METTL3-dependent MALAT1 delocalization drives c-Myc induction in thymic epithelial tumors

Background Thymic epithelial tumors (TETs) are rare neoplasms, originating from epithelial thymic cells. The oncogenic potential of these rare neoplasms is still largely undefined, and a deeper molecular characterization could result in a relevant advance in their management, greatly improving diagnosis, prognosis and treatment choice. Deregulation of N6-methyladenosine (m6A) RNA modification, catalyzed by the METTL3/METTL14 methyltransferase complex, is emerging as a relevant event in cell differentiation and carcinogenesis. Various studies have reported that altered expression of METTL3 is associated with an aggressive malignant phenotype and favors migration and invasiveness, but its role in Thymic Tumors remains unknown. Results In this study, we characterized that METTL3 contributes to Thymic Epithelial Tumor phenotype. We evidenced that METTL3 is overexpressed in tumor tissue compared to normal counterpart. Silencing of METTL3 expression in thymic carcinoma cells results in reduced cell proliferation and overall translation rate. Of note, METTL3 is responsible for the induction of c-MYC expression in TET cells. Specifically, high expression of c-MYC protein is enabled by lncRNA MALAT1, which is methylated and delocalized by METTL3. Interestingly, blocking of c-MYC by using JQ1 inhibitor cooperates with METTL3 depletion in the inhibition of proliferation and induction of cell death. Conclusion This study highlighted METTL3 as a tumor promoter in Thymic tumors and c-MYC as a promising target to be exploited for the treatment of TET. Supplementary Information The online version contains supplementary material available at 10.1186/s13148-021-01159-6.


Background
Thymic epithelial tumors (TETs) are very rare neoplasms comprising thymoma (THY) and thymic carcinoma (TC) that establish a unique group of cancers arising from the epithelial cell of the Thymus [1]. On the basis of the current World Health Organization (WHO) histopathological classification, thymomas have been classified into types A, AB, B1, B2, and B3 and rare other subtypes [2,3]. Iaiza et al. Clin Epigenet (2021) 13:173 The malignant potential of thymomas, ranging from low to very aggressive locally, is in relation to the stage of these tumors [4,5]. The metastatic potential and local relapse and/or pleural dissemination that characterize both the advanced stage of B2 and B3 types of thymomas and thymic carcinomas (TC) is similar to the other carcinoma features observable outside of mediastinum [6].
Complete surgical resection is the gold standard when feasible; chemotherapy like the combination of cisplatin, doxorubicin and cyclophosphamide (PAC) and/or radiotherapy can be administered as an induction at advanced stages or prior to surgery as adjuvant treatment in case of advanced pathological stages [7,8].
The management of TET is a paradigm of synergistic strategy to guarantee an advance in clinical and translational research in these tumors [9].
Significant advances have recently been made in the characterization of the molecular landscape at the basis of TET underlined by The Cancer Genome Atlas (TGCA) [10] data publication and the identification of the noncoding RNA contribution to TET-associated phenotype [11][12][13].
N6-methyladenosine (m 6 A), one of the most abundant internal modification of mRNA and noncoding RNA (ncRNA), is an important modification involved in RNA splicing [14], translation [15,16], stability [17] and export [18], leading to modulation of several physiologic and pathologic processes [19][20][21][22]. m 6 A modification is catalyzed in the nucleus by a multicomponent complex formed by two methyltransferases METTL3 (MT-A70) and METTL14 (or KIAA1627) [23]. The first one is the catalytic subunit of the complex, while METTL14 has a degenerate active site and plays non-catalytic roles in maintaining complex integrity and substrate RNA binding [24,25]. Other factors are necessary in vivo to catalyze m 6 A modification, like WTAP, required for complex localization into nuclear speckles [26], VIRMA (KIAA1429) and RBM15 for recruitment of the methyltransferase complex on the target site of the pre-mRNA, ZC3H13 required for RBM15 and WTAP association [27].
This RNA modification is dynamic and reversible; therefore, it can be erased by FTO and ALKBH5 demethylases [23], both belonging to the AlkB deoxygenase family and localized in the nuclear speckles [28,29]. The first readers to be identified are the members belonging to YT521-B homology (YTH) domain family (YTHDF1, YTHDF2, YTHDF3, YTHDC1 and YTHDC2), which contain an aromatic cage for specifically accommodating the m 6 A [30]. Furthermore, m 6 A can induce a structural modification in transcripts, termed 'the m 6 A-switch' , favoring or inhibiting the interaction with hnRBP [31,32].
The recent rediscovery of m 6 A is due to its crucial role in the growth and progression of many types of cancer, where different components have an alternated expression, in particular METTL3. Some roles described for the physiological functions of the cell are exploited by cancer cells to their advantage to allow the self-renewal of the cancer stem cells, to increase translation of oncogenes such as EGFR, MYC, TAZ, MAPKAPK2 and to use shelter systems to allow cell survival [33][34][35][36][37]. Moreover, an interplay between m 6 A modification and non-coding RNAs, in particular microRNAs and lncRNAs has recently emerged [38]. MALAT1, for example, a lncRNA frequently upregulated in cancer, contains several m 6 A modifications that impact on its functional activity [39,40].
In this work, we assessed that METTL3 is upregulated in thymic tumors and we investigated the biological significance of the increased levels of METTL3 in thymic carcinoma cells. Moreover, we evaluated whether METTL3 and its downstream target c-MYC may represent powerful targets to improve the treatment of thymic tumors.

METTL3 is upregulated in TETs
To explore the possible contribution of components of the protein machinery enabling N6-methyladenosine (m 6 A) methylation of RNAs to thymic cancer, we first took advantage of the thymoma GEPIA dataset to evaluate whether any modulation of these components was present between cancer and normal tissues. As shown in Fig. 1A, we observed that WTAP, METTL3 and METTL14 were significantly upregulated in cancer vs. normal tissues, with METTL3 showing the most significant deregulation. Further analysis of the TCGA dataset showed that median expression of METTL3 mRNA was higher in B1, B2, B3 and C compared to A and AB histotypes (Fig. 1B). To complement the analyses at mRNA level, we analyzed METTL3 protein expression by immunohistochemistry on thymic tumor tissues (N = 22). Interestingly, analysis of METTL3 protein staining intensity in epithelial cancer cells revealed an increase of intensity from types A/AB (score of intensity = 1) to types B1-B3 (score of intensity = 2), and reached a maximum of intensity (score = 3) in thymic carcinoma cases (C) (Fig. 1C).

METTL3 impacts on cell proliferation and translation rate
To investigate the biological output of the high expression level of METTL3 in TET, we depleted METTL3 by siRNA transfection in the thymic carcinoma cell line TC1889, to evaluate the effects on proliferation and colony forming ability. As shown in Fig. 2A-B, Additional file 1, 2: S1A and S2A, a marked decrease of proliferative potential in cells depleted of METTL3 compared to control cells was observed. Accordingly, TC1889 cells depleted of METTL3 expression for 72 h also showed a significant decrease of colony forming ability (Fig. 2C). We didn't observe any impact of METTL3 depletion on migration capacity (Additional file 1: Fig. S1B), differently from what previously reported in other cell types [41][42][43].
As METTL3-dependent m 6 A modification may influence the translation rate of targeted mRNAs [16,44], we wondered whether the observed decreased proliferation rate of METTL3-depleted cells could be attributable to reduced translation potential of cells. To address this, we evaluated the effect of modulation of METTL3 on global translation efficiency in TC1889 cells. Specifically, 72 h METTL3-depleted (siMETTL3) and control (siC) cells were challenged with puromycin, a tyrosyl-tRNA mimic that blocks translation by labeling and releasing elongating polypeptide chains from translating ribosomes, for 10, 20, 40 or 60 min. We measured the levels of newly synthesized polypeptides using an anti-puromycin antibody. We observed a strong decrease of puromycin incorporation upon METTL3 silencing compared to control cells, indicating a global decrease of protein synthesis ( Fig. 3A and Additional file 3: Fig. S3A). According to this observation, polyribosome (polysome) fractionation analysis evidenced a drop of polysome fraction in samples depleted of METTL3 for 48 h and 72 h and an increase of 80S ribosome fraction compared to control samples ( Fig. 3B and Additional file 3: Fig. S3B).
Analysis of the protein levels of crucial regulators of protein synthesis, such as p70 ribosomal protein S6 kinase 1 (S6K1), its substrate ribosomal protein S6 (S6), and eukaryotic translation initiation factor 4E-binding protein 1 (4EBP1), revealed that METTL3 depletion causes a marked and significant decrease of the phosphorylated forms of these upstream regulators, compared to control cells (Fig. 3C), indicating a role for METTL3 in the induction of translation in TET cells. Interestingly, METTL3 depletion also caused a decrease of the basal level of expression of translation regulators S6K1 and 4EBP1, while no changes in S6 levels were detected (Fig. 3D).

METTL3 regulates c-MYC mRNA and protein levels
It has been extensively reported that c-MYC, a well characterized oncogene involved in the induction of cell proliferation and growth, is targeted by METTL3 in numerous malignancies, including acute myeloid leukemia (AML) [33,45,46]. We then explored whether the METTL3-dependent increase of cell proliferation and translation rate in TET could be attributed to the induction of c-MYC.
Bioinformatic analysis of TGCA data showed a positive correlation between METTL3 and c-MYC mRNA levels in Thymoma and Thymic carcinoma samples ( Altogether, these results indicate that METTL3 controls c-MYC transcript and protein level also in TET.

METTL3-dependent m 6 A modification of MALAT1 impinges on MYC protein synthesis
We next explored potential mediators responsible for METTL3-dependent increased translation of c-MYC. Since it has been reported that long non-coding RNA MALAT1 may be subjected to m 6 A modification and may control translation in various malignancies, we investigated whether its activity was regulated by METTL3. To this end, we first evaluated by RNA FISH the subcellular localization of MALAT1 in control and METTL3depleted TC1889 cells. It has been indeed reported that MALAT1 may be localized in nuclear speckles and/or it may be diffused in the nucleoplasm [47,48], and that different localization is associated with specific functions.  Fig. S4A) led to a significant reduction of both S6K1 protein level and its phosphorylated form, while no changes in S6 and 4EBP1 were present. Interestingly, depletion of MALAT1 also caused a marked reduction of c-MYC protein (Fig. 5E). We have also evaluated, in the same experimental conditions as above, the expression level of VEGFA isoforms, which we have previously reported to be affected by MALAT1 subnuclear localization in breast cancer cells [48]. Expression of both VEGF165 and VEGF121 is decreased following MALAT1 or METTL3 depletion as shown in Additional file 4: Fig.  S4B. Next we evaluated the impact of single or combined METTL3 and MALAT1 silencing on c-MYC expression. To address this, we considered the 24 h time point of METTL3 silencing, when c-MYC protein is not yet strongly decreased (as already shown in Fig. 4C). Combined METTL3 and MALAT1 silencing led to a more marked reduction of c-MYC protein expression compared to single silencing (Additional file 4: Fig. S4C). Interestingly, silencing of METTL3 decreases c-MYC mRNA and protein, while MALAT1 silencing impinges only on c-MYC protein level (Fig. 5F).

Silencing of METTL3 combined with cisplatin or c-MYC inhibitor induces cell death in TET cells
Surgery, preceded and/or followed by chemotherapy and radiotherapy, is currently the gold standard for the treatment of thymic epithelial tumors.
To investigate whether METTL3 upregulation in TET impinges on the response to chemotherapy, we evaluated whether TC1889 cells silenced for METTL3 showed different sensibility to cisplatin, compared to control cells. Specifically, we monitored cell death and proliferation in cells transfected with control or METTL3 siRNAs and subsequently treated with cisplatin (CDDP, 7.5 μg/mL). We observed that CDDP treatment or METTL3 depletion significantly induced cell death and that the combination of CDDP and METTL3 silencing determined an additive effect, compared to individual treatments, leading to reduced proliferation (Fig. 6A, left) and increased cell death (Fig. 6A, right). This last effect was also confirmed by the analysis of the ratio between Cleaved-PARP and PARP (Additional file 5: Fig. S5B). Analysis of c-MYC protein level in the same experimental condition is shown in Fig. 6C (left) and Additional file 5: Fig. S5A.
On the basis of METTL3-dependent c-MYC protein induction, we evaluated the effect of the combination of METTL3 depletion and c-MYC inhibition on cell death and proliferation. c-MYC was inhibited using the small molecule JQ1, a selective bromodomain inhibitor able to block the c-MYC inducer BRD4 [49][50][51][52]. 24 h after transfection with siRNAs siC or siMETTL3, we treated cells with JQ1 (3 μM) for additional 48 h. We observed reduced proliferation and increased cell death in the combined compared to individual treatment ( Fig. 6B and Additional file 5: Fig. S5B). Analysis of c-MYC protein level in the same experimental condition is shown in Fig. 6C (right) and Additional file 5: Fig. S5A.
This result suggests that a block of c-MYC cooperates with METTL3 depletion, impinging on cell viability and proliferation, and highlighting this axis as a promising target to be exploited for the treatment of TET.

Discussion
This study focused on the study of a specific enzyme, METTL3, and its role in Thyimic epithelial tumors (TETs). METTL3, the main component of the m 6 A methyltransferase complex, plays a crucial role in many biological processes especially in tumorigenesis, either dependently on or independently from its m 6 A RNA methyltransferase activity [53,54].
As described in a variety of studies [55,56], METTL3 is expressed at higher level in tumor tissues compared to normal and we demonstrate here that it is up-regulated also in TETs tissues. As the role of METTL3 in Thymoma and Thymic carcinoma is currently uncharacterized, we first analyzed the effects of METTL3 silencing on the thymic tumor phenotype, using TC1889 cells obtained from a Thymic carcinoma, highlighting the requirement of METTL3 expression for proliferation and high translation rate. At molecular level, METTL3 is responsible for the high expression and phosphorylation of crucial proteins which control  in (B, B). See also Additional file 5: Fig. S5 for quantifications. *p ≤ 0.05; **p ≤ 0.005; ***p ≤ 0.0005; ****p ≤ 0.00005; p-values have been calculated by One-Way Anova translation, such as the mTOR-dependent S6K1, S6 and 4EBP1. Specifically, silencing of METTL3 decreases the basal and phosphorylated forms of S6K while reducing only the phosphorylated form of S6; this suggests that reduction of P-S6 is a consequence of S6K modulation by METTL3. It has been previously proposed that METTL3 activity is fundamental for normal cell survival; an important limitation of this study is represented by the lack of normal thymic cells analysis upon METTL3 depletion. Further studies focused on normal thymic epithelial cells will clarify whether or not the identified networks are specific to TET cells.
One of the most well-known oncogenes and key regulator of cell proliferation is c-MYC. Several reports have demonstrated that abnormal expression of c-MYC may be related to increased m 6 A modification of its mRNA in several tumors as for example in thyroid cancer [57], or in oral squamous cell carcinoma (OSCC) [46]. Consistent with these evidences, we validated c-MYC as target of METTL3-dependent methylation also in Thymic carcinoma and showed the requirement of METTL3 for an efficient c-MYC expression in TC1889 cells, suggesting that METTL3 might enhance cell proliferation by promoting c-MYC translation. Accordingly, a positive correlation between METTL3 and c-MYC expression was observed by analyzing the TCGA TET dataset. Increased c-MYC expression had been already reported in TET. Specifically, the TCGA study revealed an upregulation of the oncogenes MYC/ MAX in AB-, B-, and C-like associated molecular signatures compared to A-like subtype, in line with their known increased clinical aggressiveness in TETs [10].
Recently, different studies have reported that m 6 A modifications can also regulate the biogenesis and function of noncoding RNAs, such as miRNAs, lncR-NAs, and circRNAs [38]. In this study we found that MALAT1 is an additional target of METTL3-dependent methylation, and that this modification is able to influence the subnuclear localization of this lncRNA, leading to a diffused nucleoplasmic staining; on the contrary, MALAT1 shows a clear-cut localization in nuclear speckles in METTL3-depleted cells. Interestingly, similarly to METTL3-depleted cells, MALAT1depleted TC1889 showed decreased S6K1 protein level (including its phosphorylated form). However, siMALAT1 causes a less marked decrease of S6K and P-S6K, compared to siMETTL3, suggesting that this could be not sufficient to drive the S6 dephosphorylation observed in siMETTL3. It's also worth underlining that MALAT1 depletion by siRNA transfection isn't exactly equivalent to METTL3 depletion, which causes a relocalization of MALAT1 in nuclear speckles. This could also explain the incomplete overlap of results in the two interference conditions. Similarly, to METTL3-depleted cells, MALAT1depleted TC1889 also showed decreased c-MYC protein, suggesting that METTL3 could regulate c-MYC translation also through the delocalization of MALAT1. However, silencing of MALAT1 didn't impact on the m 6 A modification of c-MYC mRNA (data not shown), suggesting that it is not required for METTL3 enzymatic activity.
Previously, we have shown that an alternative oncogenic stimulus (e.g. the presence of mutated p53 protein and overexpressed ID4) causes a similar delocalization of MALAT1, in breast cancer cells, with consequent altered splicing [48]. The possibility that m 6 A modification could mediate also the effects previously observed in breast cancer cells is certainly an intriguing possibility that merits to be verified in future studies. Interestingly, it has been also recently reported that m 6 A modification of MALAT1 is critical for metastatic ability of cancer cells, reshaping the nuclear speckles [58].
Although the gold standard treatment for thymic epithelial tumors remains surgery, the use of several innovative agents for TET treatment is under investigation [59]. For this reason, we analyzed and established a sensitization effect of thymic carcinoma cells to a chemotherapeutic agent, like cisplatin, upon METTL3 silencing. As c-MYC overexpression has been extensively reported as a feature conferring chemoresistance to cancer cells, we can hypothesize that the reduced c-MYC protein levels observed after METTL3 silencing could represent the determinant allowing sensitization to cisplatin treatment. Of note, METTL3 depletion similarly sensitizes cells to the treatment with JQ1, a BRD4 inhibitor impairing c-MYC expression.

Conclusions
In conclusion, our study identified the oncogenic role of METTL3 also in Thymic Ephitelial Tumors, which improves cells proliferation trough the control of c-MYC expression. Moreover, our results highlight METTL3 as a powerful target whose depletion improves the response to different anticancer treatments and open to the possibility of exploiting this protein to improve the treatment of TETs.

Clonogenic assays
Cells were detached 72 h after METTL3 silencing and seeded into 6-well dishes at a concentration of 1 × 10 3 cells/well. Medium was changed every four days. 10-14 days after plating, cells were washed with PBS for three times and were stained using crystal violet dye. Cells were counted using ImageJ software.

RNA extraction and real-time qRT-PCR analysis
Total RNA was extracted using the TRIzol RNA Isolation System (Invitrogen) according to manufacturer instructions. Reverse transcription to cDNA was performed with the High-Capacity RNA-to-cDNA Kit (Applied Biosystems) and cDNA was amplified using the SYBR ™ Green PCR Master Mix (Thermo Fisher Scientific, Waltham, MA USA) on an ABI PRISM 7500 Sequence Detection System (Applied Biosystems). Relative expression levels of targets were determined with the comparative 2 −ΔΔCt method, using H3 as control. All reactions were performed in duplicate. The following oligo sequences were used:
TC1889 cells were treated with puromycin (#ant-pr-1, InvivoGen) at 10ug/ml after 72 h of transfection with siC or siMETTL3. Incorporated puromycin protein level was then analyzed 10, 20, 40 and 60 min after treatment by Western Blot analysis.

m 6 A immunoprecipitation
Cells were seeded and transfected with METTL3 and Control siRNA. After 72 h, total RNA was extracted and fragmented into ~ 100 nt long fragments in Fragmentation Buffer (100 mM Tris-HCl and 100 mM ZnCl 2 ) for 5′ at 94 °C. A fraction of fragmented RNA was collected as input control and the rest was incubated with m 6 A-specific antibody (#ab151230, Abcam) or with control rabbit IgG (Millipore) antibody for 2 h at 4 °C on rotator. Immunoprecipitated RNA was eluted and resuspended in RNasefree water. RT-qPCR on immunoprecipitated RNA was performed with primers MYC_FW and MYC_REV.

Statistical analysis
The statistical analyses were performed using GraphPad Prism (GraphPad Software, San Diego, California, USA). Statistical analysis to determine significance was performed using Student's t tests or One-Way Anova test. Differences were considered statistically significant at the p < 0.05 level.