Regulation of RNA methylation by therapy treatment, promotes tumor survival

Abstract

Our data previously revealed that chemosurviving cancer cells translate specific genes. Here, we find that the m6A-RNA-methyltransferase, METTL3, increases transiently in chemotherapy-treated breast cancer and leukemic cells in vitro and in vivo. Consistently, m6A increases on RNA from chemo-treated cells, and is needed for chemosurvival. This is regulated by eIF2α phosphorylation and mTOR inhibition upon therapy treatment. METTL3 mRNA purification reveals that eIF3 promotes METTL3 translation that is reduced by mutating a 5'UTR m6A-motif or depleting METTL3. METTL3 increase is transient after therapy treatment, as metabolic enzymes that control methylation and thus m6A levels on METTL3 RNA, are altered over time after therapy. Increased METTL3 reduces proliferation and anti-viral immune response genes, and enhances invasion genes, which promote tumor survival. Consistently, overriding phospho-eIF2α prevents METTL3 elevation, and reduces chemosurvival and immune-cell migration. These data reveal that therapy-induced stress signals transiently upregulate METTL3 translation, to alter gene expression for tumor survival.

Keywords: RNA methylation; chemosurvival; integrated stress response; mTOR inhibition; non-canonical translation.

Fig 1.. METTL3 and m6A modification on RNA are increased with chemotherapy doxorubicin treatment in breast cancer cells, and are needed for chemosurvival.

A. Western blot analysis of METTL3 (identified from Table S1), and canonical translation marker: phospho-eIF2α and eIF2α, in MCF7 cells over time of treatment with doxorubicin (250nM) chemotherapy. B. Dot blots of rRNA and tRNA depleted cellular RNA from 500nM doxorubicin and control treated cells versus cells depleted of METTL3 with METTL3 shRNA, probed with m6A antibody. Western blot of samples with control shRNA or shMETTL3 shRNA. C. 2D-TLC of rRNA and tRNA depleted cellular RNA from 500nM doxorubicin treated and untreated MCF7 cells; m6A and A nucleosides are marked, and quantitated. D. Doxorubicin chemosurvival in MCF7 cells expressing i. control shRNA or METTL3 shRNA (500nM doxorubicin for 5 days) ii. wildtype METTL3 compared to a control GFP vector (500nM doxorubicin for 3 days), and iii. m6A catalytically defective METTL3 (CD-MUT) compared to wild type METTL3 (500nM doxorubicin for 3 days). E. Western blot show METTL3 and eIF2α phosphorylation increase in hormone positive breast cancer patient samples grown as 3D tumorspheres and treated with doxorubicin (250nM Doxorubicin for 18hrs), and in F. patient derived triple negative breast cancer xenografts (pdx) that were treated with 2mg/kg of doxorubicin in vivo for 1 hr. Tubulin and Histone serve as Western blot loading controls. Data are average of 3 replicates +/−SEM. See also Fig. S1, Table S1.

Fig. 2.. eIF2α phosphorylation promotes METTL3 translation.

Western blots of MCF7 cells treated with A. poly I:C to induce eIF2α phosphorylation (2ug/ml). B. m6A dot blot analysis with poly I:C treatment. C. Treatment of hormone positive patient sample BT29 and BT30 cells with Sal003 (10uM) phosphatase inhibitor to retain eIF2α phosphorylation. D. Western blot of METTL3, METTL14 and ATF4 levels that do not increase when 500nM doxorubicin-treated patient sample cells that are also co-treated with 5uM Trazodone, or with E. with 1uM ISRIB, to override the effect of eIF2α phosphorylation. F. qPCR analysis of METTL3 and METTL14 RNA levels in 500nM doxorubicin (1day) treated MCF7 cells, compared to untreated cells. Polysome fractionation followed by qPCR analysis of polysome fractions G. for METTL3 mRNA and KI67 mRNA in 500nM doxorubicin (1day) treated MCF7 cells compared to untreated cells. H. for METTL3 mRNA (below) in Sal003 (10uM 1 day) treated MCF7 cells compared to untreated cells. I. Nascent amino acid labeling and METTL3 immunoprecipitation (Western blot of METTL3 with dark exposure of 1/10 Inputs) from 500nM doxorubicin (1day) treated versus untreated MCF7 cells to detect translation levels (quantitation below) of newly labeled immunoprecipitated METTL3 in these cells. Tubulin and Calnexin serve as Western blot loading controls. Data are average of 3 replicates +/−SEM. See also Fig. S2.

Fig 3.. EIF3 promotes METTL3 translation in doxorubicin-treated MCF7 cells.

A. i. Western blot of eIF3a, eIF3m, and eIF3d in eluates after streptavidin affinity purification by biotin-tagged antisense of METTL3 and METTL14 mRNAs compared to a scrambled control antisense, from in vivo UV and formaldehyde (HCHO) crosslinked, 250 nM and 500 nM doxorubicin (12 h and 1 day) treated or untreated MCF7 cells (scheme shown above, data in Table S2A). ii. qPCR of METTL3 and METTL14 RNAs co-purified by antisense compared to control antisense in doxorubicin treated cells. B. Western blot of METTL3 in MCF7 cells with eIF3d overexpression compared to control, which are 500nM doxorubicin (1day) treated or untreated. C. Chemotherapy survival cell counts of eIF3d overexpressing MCF7 cells compared to control cells, treated with 500nM doxorubicin for 20 h. D. Western blot of eIF3a depletion compared to control vector, transfected in MCF7 cells, followed by 500nM doxorubicin treatment to test E. chemotherapy survival by percentage surviving cell count after 3 days. Translation efficiency of Firefly luciferase reporter bearing F. the 5′UTR of METTL3 normalized to co-transfected Renilla in untreated and 500nM doxorubicin (1 day) treated MCF7 cells, normalized for reporter RNA levels, or G. bearing the 3′UTR of METTL3. H. Translation efficiency of Firefly luciferase reporter bearing the 5′UTR of METTL3 compared to an identical reporter bearing mutated GAC (m6A modification motif mutated to GUC) site upstream of the reporter start site in the 5′UTR of METTL3, normalized to co-transfected Renilla and reporter RNA levels, in 500nM doxorubicin (1 day) treated MCF7 cells. I. Translation of Firefly luciferase reporter bearing the 5′UTR of METTL3 in METTL3 shRNA MCF7 cells compared to control shRNA cells, with or without 500nM doxorubicin treatment for 1 day. Tubulin is the loading control for Western blots. Data are average of 3 replicates +/−SEM. See also Fig. S3, Table S1, S2A-B.

Fig. 4. METTL3 suppresses proliferation and anti-viral immune response, while promoting invasion genes, causing chemosurvival and limiting immune cell migration.

A. i. m6A IP RNAs from 500nM doxorubicin treated (1 day) cells but not in untreated cells (2335, Table S3) were compared with genes upregulated in shMETTL3 cells versus control shRNA cells (5168, Table S3 for RNA levels, Table S4 for protein levels) to identify RNAs that are m6A marked and downregulated by METTL3 (401, Venn Diagram, Table S5 for commonly identified genes from Table S3-S4). Upregulated genes were also compared but were fewer. ii. RNAs that are m6A modified: cell cycle RNAs (Ki67, PLK1), and anti-viral (DDX58 or RIG-I, PKR) RNAs but not control RNAs such as Actin and tRNA-lys are associated with m6A antibody and are increased upon METTL3 depletion in shMETTL3 MCF7 cells at the protein and RNA levels by TMT spectrometry and microarray respectively. iii. GSEA analysis of m6A-associated RNAs, reveals enrichment of cell cycle genes that are significantly (>=1.5 fold) upregulated upon METTL3 depletion. B. i. Western blot analysis of RIG-I anti-viral protein in shMETTL3 compared to control shRNA expressing MCF7 cells treated with 500nM doxorubicin for 1 day. ii. M6A antibody immunoprecipitation of RIG-I and PKR anti-viral gene mRNAs compared to IgG control. C. Anti-viral response to treatment with poly I: C (2ug/ml for 18 hrs) was tested by qPCR levels of RIG-I and PKR RNAs compared to control Actin mRNA by qPCR. This was tested in METTL3 overexpressing MCF7 cells compared to catalytic mutant METTL3 CD-M3 overexpressing cells. Shown Western blot of METTL3 levels in MCF7 cells with METTL3 and CD-M3 overexpression, compared to control vector. D. CASP9 and STAT1 downstream targets of RIG-I, are increased in METTL3 depleted MCF7 cells, as observed by qPCR analysis normalized for Actin mRNA. E. Inhibition of doxorubicin chemosurvival in MCF7 cells, with small molecules that override eIF2α phosphorylation (trazodone, ISRIB), compared to buffer treated control. 500nM Doxorubicin-treated MCF7 cells were co-treated with i. 1uM of ISRIB, ii. 5uM of Trazodone or DMSO buffer. Western blot analyses of METTL3 and METTL14 levels and ATF4 levels are shown with these (Trazodone or ISRIB combined with doxorubicin) treatments in Fig. 2D–E compared to buffer treated control cells. F. Ki67 cell cycle gene increases in i. Doxorubicin-treated shMETTL3 MCF7 cells compared to control shRNA cells, and in ii. Doxorubicin-treated MCF7 cells that are co-treated with 5uM of Trazodone, compared to DMSO buffer. G. i. Western blot showing decrease of invasion genes upon METTL3 depletion in shMETTL3 MCF7 cells compared to control shRNA cells. ii. Invasion assay showing decrease in invading cells upon METTL3 knockdown compared to control shRNA MCF7 cells. H. i. Bar graph shows fewer CD14+ monocytes migrate towards doxorubicin-treated MCF7 cells where METLL3 levels are increased compared to untreated MCF7 cells. ii. Bar graph shows more CD14+ monocytes migrate towards doxorubicin-treated MCF7 cells that were co-treated with ISRIB that prevents increase of METTL3 levels. Data are average of 3 replicates +/−SEM. Actin and Tubulin are loading controls for Western blots. See also Fig. S4, Table S3–5.

Fig. 5.. Metabolic genes that control methylation, are altered over time post-therapy, coincident with reduced m6A on METTL3 mRNA over time of doxorubicin treatment.

A. Metabolism genes and one-carbon gene RNAs that are m6A modified (from Table S3-S4), and are regulated in doxorubicin-treated cells compared to untreated (from Table S1), and conversely in shMETTL3 cells compared to control shRNA MCF7 cells (from Table S3-S4). B. Western blot of metabolism genes over time of doxorubicin treatment: one-carbon metabolism genes (SHMT, MAT2B), ATF4 that increases PHGDH that can enhance α-ketoglutarate (αKG) that can lead to demethylation. Western blot analysis of impact on METTL3 levels upon addition of: C. SHIN1, the inhibitor of SHMT compared to buffer control, D. αKG compared to succinate or buffer control, and E. αKG on chemosurvival. F. m6A antibody immunoprecipitation at an early time-point after doxorubicin treatment (12 h where METTL3 increases, Fig. 1A) and at a later time (48 h where METTL3 levels do not increase, Fig. S1F), compared to IgG control, followed by qPCR analysis for METTL3 mRNA. See also Fig. S5, Table S3–5.