and tumor weight results further con rmed this effect (Fig. 6B, C). In order to get a more profound understanding of the function of METTL13 in vivo, we obtain the tumor sections. Next, hematoxylin and eosin (H&E) staining and METTL13 IHC results veri ed that METTL13 expression was decreased in sh-METTL13 tumor cells compared to the control (Fig. 6D, E). The staining of Ki67, a proliferation marker, showed alleviated cell growth after METTL13 gene knockdown, which was consistent with in vitro results (Fig. 6E). Consistently, IHC staining of nude mice showed the expression of Snail protein was reduced by the silencing of METTL13 (Fig. 6E).
Collectively, these data demonstrated that METTL13 is required for tumorigenic ability of HNSCC cells in vivo.

Discussion
Translation is one of the most energy consumption processes inside the cell and dysregulation of translation mechanisms is a common cause of neoplastic diseases [31,32]. Because in cancer cells where mRNA translation is usually highly active, it is necessary to increase the rate of protein synthesis [14]. And increasing evidences revealed translation elongation factor modi cation could cause promotion or inhibition of protein synthesis [33,34]. In our study, we demonstrated that METTL13, a regulator of translation elongation, showed higher expression levels in HNSCC tumors than adjacent normal tissues and might serve as a prognostic marker for overall survival in HNSCC patients.
METTL13 has been reported to be involved in cell apoptosis, stem cell biology and various cell signal and metabolic pathways [35]. However, whether it participates in tumorigenesis activity in HNSCC remains largely elusive. Here, we documented a detailed oncogenic role of METTL13 in HNSCC cell lines. We found that METTL13 is necessary and required for proliferation and EMT of HNSCC cancer cells both in vitro and in vivo. The translation level of METTL13-depleted HNSCC cells underwent signi cantly decrease compared with control cells. Intriguingly, GESA analysis of the transcriptome showed that genes down-regulated in the METTL13-depleted HNSCC are related to EMT process. Previously, METTL13 has been shown to promote cell proliferation and metastasis by elevating the levels of TCF3 and ZEB in hepatocellular carcinoma [36]. In addition, METTL13 can participate in development of Ras-dependent epithelial tumors of the pancreas and lung [19]. In HNSCC, we found depletion of METTL13 profoundly reduced the protein level of Snail, one of the key transcription factors of EMT. More speci cally, we demonstrated that METTL13 could enhance the translation e ciency of Snail mRNA to promote EMT and HNSCC progression. It is highly possible that the inhibition of the metastatic ability of cancer cells after deletion of METTL3 may be mitigated by Snail's overexpression. Previous studies revealed that EMT could be utilized by the tumor cells to perform metastatic cascades [7], and scienti c groups began to work on the its underlying mechanisms in tumor progression. In one study, EMT signaling pathway was blocked through modulation of the key protein translation in breast cancer [37]. Thereby, our study provided a more profound understanding of EMT in HNSCC.

Conclusion
In conclusion, our data revealed the tumor promoter function of METTL13 in HNSCC. METTL13 could regulate the progression of EMT by promoting Snail mRNA translation. These results suggested that METTL13 might function as a promising drug target. METTL13 could regulate the translation of downstream proteins by affecting the stability of 80 s ribosome structure. Although deeper researches are needed to uncover the mechanism of METTL13 carcinogenesis and as a drug target for HNSCC in the future, our present research has built the foundation for better therapeutic strategies.  Figure 1 The expression of METTL13 is increased in HNSCC and its high expression is correlated with poor    overexpression were evaluated by transwell migration assay, left; right, column chart. Scale bar, 100μm. E.

Page 22/24
Cell apoptosis rate was examined by ow cytometry using the Annexin V apoptosis detection kit. Ratio of early and late apoptotic cells was collected and presented in the column chart. F. Representative images and quanti cation results of sphere-forming assay for SCC1 cell lines. G. ALDH activity was detected by the ow cytometry to examine the effect of METTL13 overexpression on the cellular stemness.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Table2.pdf Table1.pdf