METTL18 functions as a Phenotypic Regulator in Src-Dependent Oncogenic Responses of HER2-Negative Breast Cancer

Abstract

Methyltransferase-like (METTL)18 has histidine methyltransferase activity on the RPL3 protein and is involved in ribosome biosynthesis and translation elongations. Several studies have reported that actin polymerization serves as a Src regulator, and HSP90 is involved in forming polymerized actin bundles. To understand the role of METTL18 in breast cancer and to demonstrate the importance of METTL18 in HER-2 negative breast cancer metastasis, we used biochemical, molecular biological, and immunological approaches in vitro (breast tumor cell lines), in vivo (tumor xenograft model), and in samples of human breast tumors. A gene expression comparison of 31 METTL series genes and 22 methyltransferases in breast cancer patients revealed that METTL18 is highly amplified in human HER2-negative breast cancer. In addition, elevated levels of METTL18 expression in patients with HER2-negative breast cancer are associated with poor prognosis. Loss of METTL18 significantly reduced the metastatic responses of breast tumor cells in vitro and in vivo. Mechanistically, METTL18 indirectly regulates the phosphorylation of the proto-oncogene tyrosine-protein kinase Src and its downstream molecules in MDA-MB-231 cells via METTL18-mediated RPL3 methylation, which is also involved in determining HSP90 integrity and protein levels. In confocal microscopy and F/G-actin assays, METTL18 was found to induce actin polymerization via HSP90. Molecular events involving METTL18, RPL3, HSP90, and actin polymerization yielded Src phosphorylated at both tyrosine 419 and tyrosine 530 with kinase activity and oncogenic functions. Therefore, it is suggested that the METTL18-HSP90-Actin-Src regulatory axis plays critical oncogenic roles in the metastatic responses of HER2-negative breast cancer and could be a promising therapeutic target.

Keywords: HSP90; METTL18; RPL3 methylation; Src kinase; actin; breast cancer; histidine methyltransferase; invasion; metastasis; migration.

Figure 1

Clinical relevance of METTL18 in human HER2-negative breast cancer. (A) Heatmap of expression profiles of METTL18 and 30 METTL series genes. Gene expression was analyzed in normal-like, HER2-positive, and HER2-negative breast cancer samples in the TCGA database (n=1080). (B) Gene expression profile of METTL18 in breast cancer tissues and normal tissues. We used the GEPIA2 dataset and classified samples into four groups: normal tissues from HER2-negative patients (n=291), tumor tissues from HER2-negative patients (n=550), normal tissue from HER2-positive patients (n=291), and tumor tissues from HER2-positive patients (n=260). (C) Protein expression profile of METTL18 in breast tumor tissue from Samsung's cohort (n=150). Patients were divided into HER2-positive and HER2-negative. (D) Overall survival time (months) of Samsung's cohort breast cancer patients with low or high levels of METTL18 (median cutoff). (E) Kaplan-Meier curve showing the survival probability of HER2-negative breast cancer patients with high or low expression of METTL18 (median cutoff). (F) Gene expression of METTL18 in HER2-negative breast cancer patients with or without metastasis. Database from the Metastatic Breast Cancer Project [Provisional, Feb 2020] was used. (G) DMFS of HER2-negative breast cancer patients with low or high expression of METTL18 (median cutoff) (GSE25066). * P < 0.05, ** P < 0.01, ns: not significant.

Figure 2

Effect of METTL18 on metastatic responses in breast cancer cells in vitro and in vivo. (A) Protein expression of METTL18 in breast cancer cell lines. Protein levels of HER2, METTL18, and β-actin were identified by Western blotting. (B) Migration capacity of METTL18-knockdown (left panel) and METTL18-overexpressing (right panel) MDA-MB-231 cells. Knockdown and overexpression of METTL18 were confirmed by western blotting (Fig. S2A). The Tag3 vector, utilized for overexpressing METTL18, served as the vehicle control. ImageJ measured the migration range. (C) Invasive capacity of METTL18 knockdown (left panel) and METTL18-overexpressing (right panel) MDA-MB-231 cells. Knockdown and overexpression of METTL18 were confirmed by western blotting (Fig. S2A). The Tag3 was used as a control of METTL18 overexpression. (D) Enzyme activities of MMP-2 and MMP-9 in METTL18-knockdown MDA-MB-231 cells. (E) Metastasis of tumors in xenograft mice intravenously injected with shScramble- or shMETTL18-transfected MDA-MB-231 cells. The white arrows indicate metastatic tumors, visualized via ¹⁸F-FDG-PET/CT scans. The blot shown in (A) and (D) is a representative image of three independent Western blot experiments. * P < 0.05; ** P < 0.01.

Figure 3

Effect of METTL18 on the Src kinase signaling pathway. (A) Significantly altered pathways in the RNA-seq analysis of shMETTL18 #2795-expressing MDA-MB-231 cells (cutoff for p-value < 0.01). (B) Heatmap of phospho-protein array data. Samples were prepared from METTL18-overexpressing MDA-MB-231 cells. (C) Immunoblotting for the phospho- and total forms of tyrosine kinases (Src, STAT3, and p85) in Myc-METTL18-overexpressing or shMETTL18-expressing MDA-MB-231 cells. Transfection efficacy of the Myc-METTL18 construct and shMETTL18 was verified by immunoblotting with anti-Myc and anti-METTL18, respectively. β-actin was used as the loading control. (D) Western blotting results for the phospho- and total protein levels of Src in siPIMT (20 nM)-, siEEF2KMT (20 nM)-, and siPRMT1 (20 nM)-transfected MDA-MB-231 cells after 48 hours. (E) Scatterplot showing the correlation between p-Src and METTL18 in breast cancer patients from Samsung's cohort. Source data for the scatterplot were obtained from immunoblotting (Fig. S5). P-values and the correlation coefficient (R) were calculated using GraphPad Prism. (F) Kaplan-Meier curve showing the survival probability of HER2-negative breast cancer patients with high or low expression of p-Src Y419 (best cutoff). (G) Invasive capacity of MDA-MB-231 cells transfected with Myc, Myc-METTL18, shScramble, or shSrc. The transfection efficacy of the plasmids and shRNA was verified by Western blotting with anti-Myc and anti-Src. The blots shown in (C), (D), and (G) are representative images of three independent Western blot experiments. ns: not significant; * P < 0.05; ** P < 0.01; ## P < 0.01.

Figure 4

Involvement of RPL3 in the METTL18-Src regulatory mechanism. (A) Immunoprecipitation analysis of METTL18 and Src. The assay was performed with Myc, Myc-METTL18, HA, and HA-Src-overexpressing MDA-MB-231 cells. METTL18 was immunoprecipitated with anti-Myc followed by immunoblotting with anti-HA to identify Src. (B) Western blotting for the phospho- and total protein levels of Src in METTL18 WT- and METTL18 histidine 154 to lysine mutant (H154K)-transfected MDA-MB-231 cells and MCF-7. Transfection efficiency was tested by Western blotting with anti-Myc, and β-actin was used as the loading control. (C,D) The invasive capacity (C) and migrative capacity (D) of MDA-MB-231 cells and MCF-7 transfected with Myc, Myc-METTL18, or Myc-METTL18 H154K. The transfection efficacy of the plasmids and shRNA was verified by Western blotting with anti-Myc and anti-Src. (E) Methyltransferase assay with METTL18 and RPL3. For the in vitro methyltransferase assay, immunoprecipitated Myc and Myc-METTL18 prepared with lysates of MDA-MB-231 cells transfected with Myc-METTL18 or empty vector (Myc) were individually used as enzyme source. In addition, immunoprecipitated HA-RPL3 prepared with lysate of MDA-MB-231 cells transfected with HA-RPL3 or empty vector (HA) was used as a substrate. The enzyme and substrate sources were incubated with SAM (20 μM). Methyltransferase activity was determined via bioluminescence. (F) Histidine methylation level of RPL3 was evaluated with immunoprecipitated HA-RPL3 prepared with lysate of METTL18-knockdowned MDA-MB-231 cells transfected with HA-RPL3 or empty vector (HA) by immunoblotting with pan-anti-methylhistidine antibody. (G) Immunoblotting for the phospho- and total protein levels of Src in siRPL3- and METTL18-transfected MDA-MB-231 cells after 48 hours. (H) The invasion rate of MDA-MB-231 cells transfected with siRPL3 and METTL18. (I) Immunoblotting for the phospho- and total protein levels of Src in RPL3 WT and RPL3 histidine 245 to alanine mutant (H245A)-transfected MDA-MB-231 cells and MCF-7 cells. Transfection efficiency was tested by Western blotting with anti-HA, and β-actin was used as the loading control. The blots shown in (A), (B), (F), (G), and (I) are representative images of three independent Western blot experiments. ns: not significant; * P < 0.05; ** P < 0.01; ## P < 0.01.

Figure 5

Involvement of HSP90 and actin in the METTL18-Src regulatory mechanism. (A) Silver staining image showing proteins immunoprecipitated with anti-Myc. The immunoprecipitation analysis was performed with whole lysates from Myc- or Myc-METTL18-overexpressing MDA-MB-231 cells, and each protein was identified by a further MS analysis. (B) Immunoblotting for p-Src and Src in MDA-MB-231 cells. Myc- or Myc-METTL18-overexpressing MDA-MB-231 cells were treated with siScramble, siHSP90AA1, and HEP70-I (VER15508). β-actin or β-tubulin was used as the loading control. The efficacy of the siRNAs was verified by immunoblotting with an antibody against each target protein. (C,D) Immunoblotting for HSP90 or JAK2 levels in the siScramble-, siMETTL18-, or siRPL3-transfected MDA-MB-231 cells. The transfection efficacy of the siRNA was examined by immunoblotting with anti-METTL18 and anti-RPL3. (E) Immunoblotting of HSP90, phospho-Src (Y419), phospho-Src (Y530), and Src in siScramble- or siMETTL18-transfected (48 hours) and HA-RPL3-WT-transfected (24 hours) MDA-MB-231 cells. The transfection efficacy of the siRNA and RPL3 was verified by Western blotting with anti-METTL18 and anti-HA. β-actin was used as the loading control. (F) Immunoblotting of HSP90, phospho-Src (Y419), phospho-Src (Y530), and Src in siScramble- or siMETTL18-transfected (48 hours) and HSP90-transfected (24 hours) MDA-MB-231 cells. The transfection efficacy of the siRNA and HSP90 was verified by Western blotting with anti-METTL18 and anti-HA. β-actin was used as the loading control. (G,I) Confocal microscopy images of Src (green) and phospho-Src (p-Y419) (red) in MDA-MB-231 cells transfected with Myc, Myc-METTL18, siScramble, siHSP90 (G), or siActin (I). Nuclei were stained with DAPI (blue). The transfection efficacy of each plasmid and siRNA was verified by immunoblotting with anti-Myc, anti-HSP90, and anti-actin. (H) Confocal microscopy images of polymerized actin (red) in MDA-MB-231 cells transfected with siScramble, siMETTL18, siHSP90, or siActin for 48 hours. Knockdown of METTL18, Actin, and HSP90 was confirmed by western blotting (Fig. S11A). Carl Zeiss Zen blue edition calculated the relative intensity of polymerized actin. The blots shown in (B), (C), (D), (E), (F), (G) and (I) are representative images of three independent Western blot experiments. ns: not significant; * P < 0.05; ** P < 0.01.

Figure 6

Importance of biphosphorylated Src in METTL18-mediated Src activation. (A,B) Immunoblotting for phospho-Src (Y419), phospho-Src (Y530), Src, and β-actin in siScramble-, siMETTL18- (A), and siRPL3- (B) transfected MDA-MB-231 cells. The transfection efficacy of the siRNA was tested by Western blotting with anti-METTL18 or anti-RPL3. (C) Immunoblotting of phospho-Src (Y419), phospho-Src (Y530), and Src in MDA-MB-231 cells treated with cytochalasin B (CytoB) for 6 hours. GAPDH serves as an internal reference. Additionally, the efficacy of CytoB was verified by observing the suppressed levels of filamentous actin (F-actin). (D) Confocal microscopy images showing DAPI and proximity ligation assay (PLA) signals in control and METTL18-expressing MDA-MB-231 cells. The cell nuclei were stained with DAPI (blue). The PLA was performed with primary antibodies against phospho-Src Y416 (rabbit polyclonal) and phospho-Src Y530 (mouse monoclonal), and the PLA signals are shown as green dots. The anti-p-Src Y416 was employed to detect human p-Src Y419. (E-G) Immunoprecipitation analysis to determine the presence of biphosphorylated Src kinase. We used lysates from MDA-MB-231 cells treated with Myc (E), Myc-METTL18 (E), siScramble (F), siHSP90 (F), or cytoB (G). Immunoprecipitation was performed with anti-phospho-Src (Y530), and each immunoprecipitation lysate was identified by immunoblotting with antibodies against phospho-Src (Y419) and phospho-Src (Y530). Heavy chain (H.C.) and β-actin were used as loading controls. Transfection and siRNA efficacy was identified by Western blotting with anti-Myc (e) or anti-HSP90 (f). (H,I) Kinase assay with Src WT and Src mutants. For the in vitro kinase assay (H), HA, HA-Src WT, HA-Src Y530F, HA-Src Y530D, and HA-Src Y419D/530D mutants were individually used as enzymes. In addition, Myc-PI3K was used as a substrate. The enzyme and substrate sources were prepared by immunoprecipitation and incubated with ATP (200 μM). Kinase activity was determined via immunoblotting against p-PI3K. Using a Src kinase enzyme system purchased from Promega (I), the assay was performed in HEK293 cells with immunoprecipitated HA, HA-Src WT, HA-Src Y530F, HA-Src Y530D, or HA-Src Y419D/530D. (J,K) The invasive capacity (J) and migrative capacity (K) of MDA-MB-231 cells transfected with HA, HA-Src WT, HA-Src Y530F, or HA-Src Y419/530D. Western blotting with anti-HA verified the transfection efficacy of the plasmids and shRNA. The blot shown in (A), (B), (C), (E), (F), (G), (H), and (J) is a representative image of three independent Western blot experiments. ns: not significant; * P < 0.05; # P < 0.05; ** P < 0.01; ## P < 0.01.

Figure 7

The alternative Src regulatory pathway mediated by the METTL18-RPL3-HSP90-actin axis in HER2-negative breast cancers. HER2-negative breast cancer expresses more METTL18 proteins than normal tissues and other breast cancer subtypes. METTL18 is responsible for RPL3 methylation at the histidine 245 residue, and that modification prompts the maintenance of HSP90 protein integrity. The increased HSP90 protein levels lead to actin polymerization. Actin filaments interact with the SH2 domain in the Src kinase, generating biphosphorylated Src with kinase activity. Ultimately, the biphosphorylated Src kinase increases the metastatic capacity of HER2-negative breast cancer, resulting in a poor prognosis.