YTHDF1-enhanced iron metabolism depends on TFRC m6A methylation.

Background: Among head and neck squamous cell carcinomas (HNSCCs), hypopharyngeal squamous cell carcinoma (HPSCC) has the worst prognosis. Iron metabolism, which plays a crucial role in tumor progression, is mainly regulated by alterations to genes and post-transcriptional processes. The recent discovery of the N6-methyladenosine (m6A) modification has expanded the realm of previously undiscovered post-transcriptional gene regulation mechanisms in eukaryotes. Many studies have demonstrated that m6A methylation represents a distinct layer of epigenetic deregulation in carcinogenesis and tumor proliferation. However, the status of m6A modification and iron metabolism in HPSCC remains unknown.
Methods: Bioinformatics analysis, sample analysis, and transcriptome sequencing were performed to evaluate the correlation between m6A modification and iron metabolism. Iron metabolic and cell biological analyses were conducted to evaluate the effect of the m6A reader YTHDF1 on HPSCC proliferation and iron metabolism. Transcriptome-wide m6A-seq and RIP-seq data were mapped to explore the molecular mechanism of YTHDF1 function in HPSCC.
Results: YTHDF1 was found to be closely associated with ferritin levels and intratumoral iron concentrations in HPSCC patients at Sir Run Run Shaw Hospital. YTHDF1 induced-HPSCC tumorigenesis depends on iron metabolism in vivo in vitro. Mechanistically, YTHDF1 methyltransferase domain interacts with the 3'UTR and 5'UTR of TRFC mRNA, then further positively regulates translation of m6A-modified TFRC mRNA. Gain-of-function and loss-of-function analyses validated the finding showing that TFRC is a crucial target gene for YTHDF1-mediated increases in iron metabolism.
Conclusion: YTHDF1 enhanced TFRC expression in HPSCC through an m6A-dependent mechanism. From a therapeutic perspective, targeting YTHDF1 and TFRC-mediated iron metabolism may be a promising strategy for HPSCC. © The author(s).

Introduction

Head and neck squamous cell carcinoma (HNSCC) is the sixth most common cancer worldwide, with approximately 700,000 cases diagnosed per year 1. Of all the areas affected (oral, oropharyngeal, nasal/paranasal, laryngeal, hypopharyngeal, cervical esophageal locations, etc.), hypopharyngeal squamous cell carcinoma (HPSCC) has the worst prognosis. HPSCC is generally detected at an advanced stage because of a lack of biomarkers for early diagnosis 2,3.

To date, except for EGFR inhibitors, no FDA-approved targeted therapies are available. With current standard multimodality therapies, the 5-year overall survival (OS) rate remains low, varying between 28% and 41% 4,5. Increasing evidence indicates that the unique clinical and biological characteristics of HPSCC are due to anatomic location of the tumor and genetic and transcriptome alterations 6,7. For example, co-occurring CCND1 and CDKN2A mutations, and chromosomal instability markers are associated with chemoradiotherapy outcomes in advanced stage HPSCC patients 8. We found that multiple genomic alterations have been analyzed in relevance of sensitivity to chemotherapy, targeted therapy, and ionizing radiation 9. However, few specific biomarkers and therapeutic targets for HPSCC management have identified and validated. Therefore, a better understanding of the molecular mechanism of HPSCC progression is critical.

Iron plays a crucial role in tumor progression due to both its major contributions to tumor survival and reprogramming of the tumor microenvironment 10. Large-scale epidemiological studies have shown a direct correlation between systemic iron levels and increased cancer risk 11,12. Dysregulation of iron homeostasis typically manifests as upregulation of iron uptake and downregulation of iron efflux 13. First, tumor cells exhibit uncontrolled growth and proliferation and require more iron than do normal cells 14-16. In other words, excessive iron intake by cells is one of the hallmarks of carcinogenesis. Second, cytoplasmic iron is stored mainly as ferritin, which is critical for the maintenance of iron metabolism and protection of cells from oxidative damage 17; however, the levels of iron and ferritin in malignant tumors can be 5~6-fold higher than those in benign tumors 18. Excessive iron may cause redox imbalance and generate reactive oxygen species (ROS) in tumor cells, which increase genomic instability and proliferation 19. Moreover, serum ferritin is also increased in cancer patients compared to healthy individuals 20. Several studies have suggested that the reprogramming of iron metabolism is a major aspect of HNSCC. Multiple iron-related genes are significantly elevated in HNSCC relative to those in normal squamous epithelium, resulting in increased intratumoral iron accumulation and cell proliferation 21,22. To date, there is limited knowledge regarding the molecular regulation of intracellular iron metabolism in HNSCC 22.

The identification and functional characterization of proteins that specifically recognize the m6A modification in RNA (RNA m6A) revealed that this modification induces cells to accelerate mRNA metabolism and translation. These proteins include methyltransferase complexes, demethylases, and a group of specific RNA-binding proteins, also known as “writers”, “erasers”, and “readers” 23. Writers (such as METTL3/METTL14 complexes) and erasers (such as FTO and ALKBH5) determine the prevalence and distribution of m6A. Previous studies demonstrated that the m6A modification mediated by writers METTL3 and METTL14 enhanced the stability of LncAROD, thus exacerbating the malignant behavior of HNSCC cells 24. METTL14 also drives Epstein-Barr virus (EBV)-mediated tumorigenesis 25. The m6A reader IGF2BP2 was reported to be upregulated in HNSCC tumor tissues, and high IGF2BP2 expression is associated with poor prognosis for patients with HNSCC 26. Readers (such as YTHDFs) primarily recruit proteins that mediate the effects of m6A modification on mRNAs 27-29. For YTHDFs, including YTHDF1, YTHDF2 and YTHDF3; YTHDF1, are enriched in the cytoplasm and have been recognized to enhance the translational efficiency of m6A-modified mRNAs 30. YTHDF1 expression is amplified in various types of cancers, including HNSCC; furthermore, it has a critical oncogenic role 31-38. Recent findings also showed that YTHDF1-deficient mice show an elevated antigen-specific CD8(+) T cell anti-tumor response 32. YTHDF2 mediates mRNA instability 30,39, while YTHDF3 facilitates translation in conjunction with YTHDF1 and affects the decay of m6A-modified mRNAs 40.

Previous studies demonstrated that the m6A modification was correlated with carcinogenesis, tumor proliferation and chemoresistance of cancer cells 31,34,41-43. However, the effect of m6A modifications on cancer progression depends on whether the m6A target gene is an oncogene or a tumor suppressor, the extent of the change in m6A level in cancer cells, and, most importantly, the post-modification regulation of target mRNA. Despite the numerous discoveries highlighting the association between m6A modification and cancer development, the underlying regulatory mechanism of m6A modification in HPSCC, especially with regard to iron metabolism of HPSCC, remains poorly understood. Herein, we investigated the role of the m6A modification in HPSCC and identified the oncogenic role of YTHDF1. By thoroughly investigating the role of YTHDF1 deregulation in HPSCC, we found that YTHDF1 is closely associated with tumor proliferation and iron metabolism. Our study illustrates the critical role of YTHDF1 in human HPSCC carcinogenesis and iron transport pathways. These results might shed light on possible clinical therapies for treating HPSCC patients with high iron overload.

Materials and Methods

Patient specimens

Two cohorts of HPSCC patients who underwent surgery at individual medical centers between 2012 and 2020 were included in this study.

Tissue specimens from patients in cohort 1 included whole blood, freshly resected cancerous and normal tissues, and paraffin-embedded tissues, all of which obtained from Sir Run Run Shaw Hospital, College of Medicine, Zhejiang University. The blood and excised cancerous and normal tissues were stored and numbered in the Sir Run Run Shaw Hospital biological specimen bank, whereas the paraffin-embedded surgical specimens were permanently stored in the Department of Pathology at Sir Run Run Shaw Hospital.

The tissue specimens from patients in cohort 2 comprised paraffin-embedded surgical specimens, which were obtained from The Second Affiliated Hospital of Nanchang University Medical College.

None of these patients had received radiotherapy or chemotherapy prior to surgery and had no history of hepatic disease.

Iron assay and intracellular chelatable iron (Fe2+) and ROS measurements

Intracellular iron (ferrous iron and ferric iron) levels were determined using an iron assay kit from Sigma-Aldrich (MAK025). Intracellular chelatable iron (Fe2+) was measured with the fluorescent indicator Phen Green SK fluorescent probe (#P- 14313, Life Technologies, Grand Island, NY, USA). Levels of ROS were determined using DCFH-DA (Solarbio, D6470) according to the manufacturer's instructions. The detailed protocols are described in the

m6A quantification

The global m6A levels in mRNA were measured with an EpiQuik m6A RNA methylation quantification kit (Colorimetric) (Epigenetic, Farmingdale, NY) following the manufacturer's protocol.

RNA-Seq, MeRIP and anti-m6A immunoprecipitation (MeRIP-qPCR)

Total RNA was extracted using Trizol reagent (Invitrogen, CA, USA) following the manufacturer's protocol. Eluted m6A-containing fragments (IP) and untreated input control fragments were converted into the final cDNA library by the dUTP method in accordance with strand-specific library preparation. The average insert size for the paired-end libraries was ~ 100 ± 50 bp. Then, we performed paired-end 2 × 150 bp sequencing on an Illumina NovaSeq 6000 platform at LC-BIO Bio-tech Ltd. (Hangzhou, China) following the vendor's recommended protocol. Real-time PCR was carried out following m6A-IP to quantify the changes in the m6A methylation levels of a specific target gene. The detailed MeRIP and MeRIP-qPCR protocols are described in the

RNA immunoprecipitation (RIP) and high-throughput sequencing

The enrichment of RNA was normalized to the IgG level. cDNA libraries were produced by employing an NEBNext UltraRNA Library Prep Kit for Illumina (New England Biolabs) and sequenced on an Illumina NovaSeq 6000 platform at LC-BIO Bio-tech Ltd. (Hangzhou, China) following the vendor's recommended protocol. RIP was performed using an EZ-Magna RIP RNA-binding protein immunoprecipitation kit (Millipore) following the manufacturer's protocol with some modifications. The detailed protocol is described in the .

Statistical analyses

Bioinformatic analyses, including Gene Ontology (GO), Kyoto Encyclopedia of Genes and Genomes (KEGG), and gene set enrichment analysis (GSEA), were performed using the OmicStudio tools at https://www.omicstudio.cn/tool. All statistical analyses were carried out using GraphPad Prism version 7 (GraphPad Software, CA) for Windows or R software (www.r-project.org). Statistical significance was assessed by unpaired two tailed Student t-tests, analysis of variance (ANOVA) or Spearman rank correlation. Recurrence-free survival was evaluated by the Kaplan-Meier method and log-rank test. The data are expressed as the means ± SD. Statistical significance was indicated as follows: n.s. no significance, *p < 0.05; **p < 0.01; ***p < 0.001, ****p < 0.0001. Each experiment was repeated independently at least three times.

More detailed materials and methods are in the .

Results

YTHDF1 is closely correlated with iron metabolism in HPSCC

To investigate the potential factors and mechanisms involved in intracellular iron levels in the general context of HNSCC, we analyzed the expression levels of iron-related genes by evaluating RNA-sequencing (RNA-seq) data from The Cancer Genome Atlas (TCGA) HNSCC data set, which contains 504 pairs of HNSCC and matched normal specimens. The expression patterns of genes critical for iron metabolism, including aconitase 1 (ACO1, also known as IRP1), cytochrome b reductase 1 (CYBRD1), ferritin heavy chain (FTH1), ferroportin (FPN, also known as SLC40A1), hepcidin antimicrobial peptide (HAMP), SLC11A1 and transferrin receptors (TFRCs), were assessed. Among those iron-related genes, FTH1 and TFRC showed significantly elevated expression (Figure . Next, the expression patterns of m6A RNA methylation regulators were assessed, which revealed that the expression levels of multiple regulators were remarkably different between the HNSCC tumor samples and normal control samples (.

Because ferritin is encoded by FTH1 and TFRC is the most important gene for intracellular iron uptake, we concentrated on exploring the correlation between m6A modification and both ferritin level and intratumoral iron concentration in HPSCC patients. Analyses of ferritin expression and m6A modifications were then undertaken using samples from 50 primary HPSCC patients in cohort 1 with measured ferritin expression and pre-operative cervical magnetic resonance imaging (MRI). There was a significant positive correlation between serum ferritin and YTHDF1 expression, as measured by serum ferritin concentration (Figure . In addition, a significant correlation between intratumoral iron content (nmol) and relative YTHDF1 expression was observed in the HPSCC patients of cohort 1 (Figure . To validate this correlation, intratumoral iron concentration (IC) based on the relaxation rates R2 (1/T2) of the cervical contrast MRI was measured by an experienced radiologist at Sir Run Run Shaw Hospital (Figure The R2 and R2* values generated comparable estimates of non-invasive intratumoral IC in a study by Wood et al 44. Data analysis revealed a remarkable correlation between high IC values and YTHDF1 immunohistochemical (IHC) staining (Figure . However, quantification of RNA methylation revealed that the level of m6A modification was not increased in the HPSCC patients with higher serum ferritin, nor did it affect the intratumoral iron content (nmol) (.

To elucidate whether YTHDF1 plays a role in HPSCC iron metabolism and tumorigenesis, we performed transcriptome sequencing (RNA-seq) with YTHDF1-knockdown and control HPSCC FaDu cells. YTHDF1 knockdown induced by lentiviral shRNAs was confirmed at both the mRNA and protein levels (). A total of 34083 genes were identified and quantified by RNA-seq. YTHDF1 knockdown resulted in 1749 significantly altered genes (Student's t-test, p < 0.05) comprising 860 upregulated genes (49.1%) and 889(50.8%) downregulated genes ( KEGG and GO analysis showed that the following pathways were enriched with these genes: signal transduction, regulation of signaling receptor activity, protein binding, oxidation-reduction process, cellular iron homeostasis and cell proliferation (Figure GSEA of the RNA-seq data revealed gene signatures relating to tumor invasiveness, migration and Myc proto-oncogene protein (MYC, a known major contributor to TFRC upregulation in cancer cells), which were enriched in the control HPSCC cells compared with the YTHDF1-knockdown cells, indicating a role for YTHDF1 in HPSCC tumorigenesis and proliferation 45 (Figure . Taken together, these data suggest that YTHDF1 promotes tumor progression and iron metabolism in HPSCC.

YTHDF1 promotes iron metabolism in HPSCC cells

According to the literature, iron accumulation in the cell via extracellular transport into cells or ferritin release result in the formation of a common labile iron (Fe2+) pool (LIP) that is available for redox cycling 46,47. The LIP can readily contribute to steady-state levels of ROS in cancer cells through Fe2+-dioxygen biochemistry and Fenton reactions, increasing genomic instability and proliferation 48-50. Therefore, the LIP is an active hub in the genesis of cancer that links iron metabolism to the hallmarks of cancer 51. In vitro experiments showed that YTHDF1 knockdown significantly decreased total HPSCC intracellular iron levels (Figure , whereas in vivo experiments indicated that tumor xenografts with YTHDF1 knockdown had lower ferritin expression, as determined by IHC analysis (Figure To determine whether YTHDF1 directly promotes HPSCC iron metabolism, we conducted intracellular Fe2+ and ROS analyses. As predicted, knocking down YTHDF1 remarkably reduced the intracellular Fe2+ and ROS levels, as determined by Phen Green SK fluorescence iron staining and dihydroethidium (DHE) ROS staining, respectively (Figure . To elucidate whether YTHDF1-regulated iron metabolism is dependent on m6A modifications, we generated two point mutations, K395 and Y397, in the YTH domain and added a FLAG tag. Mutating the YTH domain abrogated the binding capacity of YTHDF1 with mRNA (Figure , as reported in previous studies 52,53. Both YTHDF1 wild-type (YTH1-WT) and mutant (YTH1-Mut) recombination plasmids were successfully transfected into FaDu cells (. We found that ectopic expression of YTH1-WT increased intracellular Fe2+ and ROS levels, while abrogating the activity of the YTH domain dramatically decreased both intracellular Fe2+ and ROS levels in FaDu cells (Figure . Considering these results, we suggest that YTHDF1 promotes HPSCC iron metabolism in an m6A-dependent manner.

YTHDF1-induced HPSCC proliferation depends on the promotion of iron metabolism

Shown by the cBioPortal network, YTHDF1 is frequently amplified and mutated in various squamous cell carcinomas (cervical, lung, head and neck, and esophageal cancers) (Figure . In addition, based on a GEO dataset (GSE79637) of HNSCC specimens, YTHDF1 expression was more prominent in highly metastatic lines (FaDu origin: hypopharynx, Detroit 562 origin: pleural effusion) than in nonmetastatic lines (YCU-OR891 origin: oral floor, YCU-MS861 origin: maxillary sinus) (Figure ). The function of YTHDF1 in HPSCC has never been reported. Our functional validation data showed that, compared with the shRNA control cells, Detroit 562 and FaDu cells with YTHDF1 knockdown showed significant reductions in viability (Figure , colony formation (Figure , and migration (Figure in vitro. Knocking down YTHDF1 dramatically mitigated tumor growth, as reflected by tumor volumes and weights in the xenograft mouse models (Figure . Furthermore, overexpression of YTHDF1-WT increased the proliferation (, colony formation ( and migration ( of Detroit 562 cells. YTHDF1 overexpression also remarkably increased xenograft tumor volumes ( and weights (; however, the same significant oncogenic effect was not observed in cells overexpressing YTHDF1-Mut in vitro ( or in vivo (. In addition, treatment with deferiprone (DFP, a new generation of intracellular iron chelators in clinical trials 48,54,55) significantly blocked YTHDF1-induced HPSCC iron accumulation, growth and colony formation in vivo and in vitro (Figure . The data strongly suggest that YTHDF1 plays a pivotal role in cancer progression by regulating intracellular iron metabolisms. We also confirmed that the YTHDF1 protein in the xenograft mouse models were not significantly affected upon DFP treatment by Western blot analysis (, excluding the possibility that cellular iron level regulated YTHDF1 levels. These data strongly suggest that YTHDF1 plays a pivotal role in cancer progression by regulating intracellular iron metabolism.

Transcriptome-wide m6A-seq, RNA-seq and RIP-seq assays identify potential targets of YTHDF1 in HPSCC

To identify and localize m6A sites, YTHDF1-knockdown and control FaDu cells were subjected to transcriptome-wide m6A-sequencing (m6A-seq and MeRIP-seq) assays. Principal component analysis (PCA) showed that two repeats (shCON: control1 and control2; shYTHDF1: shY11 and shY12) of each sample clustered together, suggesting good repeatability among the two replicates of each group (. By applying the HOMER motif discovery tool, we found that the “GGAC” consensus sequence was the primary motif enriched in the m6A peaks (Figure In agreement with previous reports, we found that the m6A peak density was not significantly changed (Figure . Peaks were located in protein-coding transcripts and enriched in the 5'UTR and 3′UTR, especially near stop codons (Figure .

Upon analysis of the RNA-seq data, the general transcription level did not change after YTHDF1 was knocked down in HPSCC cells, suggesting that the loss of YTHDF1 did not change the RNA abundance (. A total of 8846 (29.2%) m6A-modified transcripts overlapped with the RNA-seq data. The general m6A level of transcripts was not significantly changed, as 51.5% (7505 of 14562) of the genes were downregulated, and 48.5% (7050 of 14562) of the genes were upregulated. However, the general gene expression of YTHDF1 targets was downregulated in the YTHDF1-knockdown FaDu cells compared with that in the shCON cells (Figure . We verified this result by quantitative analysis of global RNA m6A levels in YTHDF1-knockdown and control FaDu cells, as measured with an EpiQuik m6A quantification assay. Indeed, no noticeable difference was found in the m6A levels of the two samples (. These results were in line with previous findings showing that YTHDF1 does not affect the RNA abundance of its targets but rather regulates protein synthesis by interacting with m6A-methylated mRNAs 32,35.

As a crucial m6A reader, YTHDF1 promotes the translation of m6A-methylated mRNAs and recruits translation initiation factors, thereby significantly improving translation efficiency 56. Therefore, we sequenced RNA obtained from the immunopurified complex of YTHDF1 (RIP-seq) to identify YTHDF1-bound mRNAs. RIP-seq revealed 2450 mRNAs as candidate targets of YTHDF1 (. Overlapping genes found through the RNA-seq, m6A-seq, and RIP-seq data analyses showed that 706 genes bound by YTHDF1 were marked with m6A (Figure . Among these 706 genes, 264 were downregulated. YTHDF1 is known to bind and affect m6A-methylated transcripts 30,57. Therefore, mRNA transcripts of proteins that were downregulated in YTHDF1-knockdown cells were likely potential targets. Moreover, GSEA showed that these 264 downregulated genes are involved in RNA metabolic processes, including GO_RNA binding, Reactome_Metabolism of RNA, GO_cellular iron homeostasis, and GO_Regulation_of_Transmembrane_Transport (. Six genes (CIAO1, TFRC, TET2, CYP2U1, EPAS1, and SLC25A28) with significantly decreased expression in the YTHDF1-knockdown cells (p < 0.05, fold change > 1.5) were selected as potential candidates (Figure . Noticeably, the m6A peaks fit well with YTHDF1-binding enrichment sequences (m6A-seq+RIP-seq) in these transcripts, as shown by Integrative Genomics Viewer (IGV) software (Figure . Significant m6A peaks as well as YTHDF1 binding enrichment in the TFRC were observed (Figure .

TFRC is overexpressed on the extracellular surface of the plasma membrane in a variety of solid cancer cells, where they enable increased iron uptake 45. Upon release from the TF-TFRC complex into the cytosol, ferric iron is reduced to ferrous iron by ferrireductases within endosomes 58,59. Moreover, TFRC is involved in the gene enrichment of RNA binding, iron homeostasis, and WEI_mycn targets with e-boxes. Several studies have suggested that the group of transcription factors regulating TFRC has oncogenic implications 48. Consistent with our results, previously reported m6A, RIP and CLIP data from the m6A2Target Database (http://m6a2target.canceromics.org/#/search/TFR) identified TFRC mRNA as a potential target of m6A readers (GSE78030, GSE92021, GSE78507, and GSE86214) on the 3'UTR, exons or promoter-TSS sites. In conclusion, TFRC is a functional target of YTHDF1 in HPSCC cells.

YTHDF1 regulates TFRC expression in HPSCC in an m6A methyltransferase- dependent manner

The best-studied m6A-dependent functions of the YTH family include regulation of mRNA stability, translation, splicing and lncRNA-mediated gene silencing 56. YTHDF1 facilitates the translation of m6A-modified mRNAs and induces a rapid gene expression response and controlled protein production capacity without changing the overall of mRNA levels 56. To examine whether TFRC is a direct target of YTHDF1, we assessed the transcription and translation of TFRC upon YTHDF1 knockdown. In line with the documented functions of YTHDF1, knocking down YTHDF1 decreased the protein level of TFRC without affecting mRNA expression in both Detroit 562 and FaDu cells (Figure

Moreover, RIP using an antibody against FLAG followed by qPCR (RIP-qPCR) revealed that the 3'UTR and 5'UTR of TFRC mRNA were immunoprecipitated effectively in both Detroit 562 and FaDu HPSCC cells transfected with YTHDF1-WT (Figure Moreover, we found that cells transduced with YTHDF1-WT but not YTHDF1-Mut exhibited increased protein expression of TFRC (Figure . Additionally, we constructed both WT and mutant TFRC luciferase reporter plasmids, which contained the WT 3'UTR and 5'UTR of TFRC or mutated sequences in which an adenosine residue that undergoes the m6A modification was replaced by a T residue (TFRC-m6A Mut) (Figure . As expected, compared with shCON cells, shYHDF1 cells showed substantially reduced luciferase activity of the individual reporter constructs carrying the WT 3′UTR and 5'UTR fragments of TFRC, and this decrease was completely abrogated in cells expressing the 3′UTR and 5'UTR mutant variants (Figure . Furthermore, overexpression of WT YTHDF1 but not mutant YTHDF1 significantly increased the luciferase activity of the individual reporter constructs carrying the WT 3′UTR and 5'UTR fragments of TFRC in FaDu cells (Figure . Then, we assessed the m6A modification status of TFRC mRNA by a gene-specific m6A assay, but no significant enrichment of TFRC mRNA was observed in HPSCC cells (Figure . To test whether YTHDF1 contributed to TFRC protein degradation, control and YTHDF1-knockdown Detroit 562 cells were treated with the protein translation inhibitor cycloheximide (CHX) 60,61. Western blot analysis revealed that knocking down YTHDF1 had no effect on the stability of TFRC protein in HPSCC cells (.

Next, we performed the polysome profiling assay. Polysome profiling supported the decreases of 80S monosome assembly and polysomes in shYTHDF1 Detroi562 cells (. The qRT- PCR showed that YTHDF1 knockdown resulted in significant lower TFRC mRNA in translation fractions (. These results support the notion that YTHDF1 regulates protein synthesis while excluding the possibility that YTHDF1 affects TFRC protein stability. Noticeably, TFRC is predominantly regulated by the iron regulatory protein (IRP/IRE) system at both the transcriptional and post-transcriptional levels 62. To determine whether YTHDF1 regulates the IRE/IRP system, we reanalyzed the sequence data. Neither different m6A peaks in the mRNA nor significant gene expression of these proteins was observed in the m6A-seq data. RIP-seq data indicated that IREB2 and YTHDF1 could bind each other, but RIP-qPCR failed to validate this finding (data not shown). These results confirmed that YTHDF1 enhanced TRFC expression via its methyltransferase domain at translational level.

TFRC is a crucial target gene for the YTHDF1 promotion of iron metabolism

The major mechanism by which m6A exerts its effects is determined by m6A-binding proteins 56. Therefore, we conducted functional experiments to investigate whether TRFC participates with YTHDF1 in promoting iron metabolism. WT and YTHDF1-knockdown HPSCC cells were transfected with control or TFRC-overexpressing plasmids. The overexpression efficiency of TFRC in Detroit 562 and FaDu cells was confirmed by Western blot analysis (. Ectopic expression of TFRC partially restored the viability (Figure , colony formation ability (Figure and xenograft tumor growth (Figure of YTHDF1-knockdown cells. The efficiency of TFRC shRNAs was confirmed by qPCR and Western blot analysis (. TFRC downregulation also led to significantly impaired YTHDF1-induced increases in cell viability and migration (. These results strongly indicated that TFRC is a critical target gene of YTHDF1 in HPSCC cells.

The intracellular iron level and ROS were synchronously assessed. As expected, TRFC overexpression restored the reductions in intracellular iron content, Fe2+ levels and ROS levels in YTHDF1-knockdown cells (Figure Furthermore, the downregulation of TFRC significantly reduced YTHDF1-mediated increases in iron content, intracellular Fe2+ level and ROS level in FaDu cells (. As TFRC expression led to increased iron uptake, we next measured the expression of STEAP4, an important ferrireductase that facilitates increased iron uptake by cells 63. FTH1 is a protein with ferroxidase activity that facilitates the conversion of excess Fe2+ to Fe3+ and then stores it in ferritin. Western blot analysis showed that TFRC downregulation also reduced the YTHDF1-induced increase in the protein levels of STEAP4 and FTH1, TRFC overexpression restored the protein levels of STEAP4 and FTH1 in YTHDF1-knockdown cells (. Taken together, the data show that TRFC mediates the iron regulatory function of YTHDF1 in HPSCC cells.

YTHDF1 is linked to poor prognoses of HPSCC patients receiving CCT/RT treatment

We next performed IHC analyses of HPSCC and paracancerous tissues from patients in cohort 1 and cohort 2. The percentage of YTHDF1-expressing cells was significantly higher in HPSCC tissues (65.7%; 67/102) than in normal epithelium tissues (17.1%; 12/70) (Figure . A significant difference was found between the clusters of T stage and N stage tissues (Figure In addition, samples with higher YTHDF1 expression showed intense TFRC staining, while samples with low YTHDF1 expression displayed lower levels of TFRC staining (Figure Surprisingly, Kaplan-Meier analysis showed that elevated TFRC expression in HPSCC cells was associated with poor prognosis; however, patients with high YTHDF1 expression did not have shorter disease-free survival, in contrast to our expectations (Figure . We obtained similar results when analyzing the TCGA head and neck dataset (Figure

Then, we reanalyzed the clinical characteristics and treatments for patients in cohort 1 and cohort 2. We found that 72.5% of the patients were treated with postoperative platinum-based chemoradiotherapy (CCT) or radiation (RT) alone. High TFRC expression correlated with a reduced response to chemoradiotherapy because it enhanced iron uptake and storage 10,64. Based on the significant increase in intratumoral IC and serum ferritin levels in HPSCC tissues, we suspected that YTHDF1 might also ameliorate the response to radiation because of TFRC-induced iron accumulation. Then, we reassessed the association between YTHDF1 staining and recurrence-free survival in patients treated with adjuvant CCT/RT. As expected, YTHDF1 did not differ between patients who receives CCT/RT treatment and those who did not (Figure , but TFRC expression and serum ferritin levels were significantly upregulated in patients who received adjuvant CCT/RT treatment (Figure High YTHDF1 and TFRC levels were associated with poor prognosis in patients receiving adjuvant CCT/RT treatments (Figure Taken together, these data suggest that YTHDF1 enhances TFRC expression in HPSCC through an m6A- dependent mechanism.

Discussion

Although the m6A modification has been researched for decades, the potential involvement of m6A modification on RNA is poorly defined in the context of human HNSCC and iron metabolism. Data from epidemiological studies and in vivo and in vitro models have corroborated the critical role of iron in HNSCC 21,22,65,66. Because of its high invasive capacity at all subsites, HPSCC was the focus of our study. Here, through bioinformatics analysis, sample detection, transcriptome-wide high-resolution m6A-seq, RNA-seq and RIP-seq assay analyses, and cell biological analyses, we demonstrated that the m6A reader YTHDF1 regulates HPSCC iron uptake via m6A-modified TFRC mRNA, thereby regulating tumorigenesis and tumor proliferation. Moreover, high expression of YTHDF1 might also be linked to radiation resistance by promoting intratumoral iron accumulation (Figure . Through bioinformatics analyses, we demonstrated that the expression of iron-related genes, especially FTH1 and TFRC (which regulate ferritin and the intratumoral iron levels, respectively), were elevated in cancer tissues. Noticeably, among the m6A readers, YTHDF1 had remarkably different expression levels between HNSCC tumor samples and normal control samples. Both upregulated transcription and amplified DNA copy number of YTHDF1 have been observed in various human cancers, making this protein a robust and reproducible cancer biomarker 37. YTHDF1 was also found to maintain intestinal stem cell features during Wnt-driven intestinal regeneration and tumorigenesis 37. Our findings supported the notion that YTHDF1 is the major effector of m6A during tumorigenesis. In addition, we firstly demonstrated the relationship between iron metabolism and YTHDF1.

Despite exhibiting generally comparable m6A RNA methylation levels, HPSCC patients with higher serum ferritin and intratumoral iron levels show highly upregulated m6A reader YTHDF1 expression compared to those patients with lower serum ferritin and intratumoral iron. Combining the results from GO, KEGG and GSEA demonstrated that cellular iron homeostasis and cell proliferation were significantly enriched in response to YTHDF1 knockdown in HPSCC cells. In vitro and in vivo experiments indicated that downregulating YTHDF1 suppressed cancer growth, colony formation, and immigration and reduced intracellular iron content, Fe2+ and ROS levels in HPSCC cells. Importantly, YTHDF1 mutant plays dominant-negative effect to inhibit the function. The current data indicate that m6A can be recognized by proteins that contain a YTH (YT521B homology) domain 56. Therefore, YTH domain functional deletion would interrupt the binding of YTHDF1 to its m6A- methylated targets in our study, deletion of the YTH domain of YTHDF1 blocked YTHDF1-induced iron metabolism in HPSCC cells. Overexpression YTHDF1 using YTHDF1-WT virus xenograft mouse models increased the tumorigenesis in vivo (Figure . But overexpression YTHDF1-mut did not have such effect (. In addition, DFP (intracellular iron chelators) treatment significantly blocked YTHDF1-induced iron uptake and tumorigenesis in vivo (Figure . These results elucidated that YTHDF1-induced HPSCC iron metabolism depends on its RNA m6A function, consistently with other data showing that during tumor development, YTHDF1 activates the translation of methylated mRNAs for sufficient protein production 67. A recent study by Shi, Y et al investigated the function of YTHDF1 in non-small-cell lung cancer (NSCLC) under hypoxia 31. Downregulated YTHDF1 in NSCLC has a protective role against cellular stresses, such as hypoxia and DDP treatment, through the Keap-Nrf2 axis. We found that YTHDF1 increased ROS levels upon regulating iron metabolism, expanding the function of YTHDF1 in intracellular ROS regulation.

YTHDF1 was only once reported to bind to the m6A site in CDS region of Snail mRNA and enhance its translation by recruiting the translation elongation factor eEF-2. They also firstly assessed the most common sites 3'UTR, but they found that it was not involved in m6A modification regulated target gene expression. They speculated that YTHDF1 could function via regulating translation elongation as well as translation initiation. Here in our study, m6A methylation on 5'UTR and 3'UTR are the main locations of m6A modification. YTHDF1 increases the expression of the TFRC gene mainly at the 3′UTR and 5′UTR, which in turn leads to the upregulation of TFRC at the protein level but not at the mRNA level. Notably, mutagenesis assays indicate that the m6A sites in the 5'UTR and 3'UTR of TFRC mRNA are essential for YTHDF1 to post-transcriptionally regulate TRFC expression. This finding was further supported by RIP-qPCR and Western blot data. Polysome profiling assay supported that YTHDF1 regulates TFRC translations efficiency. mRNA transcripts with m6A modifications have differing fates primarily based on their interaction with different m6A readers, which contain a YTH domain, IGF2BPs or eukaryotic initiation factor 3 (eIF3) 56. YTHDF1 did not have a substantial effect on mRNA stability, as indicated by an analysis of DF1-depleted cells 30. In contrast to YTHDF1, YTHDF2, YTHDF3, and YTHDC2 tend to reduce gene expression by promoting m6A-modified mRNA decay, while IGF2BPs regulate gene expression by promoting mRNA stability 68. In this study, we did not observe a change in m6A modification in TFRC mRNA upon alterations in YTHDF1 expression. In addition, the protein level but not the mRNA level of TFRC declined upon YTHDF1 knockdown. Importantly, our polysome profiling assay showed that YTHDF1 knockdown resulted in significant lower TFRC mRNA in translation fractions, supporting the notion that YTHDF1 regulates protein synthesis.

Whether YTHDF1 interacts with other m6A writers and erasers in HPSCC cells remains unknown and needs further investigation. We speculated that YTHDF1 enhanced TFRC expression by promoting translation, which was revealed by an analysis of previously reported transcriptome- wide m6A, RIP and CLIP data (GSE78030, GSE92021, GSE78507, and GSE86214). YTHDF1 was shown to promote translation elongation and translation initiation in multiple studies 32,35,53,69. Further studies should focus on initiation factor identification and determine how YTHDF1 promotes translation upon the binding of these initiation factors 56.

We found that high YTHDF1 and TFRC levels were associated with poor prognosis in patients treated with CCT/RT. Interestingly, increased ferritin concentrations were also observed in this group. A previous study demonstrated that cancer cells preferentially exhibit elevated expression of TFRC, which can increase iron uptake 48. Logically, by regulating TRFC-induced iron uptake, YTHDF1 is expected to cause CCT/RT resistance in HPSCC, thereby causing poor prognosis. Thus far, we have shown that YTHDF1 may function as an oncogene by enhancing TFRC protein expression in HPSCC cells, which results in increased iron uptake. Ultimately, imported iron enters the bioactive LIP for proliferation and metabolic purposes 10. However, we cannot definitively point to a direct link between YTHDF1 and LIP. The amount of LIP is detected by post-transcriptional mechanisms of IRP1 and IRP2 62. In our pilot experiments, we failed to observe binding between YTHDF1 and IRP1/IRP2 in our culture systems (data not shown). However, we hypothesized that TFRC is a crucial target gene for YTHDF1-mediated promotion of iron metabolism in HPSCC. Previous studies have demonstrated that TRFC-induced iron uptake is the most important mechanism by which cancer cells internalize iron 70,71. Here, we first revealed that YTHDF1 enhanced TRFC expression via m6A recognition to promote iron metabolism. Multiple strategies for anticancer therapies, including utilization of TFRC-mediated cytotoxic drug conjugates and iron chelators, have been designed to disrupt the intracellular iron balance needed to meet the urgent metabolic demand of cancer cells 62. In our study, DFP treatment significantly blocked YTHDF1- induced HPSCC growth and colony formation in vitro and in vivo. Fe2+ is an essential cofactor for the delivery of oxygen to cells, protecting cancer cells from hypoxia-induced stress. However, excessive intracellular iron accumulation can cause cells to experience extreme oxidative stress and may induce tumor death. Iron-dependent cancer death, also known as “ferroptosis”, can be inhibited by iron chelators. The functions and mechanisms of TFRC in cancer cell ferroptosis have been observed, but their relevance is still unclear 72. In other words, TFRC regulation may mark a watershed event during iron accumulation. Depending on primary recognition by the various m6A readers, mRNA transcripts with m6A modifications undergo different fates 56. Therefore, understanding the functional mechanism of YTHDF1 may enable the reconstitution of TFRC activity, thus enabling the development of specific therapeutic targets according to different intratumoral iron levels.

Conclusions

In summary, we found that YTHDF1 is closely correlated with iron metabolism and tumor progression in HPSCC. Mechanistically, TFRC was identified as a direct target of YTHDF1 in HPSCC cells and promoted iron metabolism, thereby increasing tumor growth and proliferation. From a therapeutic perspective, targeting YTHDF1 and TFRC-mediated iron metabolism may be a promising strategy for HPSCC.

Supplementary figures and tables.

Click here for additional data file.