N6-methyladenosine regulates glycolysis of cancer cells through PDK4

Abstract

Studies on biological functions of N⁶-methyladenosine (m⁶A) modification in mRNA have sprung up in recent years. We find m⁶A can positively regulate the glycolysis of cancer cells. Specifically, m⁶A-sequencing and functional studies confirm that pyruvate dehydrogenase kinase 4 (PDK4) is involved in m⁶A regulated glycolysis and ATP generation. The m⁶A modified 5'UTR of PDK4 positively regulates its translation elongation and mRNA stability via binding with YTHDF1/eEF-2 complex and IGF2BP3, respectively. Targeted specific demethylation of PDK4 m⁶A by dm⁶ACRISPR system can significantly decrease the expression of PDK4 and glycolysis of cancer cells. Further, TATA-binding protein (TBP) can transcriptionally increase the expression of Mettl3 in cervical cancer cells via binding to its promoter. In vivo and clinical data confirm the positive roles of m⁶A/PDK4 in tumor growth and progression of cervical and liver cancer. Our study reveals that m⁶A regulates glycolysis of cancer cells through PDK4.

Fig. 1. m⁶A regulates glycolysis and ATP generation of cancer cells.

a–c The glucose consumption (a), lactate production (b), and ATP levels (c) in Mettl3^Mut/- HeLa, sh-Mettl3 Huh7 and their corresponding control cells. d, e The cellular ECAR (d) and OCR (e) were measured in wild type and Mettl3^Mut/- HeLa cells. f, g The cellular ECAR (f) and OCR (g) were measured in sh-control and sh-Mettl3 Huh7 cells. b The glucose consumption, lactate production, and ATP levels in Huh7 cells transfected with vector control or ALKBH5 constructs. Data are presented as the mean ± SD from three independent experiments. **p < 0.01 by two tailed t test for (a) (p = 0.005 and p < 0.0001), (b) (p < 0.0001), (c) (p < 0.0001), and (h) (p < 0.0001).

Fig. 2. PDK4 mediates m⁶A regulated glycolysis and ATP generation of cancer cells.

a Gene ontology analysis was performed on a subset of downregulated genes in Mettl3^Mut/- HeLa cells. Log2 fold change, (KD: WT) < −0.5, was applied as the threshold cutoff. Several biological processes involved in metabolic processes were enriched and highlighted in bold; Significance shown as –Log10 Bonferroni p-value after multiple hypothesis correction. b GSEA reveals negative enrichment of genes in glycolysis gluconeogenesis sets of Mettl3^Mut/- HeLa cells. c Venn diagram shows substantial and significant overlap among metabolic genes, variated genes in Mettl3^Mut/- HeLa cells (>2 folds), and m⁶A enriched genes in wild type HeLa cells (>3 folds than input). d m⁶A peaks were enriched in 5′UTR and 3′UTRs of PDK4 genes from m⁶A RIP-seq data; e m⁶A RIP-qPCR analysis of PDK4 mRNA in wild type and Mettl3^Mut/- HeLa cells. f m⁶A RIP-qPCR analysis of PDK4 mRNA in sh-Con and sh-Mettl3 Huh7 cells. g The expression of PDK4 in Mettl3^Mut/- HeLa, sh-Mettl3 Huh7, or over expression of ALKBH5 and their corresponding control cells were measured by western blot analysis. h Cells were transfected with vector control or Mettl3 construct for 24 h, the expression of PDK4 was measured. i The mRNA of PDK4 in Mettl3^Mut/- HeLa, sh-Mettl3 Huh7 and their corresponding control cells were measured by qRT-PCR. j The glucose consumption, lactate production, and ATP levels in wild type or Mettl3^Mut/- HeLa cells transfected with PDK4 constructs for 24 h. Data are presented as the mean ± SD from three independent experiments. A representative from a total of two to three independent experiments is shown for (g) and (h). **p < 0.01, NS, no significant, by random permutation test for (b), by two-way ANOVA for (e) (p = 0.0001 and p = 0.0004, respectively) and (f) (p < 0.0001 and p = 0.0007, respectively), two-tailed unpaired Student’s t test for (i) (p < 0.0001), and one-way ANOVA for (j) (p < 0.0001 and p = 0.0008 for glucose consumption, p < 0.0001 and p = 0.0011 for lactate production, and p < 0.0001 for ATP levels, respectively).

Fig. 3. m⁶A regulates the mRNA stability and translation of PDK4 in cancer cells.

a Cells were transfected with pGL3-Basic-PDK4-luc reporter and pRL-TK plasmid for 24 h. Results were presented as the ratios between the activity of the reporter plasmid and pRL-TK. b The relative levels of nuclear versus cytoplasmic PDK4 mRNA in wild-type and Mettl3^Mut/- cells, or sh-control and sh-Mettl3 Huh7 cells. c After treatment with Act-D to inhibit transcription, the precursor mRNA levels of PDK4 were checked in wild type and Mettl3^Mut/- cells. d After treatment with Act-D for the indicated times, the mature mRNA levels of PDK4 were checked in wild type and Mettl3^Mut/- cells. e HeLa cells were pre-transfected with vector control or Mettl3 construct for 24 h and then further treated with CHX (10 μg/ml) or MG-132 (5 μM) for 6 h, the expression of PDK4 was detected by western blot analysis (left) and quantitatively analyzed (right). f Cells were treated with 10 μg/ml CHX for the indicated time periods, the expression of PDK4 was detected by western blot analysis (left) and quantitatively analyzed (right). g Wild-type or Mettl3^Mut/- HeLa cells were transfected with pmirGLO-PDK4 reporter for 24 h. The translation outcome was determined as a relative signal of F-luc divided by R-luc, the mRNA abundance was determined by qRT-PCR of F-luc and R-luc, and the translation efficiency of PDK4 is defined as the quotient of reporter protein production (F-luc/R-luc) divided by mRNA abundance. h qRT-PCR checked the mRNA levels of PDK4 in non-ribosome portion (<40S), 40S, 60S, 80S, and polysome fractions in wide type and Mettl3^Mut/- HeLa cells. Data are presented as the mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, NS, no significant, by two-tailed unpaired Student’s t test for (a) and (c) (p < 0.0001 for all comparisons), and two-way ANOVA for (b) (p > 0.05), and by one-way ANOVA for (e) (p = 0.005) and (h) (p < 0.0001 for all comparisons).

Fig. 4. Methylation sites of PDK4 involved in m⁶A regulated expression of PDK4.

a Schematic representation of positions of m⁶A motifs within PDK4 mRNA. b The m⁶A in 5′UTR or 3′UTR of PDK4 in wild type or Mettl3^Mut/- HeLa cells were analyzed by m⁶A-RIP-qPCR using fragmented RNA. c Schematic representation of mutated (GGAC to GGCC) 3′UTR of pmirGLO vector to investigate the roles of m⁶A in 3′UTR in PDK4 expression. d Wild-type or Mettl3^Mut/- HeLa cells were transfected with pmirGLO-PDK4-3′UTR-WT reporter for 24 h. The protein, mRNA and translation efficiency were determined. e The relative luciferase activity of F-Luc/R-Luc of pmirGLO-3′UTR-WT, or pmirGLO-3′UTR-Mut-1/-2/-3 in wild type and Mettl3^Mut/- HeLa cells; f Schematic representation of mutation in 5′UTR to investigate the m⁶A roles on PDK4 expression. g Wild-type or Mettl3^Mut/- HeLa cells were transfected with pGL3-PDK4-5′UTR-WT or pGL3-PDK4-5′UTR-Mut1/2 reporter for 24 h. The protein, mRNA and translation efficiency were determined. h pcDNA-PDK4-5′UTR-WT or pcDNA-PDK4-5′UTR-Mut1/2 was transfected into wild type or Mettl3^Mut/- HeLa cells for 24 h. Protein expression was measured by western blot analysis (left) and quantitatively analyzed (right). i, j pcDNA-PDK4-5′UTR-WT (i) or pcDNA-PDK4-5′UTR-Mut1 (j) was transfected into wild type or Mettl3^Mut/- HeLa cells for 24 h and then further treated with Act-D for the indicated times. The mRNA of PDK4 was checked by qRT-PCR. Data are presented as the mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, NS, no significant, by two-tailed unpaired Student’s t test for (b) (p < 0.0001), (d) (p > 0.05), and (h) (p < 0.0001 for all), and two-way ANOVA for (c) and (i) (p < 0.001 for all).

Fig. 5. Factors involved in m⁶A regulated expression of PDK4.

a RIP-qPCR analysis of PDK4 mRNA in wild type HeLa cells by use of antibody of YTHDF2, YTHDF3, and IGF2BP1~3. b IGF2BP3 RIP-qPCR analysis of PDK4 mRNA in wild type or Mettl3^Mut/- HeLa cells. c Wild type or Mettl3^Mut/- HeLa cells were transfected with si-NC or si-IGF2BP3 for 24 h, the expression of PDK4 was checked by western blot analysis (left) and quantitatively analyzed (right). d, e HeLa cells were transfected with si-NC, si-IGF2BP3, pcDNA-PDK4-5′UTR-WT (d) or pcDNA-PDK4-5′UTR-Mut1 (e) for 24 h and then further treated with Act-D for the indicated times. The mRNA of PDK4 was checked by qRT-PCR. f YTHDF1 RIP-qPCR analysis of PDK4 mRNA in wild type or Mettl3^Mut/- HeLa cells. g Binding of YTHDF1 with the 5′UTR or 3′UTR in wild type or Mettl3^Mut/- HeLa cells were analyzed by YTHDF1 RIP-qPCR using fragmented RNA. h Wild type or Mettl3^Mut/- HeLa cells were transfected with vector or YTHDF1 construct for 24 h, the expression of PDK4 was checked by western blot analysis (left) and quantitatively analyzed (right). i HeLa cells were transfected with vector, YTHDF1 construct, pcDNA-PDK4-5′UTR-WT, and pcDNA-PDK4-5′UTR-Mut1 for 24 h, the expression of PDK4 was checked by western blot analysis (left) and quantitatively analyzed (right). j eEF-1 and eEF-2 RIP-qPCR analysis of PDK4 mRNA in wild type or Mettl3^Mut/- HeLa cells. Data are presented as the mean ± SD from three independent experiments. *p < 0.05, **p < 0.01, NS, no significant, by two-tailed unpaired Student’s t test for (a) (p < 0.0001) and (g) (p < 0.0001 and p = 0.11, respectively), and two-way ANOVA for (b) (p < 0.0001 and p = 0.004, respectively), (c, d) (p < 0.0001 for all), (f) (p = 0.0003 and p = 0.001, respectively), (h–j).

Fig. 6. Targeting m⁶A of PDK4 by dm⁶ACRISPR re-programs metabolic of cancer cells.

a Schematic representation of positions of m⁶A site within PDK4 mRNA and the regions targeted by three gRNAs, respectively. b The threshold cycle (Ct) of qPCR showing SELECT results for detecting m⁶A site in the potential m⁶A site of PDK4 5′UTR in Mettl3^Mut/- HeLa, sh-Mettl3 Huh7 and their corresponding control cells. c The threshold cycle (Ct) of qPCR showing SELECT results for detecting m⁶A site in 5′UTR of PDK4 in HeLa cells transfected with dCas13b-ALKBH5 combined with gRNA control or gRNA1/2/3, respectively, for 24 h. d, e The mRNA (d) or protein (e) expression of PDK4 in HeLa cells transfected with dCas13b-ALKBH5 combined with gRNA control or gRNA1/2/3, respectively, for 24 h. f RIP-qPCR analysis of PDK4 mRNA in HeLa cells transfected with dCas13b-ALKBH5 combined with gRNA control (dC-A5) or gRNA for PDK4 (dC-A5 + gRNA) for 24 h by use of antibodies against YTHDF1 and IGF2BP3, respectively. g HeLa cells were transfected with gRNA control, gRNA1 for PDK4, and dCas13b-ALKBH5 for 24 h and then further treated with Act-D for the indicated times. The mRNA of PDK4 was checked by qRT-PCR. h–j The glucose consumption (h), lactate production (i), and ATP levels (j) in HeLa cells transfected with gRNA control, gRNA1 for PDK4, and dCas13b-ALKBH5 or dCas13b-ALKBH5-Mut for 24 h. Data are presented as means ± SD from three independent experiments. **p < 0.01, NS, no significant, by two-tailed unpaired Student’s t test for (b) (p = 0.001 and p = 0.002, respectively), (h–j) one-way ANOVA for (c) (p < 0.001), (d) (p < 0.001), and (e) (p < 0.001), and two-way ANOVA for (f) and (j) (p < 0.001).

Fig. 7. TBP is responsible for the upregulation of Mettl3 in cervical cancer cells.

a–c The protein (a), mature mRNA (b), or precursor mRNA (c) of Mettl3 in ECT1/E6E7, HeLa, and SiHa cells were checked. d The promoter activities of Mettl3 in ECT1/E6E7, HeLa, and SiHa cells were checked by dual-luciferase assay. e Venn diagram shows the overlap of transcription factors of Mettl3 predicted by PROMO and ChIPBase, respectively. f The mRNA expression of potential transcription factors of Mettl3 in ECT1/E6E7, HeLa, and SiHa cells were checked by qRT-PCR analysis. g The protein expression of NRF1 and TBP was checked by western blot analysis (left) and quantitatively analyzed (right). h, i Cells were transfected by siRNA negative control (si-NC) or siRNAs of TBP for 24 h, the mRNA (h) and protein (i) levels of Mettl3 were checked. j The binding between TBP and promoter of Mettl3 was checked by ChIP-PCR using IgG or TBP antibody. k Binding between TBP transcriptional factor and the promoter of Mettl3 at the potential binding site 1 and 2 or negative site BLK was checked by ChIP-PCR. l Schematic representation of the mutated promoter in pGL3-Basic-Mettl3-luc reporter to investigate the role of TBP in Mettl3 expression. m HeLa cells were co-transfected with pGL3-Mettl3-WT-Luc, pGL3-Mettl3-Mut1-Luc, pGL3-Mettl3-Mut2-Luc, pRL-TK plasmid and si-NC or si-TBP-1 for 24 h. Results were presented as the ratio between the activity of the reporter plasmid and pRL-TK. n Correlation between Mettl3 and TBP in cervical cancer patients (n = 309) from ChIPBase database. Data are presented as the mean ± SD from three independent experiments. **p < 0.01. NS, no significant, by one-way ANOVA for (b) (p < 0.001 for all), (c) (p < 0.001 for all), (d) (p < 0.001 for all), (f, g) (p < 0.001 for all), and (h) (p < 0.001 for all), two-tailed unpaired Student’s t test for (j) (p < 0.001 for all), (k) (p < 0.001 for all), and (n) (p = 0.0005), and by two-way ANOVA for (m) (p < 0.001 for all).

Fig. 8. PDK4 is involved in m⁶A regulated cancer progression.

The relative cell proliferation of wild type and Mettl3^Mut/- HeLa cells stably transfected with vector control or PDK4 constructs; The relative cell proliferation of sh-control and sh-Mettl3 Huh7 cells stably transfected with vector control or PDK4 constructs; Wild type and Mettl3^Mut/- HeLa cells stably transfected with vector control or PDK4 constructs were treated with increasing concentrations of Dox for 24 h, and the cell proliferation was tested. a The tumor growth curves of wild type and Mettl3^Mut/- HeLa cells stably transfected with vector control or PDK4 constructs. b IHC (Mettl3 and PDK4)-stained paraffin-embedded sections obtained from wild type and Mettl3^Mut/- HeLa cells. The scale bar is 50 μM. c Correlation between Mettl3 and PDK4 in cervical cancer patients (n = 169) from TCGA database. g–i The relative mRNA expression of Mettl3 (G), IGF2BP3 (H), or YTHDF1 in Oncomine datasets. j, k The Kaplan–Meier survival curves of DFS based on Mettl3 (j) or TBP (k) expression in cervical cancer patients from TCGA data base. l, m The Kaplan–Meier survival curves of DFS (l) or OS (m) based on PDK4 expression in cervical cancer patients from TCGA data base. Data are presented as means ± SD from three independent experiments. A representative from a total of three independent experiments is shown for (e). NS, no significant, by one-way ANOVA for (a), (b), and (d), by two-tailed unpaired Student’s t test for (f–i), and by two-sided log-ranjk test for (j–m).