Intratumor microbiota Delftialacustiris correlated with METTL3 to control development of papillary thyroid carcinoma

Abstract

Background

Papillary thyroid carcinoma (PTC) exhibits epitranscriptomic dysregulation and tumor-resident microbiota alterations, yet potential links between these processes remain unclear. We investigated whether intratumoral bacterial features are associated with METTL3-dependent programs in PTC.

Methods

We profiled intratumoral microbiota from PTC and adjacent thyroid tissues and integrated these data with tumor gene-expression analyses. Candidate genes associated with the abundance of Delftia lacustris were prioritized by a prespecified scheme (FDR rank, effect size/consistency, pathway plausibility, and experimental tractability). Functional assays (gain/loss of function, rescue) tested METTL3 and downstream DMTF1 effects on PTC cell phenotypes (proliferation, colony formation, apoptosis/cell cycle, migration/invasion). Appropriate statistical tests with multiple-comparison correction were applied; exact P values and effect sizes are reported in the figure legends and Source Data.

Results

Reduced intratumoral D. lacustris abundance was inversely associated with METTL3 expression and with PTC occurrence. Based on prioritization, METTL3 was selected for mechanistic interrogation. METTL3 overexpression enhanced proliferation and clonogenicity, whereas METTL3 knockdown produced the opposite effects; DMTF1 re-expression rescued METTL3-knockdown phenotypes, delineating a METTL3-DMTF1 proliferative axis in PTC cells. These tumor-intrinsic results, together with the microbiome association, nominate D. lacustris as a tumor microenvironment (TME)–coupled biomarker of a METTL3-high epitranscriptomic state.

Conclusions

We identify a METTL3-DMTF1 axis that promotes PTC cell growth and show that lower intratumoral D. lacustris abundance is associated with higher METTL3 levels and with PTC. Given the cross-sectional design, these relationships should be interpreted as associative (not causal). The data generate testable hypotheses for microbiome–epitranscriptome crosstalk in the PTC TME, motivating spatial, longitudinal, and perturbational studies to resolve directionality.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12935-026-04183-9.

Introduction

In recent years, the existence of intratumor microbiota has been reported in some studies [1]. These microbiota are a group of microorganisms that live in the tumor microenvironment and may influence tumor development and progression [2]. The most common intratumor microbiota reported are bacteria, including but not limited to Fusobacterium nucleatum, Bacteroides fragilis, and Escherichia coli. Some researchers suggest that these bacteria may contribute to the stimulation of the host immune response or the production of metabolites that promote tumor growth [2]. However, there is still much to be explored in this field, and further research is needed to fully understand the role of intratumor microbiota in tumor biology [3]. The complex ecosystem within a patient’s tumor is composed of malignant cells and a network of normal cells, including fibroblasts, endothelial cells, and immune cells. In addition, the bacterial community has been identified as a component of the tumor microenvironment (TME), and can directly or indirectly affect cancer progression. Early studies have shown that certain bacteria present in tumor tissue may impact the occurrence and development of tumors. Similarly, fungi in tumor tissue have also been studied, with evidence indicating their association with tumor occurrence and development. For example, some fungi can promote tumor growth by inhibiting the immune system, while others can produce toxins harmful to the human body. The tumor microbiota has been acknowledged as an important participant in potential cancer metastasis, possibly involving various stages of cancer metastasis. Using techniques like metagenomic sequencing, researchers have identified the composition of microbial communities in different tumors, such as colon, prostate, and pancreatic tumors, and have discovered the presence of microorganisms like bacteria and mycoplasma. The tumor microbiota can affect tumor growth and development by creating the tumor microenvironment, generating gene toxins, inducing chronic inflammation, and also plays a crucial role in the process of chemotherapy and other therapeutic behaviors.

Thyroid nodules are prevalent disease in global patients, which has a large proportion of individuals with malignant cancer [4]. Papillary Thyroid Carcinoma (PTC) is the most prevalent subtype of thyroid cancer, according to epidemiological research [5]. Thyroid cancer of this kind has a low malignancy level and a five-year survival rate of 90% after radical surgery. Although PTC has a fair prognosis, it has a high cancer incidence rate and can affect people of any age. Resolving the underlying molecular mechanisms of PTC oncogenesis and development can serve as a theoretical foundation for new therapeutic strategies and disease prevention [6].Recent studies have reported that there might be intratumor microbiota in papillary thyroid carcinoma (PTC), but this area requires further investigation to confirm. One study revealed that PTC samples showed significantly higher bacterial load and diversity compared to benign thyroid nodules and normal thyroid tissue. The most common bacterial species detected in PTC samples were Propionibacterium acnes, Staphylococcus epidermidis, and Lactobacillus gasseri. However, the presence of bacterial DNA in tumor samples does not necessarily indicate that bacteria are actively contributing to the development or progression of PTC. The presence of microbiota within the thyroid microenvironment has been suggested to play a role in regulating the immune response, and further research is needed to explore the possible mechanisms of bacterial effects on PTC progression. It is important to note that medical questions should be addressed by healthcare professionals.

RNA m⁶A methylation has been discovered as a novel pivotal mechanism to govern the oncogenesis and progression of cancer in recent research [7–10]. This type of methylation is the most prevalent RNA epigenetic modification in eukaryotes. RNA m⁶A methylation has been discovered to be a target for tumor therapy and medication discovery in a number of clinical investigations. One of the causes of many diseases is aberrant expression of regulatory factors of RNA m⁶A in different tissue and organ during development and environment response. METHYLTRANSFERSE 3/14 (METTL3/METTL14), and WT ASSOCIATED PROTEIN (WTAP), which form a methyltransferase complex protein and altered RNA methylation status on N6 methyl adenosine, have been discovered as methylation modified methyltransferases that control RNA m6A in the human genome. METTL3 has been characterized as a vital factor in many cancers. In breast cancer, the METTL3/IGF2BP3 signaling pathway inhibit cancer immune microenvironment by enhance the m6A of PD-L1 mRNA [8]. Exosomal circLPAR1 suppresses BRD4 via the METTL3-eIF3h relationship in colorectal cancer detection and carcinogenesis [11]. YTHDF2 promotes the malignant evolution of glioma by facilitating UBXN1 mRNA degradation by detecting METTL3-mediated mA alteration and activating NF-B [12]. METTL3 enhances ESCC formation by lowering APC expression, which is mediated by YTHDF binding to APC mRNA N-methyladenosine [13]. There is also evidence supported the role of METTL3 in PTC development [14]. METTL3 is a critical factor that has been discovered to influence thyroid cancer cell migration by altering the TCF1/Wntpathway and is a regulator of thyroid cancer development.

However, the relationship between METTL3 and microbiome in thyroid cancer is still not fully unresolved, and its gene-level regulation requires validation in clinical cohorts. In this study, the bacteriaof thyroid cancer were examined, and the major species associated with the expression of the METTL3 gene were identified. METTL3’s m⁶A alteration of important genes was confirmed, and its probable mechanism of thyroid cancer regulation was investigated.

Materials & methods

Patients

Inclusion criteria: (1) PTC was confirmed by pathology or cytology; (2) First diagnosis and treatment; (3)Did not receive radiotherapy, chemotherapy and immunotherapy; (4) Age ≥ 18 years old; (5) Clinical resources Material integrity; (6) Sign informed consent. Exclusion criteria: (1) combined thyroid Hyper function, diabetes and other endocrine system diseases; (2) Merge other departments Malignant tumor; Concurrent malignancies in other organs; (3) Persons with mental illness; (4) Pregnancy or lactation woman.

RNA extraction

The extraction reagent used to extract total RNA in this experiment is Nucleozol (Biotech, Shanghai, China, Catalog No.: 740404.200). The operation steps are as follows: take a new 1.5 mL EP tube from 100 mg tissue of patients cryopreserved in − 80 ℃ refrigerator; Join 500 µL of Nucleozol, blow it with a pipette gun or shake it violently to fully mix the Nucleozol with the peripheral blood sample, and split it at room temperature for 15 min; Join 200 µL of DEPC water (RNase free) into the lysate, shake vigorously for 15 s and incubate at room temperature for 15 min; 12,000 g centrifugation for 15 min. after centrifugation, suck the supernatant into a new centrifuge tube, and do not absorb the bottom sediment; Add isopropanol in the same proportion as the supernatant absorbed in the previous step, mix well and place for 10 min; 12,000 g (10 min) 20℃ centrifugation, and the supernatant was discard; add 500 µL ethanol(75%), gently blow to clean the precipitate, then centrifuge at room temperature under 8000 g centrifugal force for 3 min, and discard the supernatant; Repeat step; Open the EP tube cover to fully volatilize the ethanol, and then add an appropriate amount of DEPC water to resuspend the RNA; The obtained RNA solution was stored at − 80℃ for < 6 months.

DNA extraction

Sample collection and preparation - collect the sample, and depending on the source material, it may need grinding, homogenization, or cell lysis; Cell lysis - homogenize the sample to break down the cell wall, cell membrane, and release the DNA; Protein removal - use protease to remove proteins from the sample; DNA separation - use isopropanol or ethanol to separate DNA from the solution; DNA purification - use washing steps to remove impurities and enhance the purity of the DNA; DNA quantification and quality assessment - use spectrophotometry or gel electrophoresis to determine the concentration and quality of DNA.

Reverse transcription

Hiasen’s Hifair Ⅱ 1 st strand cDNA synthesis Kit reverse transcription Kit (Biotech, Shanghai, China, Catalog No.: 11121es60)was used in this experiment ^®, the operation steps are as follows: Remove the residual genomic DNA in 200 according to 10ul system µ Prepare the mixture (RNase free ddH₂O, 5) in the PCR tube of L × GDNA digester buffer, gDNA digester, total RNA), and then incubated on the PCR instrument at 42℃ for 2 min; Then configure 20ul reverse transcription system (RNase free ddH2O, 5 × Hifair ^® Ⅱ Buffer plus, Enzyme Mix, Oligo (dT)18) mix the above system and put it into the PCR instrument at 25℃ for 5 min; 42℃, 30 min; 85℃, 5 min.

Quantitative PCR

The kits used in our laboratory are UltraSYBR mixture (product No.: cw0957h), and the operation steps are as follows: the primers are designed mainly by NCBI website, and the appropriate primers are selected according to the primer specificity and parameter characteristics designed by the website, and then sent to Biotechnology (Shanghai) Co., Ltd. for synthesis; Configure PCR reaction system (20) µL system). After the preparation of the above system, put it into the PCR instrument at 95 ℃ 10 min, annealed/extended 60 ℃ 1 min, melting 95℃ 30 s, 40 cycles. After the completion of qPCR program, observe the amplification curve and melting curve, remove the abnormal data value, and export the CT value for data analysis. Primers were designed using NCBI Primer-BLAST and synthesized by Biotechnology Co., Ltd. (Shanghai). PCR primers list was shown in Table S1. β-actin was used as housegene.

CCK-8 assay

Preparation of cell suspension: collect the cell suspension in good condition after miRNA mimic transfection into the centrifuge tube, centrifuge 350 g for 5 min, supernatant was discarded, then add an appropriate amount of culture medium to resuspend; Cell planking: count the cell suspension obtained in the previous step and prepare a cell suspension with a concentration of 60,000 − 10,000 cells/ml; 6 multiple holes are set for each group, and 100 holes are added for each hole µL cell suspension, add 100% into the six holes in each column before the experimental group and NC group µ The complete medium of l was used as the blank control, and the same volume of PBS was added to a circle of holes around the cells to prevent the evaporation of the edge cell medium; Absorbance measurement: put the planted plate back into the incubator for culture. After the cell paste is stable, add 10 in each of the three columns of blank group, control group and experimental group on day 0 µL of CCK8 solution, cultured in a 37 °C for 2 h, and then measured absorbance at the wavelength of 450 nm with a microplate reader to obtain the OD values of each group on day 0, and the other columns were measured at 24 h, 48 h and 72 h respectively; Result analysis: import the obtained data into prism software for analysis, and then get the relevant results of the impact on cell proliferation.

Colony formation assay

Take the cells growing in logarithmic phase and count the seed plate. Inoculate on a 6-well plate according to 1000 cells/well. The culture system is 100 µL. the culture is routinely cultured for 1–2 weeks. When cell clusters are observed in the culture plate by naked eyes, the culture is terminated immediately. Remove the original culture medium twice and wash it gently to avoid the formation of PBS. Add an appropriate amount of 4% polymethanol to fix for 20 min, pour out the fixing solution, and then add an appropriate amount of crystal violet staining solution for dyeing. The dyeing time is about 15–30 min. Turn on the faucet slightly to form slow running water, put the culture plate under the running water, gently wash off the staining solution, put it in the air, and take photos after it is dry to determine the number of cloned cells.

Cell cycle assay

Take the cells in logarithmic growth period and press 1 × 10^6 cells/ml inoculate 24 well plate with 1 ml or 6 well plate with 2 ml, carry out the required treatment (such as adding drugs), terminate the culture after a specific time, and carry out the next experiment. Cell fixation: centrifuge at 800 rpm for 5 min, collect cell precipitation, discard supernatant, wash twice with precooled PBS, add precooled 75% ethanol, and fix at 4℃ for more than 4 h. Cell staining: centrifuge at 1500 rpm, wash once, add 400ul ethidium bromide (PI, 50ug/ml), 100ul RNase A (100ug/ml), and incubate at 4℃ in dark for 30 min. With standard program, flow cytometry was used to determine, and 20,000–30,000 cells were generally counted. The results were analyzedby software ModFit.

Statistical analysis

All cellular experiments were performed in at least three independent biological replicates unless otherwise indicated. Data are presented as mean ± SD (or mean ± SEM where noted). For two-group comparisons we used two-tailed unpaired Student’s t-tests; if normality, we used the Mann–Whitney U test. For ≥ 3 groups (one factor) we used one-way ANOVA with Tukey’s post hoc test; when assumptions were not met, Kruskal–Wallis with Dunn’s post hoc test was applied. For two-factor designs, we used two-way ANOVA, reporting main effects and interactions; for longitudinal viability curves we additionally summarize group differences by area-under-the-curve (AUC) with one-way ANOVA. Proportions were compared by one-way ANOVA (or Kruskal–Wallis) across groups, with planned pairwise contrasts corrected as below.Multiple comparisons were controlled by Benjamini–Hochberg false-discovery rate (FDR), with q < 0.05 considered significant unless otherwise specified (P values are also reported). Analyses were performed in GraphPad Prism (v9/10) and R (v4.x).

Results

The abundance of delftialacustirisis correlated with METTL3 in PTC

The DNA sequences of potential intratumoral bacteria in PTC were analyzed from the database. The findings demonstrate that Bacillus cereus and Staphylococcus aureus were the primary bacteria coexisting in PTC, and both were also found in the normal thyroid tissues (Fig. 1A). Bacillus cereus was observed to be more prevalent compared to other types. Nonetheless, no notable difference was observed in biodiversity between the normal thyroid tissues and PTC, as measured by the Simpson diversity index (Fig. 1B). We analyzed in detail the interspecies differences of the top 20 representative bacteria. The heat map shows that some species of bacteria have differences in abundance (Fig. 2A). Especially, Delftialacustiris has significantly (P = 0.0057) lower abundance in PTC than in normal thyroid tissues (Fig. 2B).

Fig. 1.

The structure of intratumoral bacteria in PTC.(A) The relative abundance of top 20 bacteria species in PTC.N group indicates para-tumor control; T indicates tumor. (B) Simpson diversity. Note that no significant differences

Fig. 2.

The abundance of Delftialacustiris is correlated with METTL3 in PTC. (A)The heatmap of top 20 bacteria species in PTC. (B) The relative abundance of Delftialacustirisin para-tumor control and tumor. (C) The correlation between the abundance of Delftialacustirisand RNA m6A regulators. (D) The correlation between the abundance of Delftialacustirisand METTL3

After obtaining the results mentioned above, we partitioned 30 clinical samples of PTC into two distinct segments. The first part underwent PCR analysis for the evaluation of the relative abundance of Delftialacustiris, whereas the second part underwent RNA extraction to measure the expression levels of m6A methyltransferase and demethylase genes, including METTL3, METTL14, WTAP, FTO and ALKBH5 (Fig. 2C). Subsequently, we examined the correlation between the abundance of Delftialacustiris and the expression levels of m6A regulatory factor genes, and our findings indicated a significant negative correlation (r=−0.48, P = 0.013) between the abundance of Delftialacustiris and the expression levels of METTL3 (Fig. 2D).

DMTF1 expression are highly co-expressed with METTL3

In TCGA thyroid cancer transcriptome database, comparing the gene expression between cancer and adjacent cancer, it was found that DMTF1 and METTL3 were up-regulated in cancer (Fig. 3A), and DMTF1 was one of the genes with top 10 correlation with METTL3 expression (Supplemental Table 1). According to the figure, the correlation coefficient between the two is R = 0.87, which is very significant (P = 0) (Fig. 3B). We collected cancer and adjacent tissues from 30PTC cases and detected the transcriptome. The results demonstrated that the level of DMTF1 and METTL3 were also enhanced in cancer, and the correlation coefficient between them was r = 0.52. The expression level of the both genes was further determined by qRT-PCR, the expression in and adjacent to cancer is consistent with the change trend obtained by the above method (Fig. 3C), and the correlation factor was r = 0.52 (Fig. 3D).

Fig. 3.

METTL3 is co-expressed with DMTF1 in PTC. (A) The expression level of METTL3 and DMTF1 in tumor and para-tumor, and (B) Scatter plots show co-expression between METTL3 and DMTF1according to TCGA database. (C) The expression level of METTL3 and DMTF1 in tumor and para-tumor, and (D) Scatter plots show co-expression between METTL3 and DMTF1 according to our clinical samples. An asterisk (*) indicates statistical significance at P < 0.05

METTL3 inducedm⁶AlevelonDMTF1 mRNA

Because the expression levels of METTL3 and DMTF1 are consistent, there may be a regulatory relationship between them. We hypothesized that METTL3 can regulate the expression of DMTF1 mRNA by regulating the m⁶A level. Therefore, in TPC1 cells, we changed level of METTL3 by transient overexpression and silencing, and checked the m6A modification on DMTF1 mRNA by mIP qPCR method. The results showed that overexpression of METTL3 significantly increased the m6A modification level of DMTF1 mRNA. After silencing METTL3, the m⁶A modification level of DMTF1 mRNA decreased significantly (Fig. 4A). In addition, through RiP assay, it was found that DMTF1 mRNA was more enriched in METTL3 antibody group than in IgG antibody group (Fig. 4B). The above results suggest that METTL3 can regulate its m⁶A level by binding DMTF1 mRNA in TPC cells.

Fig. 4.

METTL3 increases PTC cell growth. (A) METTL3 regulates m6A level on DMTF1. (B) METTL3 interact with DMTF1 mRNA. (C and D) CCK-8 assay detect the cell growth of TPC-1 and BP-RAP with overexpressed METTL3. (E and F) cell colony assay and cell cycle of PTC cell with overexpressed or silenced METTL3. An asterisk (*) indicates statistical significance at P < 0.05

METTL3 increase the PTC cell growth

Aim to study the roleof METTL3 in PTC, we changed the expression of METTL3 in TPC-1 and B-CPAP cell lines by transient expression, and then observed the cell growth. The results of CCK-8 experiment displayed that ox-METTL3 significantly enhanced the proliferation of these two cells (Fig. 4C and D). Through the detection of cancer cell cloning experiment, it was suggested that ox-METTL3 significantly induced the number of clones, more than 1.5 times compared with the control group; Si-METTL3 significantly reduced the number of clones, more than twice that of the control group (Fig. 4E). Then we detected the cell cycle of each group in TPC-1. The results displayed that the S phase increased in the ox-METTL3group and decreased in the silencing group (Fig. 4F).

DMTF1 increase the PTC cell growth

In order to analyze whether the effect of DMTF1 on PTC cells is consistent with METTL3, we used the same method to change the expression of dmtf1 in TPC-1 and B-CPAP cell lines, and then observed the cell growth. The results revealed that the cells in the ox-DMTF1 significantly increased the proliferation rate compared to control (Fig. 5A and B). Through the detection of cancer cell cloning experiment, it was found that the overexpression of DMTF1 group significantly increased the number of cell clones; Silencing DMTF1 significantly decreased the number of cell clones (Fig. 5C). Then we detected the cell cycle of each group. The results showed that the S-phase cells in the overexpression group were more than control group, on the contrary, those in the silencing group were less (Fig. 5D).

Fig. 5.

DMTF1 induces PTC cell growth. (A and B) CCK-8 assay detect the cell growth of TPC-1 and BP-RAP with overexpressed DMTF1. (C and D) Overexpressed or silenced DMTF1 were check by cell colony assay and cell cycle. An asterisk (*) indicates statistical significance at P < 0.05

METTL3 regulate PTCcellvia controlling DMTF1expression

The previous experimental results show that the expression of DMTF1 is consistent with METTL3. Both genes promote the growth of PTC cells, and METTL3 can regulate the m6A level of DMTF1 mRNA and then affect the expression. Based on the above data, we hypothesized that METTL3 mainly affects the growth of PTC cells by regulating the expression of dmtf1. Similarly, we used the method of recovery experiment in TPC-1 and B-CPAP cell lines. The results showed that si-METL3 + ox-DMTF1 group increased cell proliferation compared with si-METTL3 group (Fig. 6A and B). At the same time, cell cloning experiments showed that the number of cell clones in si-METL3 + ox-DMTF1 group increased compared with si-METTL3 group (Fig. 6C). The cell cycle was detected by flow cytometry. It was found that S-phase cells in si-METL3 + ox-DMTF1 group was higher than that in Si-METTL3 group (Fig. 6D).

Fig. 6.

METTL3 regulate PTC cell via DMTF1. (A and B) cell proliferation of TPC-1 and BP-RAP with silenced METTL3 was recovered by overexpression DMTF1. (C and D) Silenced METTL3 plus DMTF1 overexpression were check by cell colony assay and cell cycle. An asterisk (*) indicates statistical significance at P < 0.05

Discussion

In this study we observed that lower intratumoral abundance of Delftia lacustris was negatively associated with papillary thyroid carcinoma (PTC) and with higher METTL3 expression, and we further identified a METTL3–DMTF1 axis that promotes PTC cell proliferation. These findings are hypothesis-generating and do not establish the direction of effect between D. lacustris and PTC. Two non-exclusive models remain plausible: (i) a microbe-to-tumor model, in which loss of D. lacustris (or its metabolites) permits or reinforces pro-tumorigenic epitranscriptomic states such as METTL3–DMTF1 activation; and (ii) a tumor-to-microbe model, in which evolving tumor ecology secondarily reduces D. lacustris detectability. Because our design is cross-sectional and relies on DNA-based detection from bulk tissue, it cannot distinguish these possibilities or prove that changes in D. lacustris occur upstream of tumor initiation.

Several factors could confound the observed association, including sampling heterogeneity, variable tumor cellularity, and technical biases that affect low-biomass microbiome measurements. To mitigate over-interpretation, we deliberately refrain from causal language and instead emphasize a correlative link among D. lacustris, METTL3, and DMTF1 [15, 16]. Mechanistically, our cell-based data support the biologic plausibility that METTL3 drives PTC growth via DMTF1, while the microbiome analysis nominates D. lacustris as a putative biomarker of this axis rather than a proven driver. Future work that could clarify directionality includes: longitudinal sampling prior to surgery, spatial profiling to co-localize bacteria and METTL3-high tumor regions, and perturbational models that test whether D. lacustris modulates METTL3–DMTF1 activity. Although such experiments are beyond the scope of this study, our data provide a clear testable hypothesis and a prioritized pathway to evaluate. Our study is limited by (i) the cross-sectional design; (ii) DNA-based detection that does not confirm microbial viability; and (iii) lack of comprehensive clinicopathologic covariates that might influence both microbiota and METTL3 levels. These limitations preclude causal inference and motivate the cautious interpretation adopted here. Within these constraints, D. lacustris abundance may serve as a contextual marker of an epitranscriptomic state characterized by high METTL3 and DMTF1 in PTC. Clinically, this raises the possibility that tumor-resident microbiota profiles could complement molecular assays when stratifying patients by METTL3-related biology; however, validation in independent cohorts and longitudinal designs is required.

Our data identify reduced intratumoral Delftia lacustris as a salient feature in PTC and show an inverse association with METTL3 expression. Because the tumor microbiome is embedded within a dynamic tumor microenvironment (TME), we consider mechanistic couplings that may link D. lacustris abundance to the METTL3–DMTF1 epitranscriptomic program [17]. D. lacustris could restrain METTL3 activity indirectly via microbial metabolites or pathogen-associated molecular patterns (PAMPs) that modulate stress, MAPK, or NF-κB pathways upstream of m⁶A regulation. Loss of D. lacustris would then remove this brake, enabling METTL3-DMTF1 activation and proliferation. This model predicts that conditioned media or purified metabolites from D. lacustris reduce METTL3 levels, global m⁶A abundance, or DMTF1 expression in PTC cells, and that these effects are ligandable. A METTL3-high, DMTF1-dependent tumor state may reshape the TME, creating conditions that are less permissive for D. lacustris persistence. This model predicts spatial anti-correlation between METTL3-high tumor niches and D. lacustris signals, with accompanying immune/stromal features that disfavor this bacterium. Upstream TME factors (hypoxia, redox, local immunity, thyroid tissue biochemistry) could simultaneously lower D. lacustris detectability and raise METTL3. This predicts that microenvironmental scores co-vary with both variables even when considered jointly. These models move beyond a simple correlation by situating D. lacustris as a TME-linked protagonist of the biology we observe. While our study is cross-sectional and cannot establish directionality, it yields actionable hypotheses: (i) spatial co-localization (16 S-FISH or probe-based imaging with METTL3/DMTF1 IHC) to test niche-level anti-correlation; (ii) co-culture/conditioned-media assays to probe microbe→tumor effects on METTL3–DMTF1; (iii) metabolomics of tumor regions stratified by D. lacustris signal; and (iv) integrated modeling (e.g., partial correlations or mediation analysis) to examine whether TME features account for the D. lacustris–METTL3 association. Consistent with our Abstract, we therefore interpret D. lacustris not merely as a correlate of METTL3 but as a microenvironment-coupled biomarker that nominates specific tumor–microbe–epitranscriptome circuits for future testing.

RNA m⁶A modification is an evolutionarily conserved RNA modification in eukaryotes, which refers to a methyl at the N6of adenoglycine. In eukaryotic mRNA, on average, each transcript has 2–3 m6A modification sites, and 0.1% − 0.4% of adenylate is modified by m⁶A [18]. The biological function of this modification is mainly to affect RNA transcription, processing, translation and metabolism. Three main components of methyltransferase complex have been clarified: methyltransferase like 3/14 (METTL3/14), and Wilm’s tumor 1-associated protein (WTAP). METTL3 is the core component, METTL14 and METTL3 bind to form a stable heterodimer. These three components have been confirmed to be involved in the pathogenesis of cancer. Studies have shown that METTL3 is significantly up-regulated in human hepatocellular carcinoma (HCC), acute myeloid leukemia (AML) and multiple solid tumors [19–22]. Knockdown or knockout of METTL3 can significantly inhibitthe proliferation and migration of tumorcells, and affect m6Aa modification through a variety of mechanisms to promote cancer. In bladder cancer, METTL3promotesthe processing of pri-miR221/222 by recognizing DGCR8 (DI George CIR tic region 8) [15]. Mature miR221/222 inhibits tumor suppressor gene PTEN and promotes cancer cell growth; In cutaneous squamous cell carcinoma (CSCC), up regulation of METTL3 promotes the expression of Np63, thereby promoting the proliferation of CSCC cells and tumor growth [23]. Of course, some cancer studies have found that METTL3 is not entirely cancer promoting. In pancreatic cancer cells, the knockdown of METTL3 has no effect on cell proliferation, but will improve the sensitivity of cells to anticancer drugs [24, 25]. Even in some studies, METTL3 has a controversial role in cancer. In glioblastoma stem cells (GSCs), knock-down METTL3 could enhanced cell growth, self-renewal and tumor progression of GSCs [26]. On the contrary, the expression of METTL3 was up-regulated in human glioblastoma. By binding with Sox2 mRNA in 3’UTR, m6A modification was induced. Silencing METTL3 could inhibit the expression of Sox2 and enhance the sensitivity of tumor cells to γ. The in vitro sensitivity of radiation inhibits the growth of mouse glioblastoma tumor, thus playing a carcinogenic role [27].

Multiple studies in PTC have reported elevated METTL3 expression in tumor tissue relative to adjacent non-tumor thyroid and associations with adverse clinicopathologic features (e.g., higher stage and lymph-node metastasis). Functional perturbations in established PTC cell lines consistently show that METTL3 promotes malignant phenotypes—including increased proliferation/colony formation, migration/invasion, and resistance to apoptosis—while METTL3 knockdown produces the opposite effects; in vivo xenograft work likewise supports a pro-tumor role for METTL3. Mechanistically, these PTC studies converge on m^6A-dependent regulation of oncogenic programs: METTL3 modifies specific transcripts, with consequences for mRNA stability and/or translation via readers such as YTH-family and IGF2BP proteins, and impacts pathways commonly dysregulated in thyroid cancer Collectively, the existing PTC literature positions METTL3 as a key epitranscriptomic amplifier of tumor progression rather than a benign bystander. The reasons for the above disputes may be related to the different target mRNA labeled by m6A. In our study, METTL3 has the same effect on thyroid cancer cells and hepatocellular carcinoma, and promotes the growth of tumor cells. In addition, it is reported that it can also promote the invasion of thyroid cancer cells. Because there are many m⁶A methylation modified genes,METTL3 can affect a wide range of genes. How to find the key factors in many genes is the basis of in-depth research. Through the research of the database and its own cases, we found the genes that are consistent with the expression of mettl3, so as to explore the most possible influence pathway, which provides a certain idea for further research in the future.

DMTF1 is a cyclin D binding protein which was characterized as a tumor suppressor. It plays as a transcription factor to activate ARF/P53 signaling and regulate cell proliferation. In many cancers, studies have proved that DMTF1 has critical role and is a potential diagnosis biomarker. miR-675-3p turn down the level of DMTF1 mRNA in colorectal cancer cell to promote cell proliferation [28].In renal cell carcinoma, which is different with colorectal cancer cell, DMTF is silenced by miR-6838-5p to enhance cell proliferation [29]. Another miRNA, miR-155 was demonstrated as DMTF1 silence regulator to induce bladder cancer cell growth [30]. In myeloid leukemias, DMTF1 was also identified as potential drug target [31]. Though knock-out DMTF1 mutant, murine model of lung cancers reveals the role of DMTF1 in tumorigenesis [32]. These results suggest not only the loss of DMTF1, but also expression level of DMTF1 is critical to cancer cell proliferation. This is consistent with our results that METTL3 control the thyroid cancer cell cycle through regulating DMTF1 expression. Furthermore, studies found that DMTF1 generated three different mRNA isoforms which encode α, β and γ proteins. In breast cancer, it was found that the ratio of different isoforms could affect the disease development [16]. Different isoforms such as DMP1α and DMP1β, have revise role in breast cancer. DMP1α and DMP1β are tumor suppressor and promoting, respectively [33]. This suggestion that the expression level of DMTF1 could be the critical point to control the cell proliferation in cancer cell. There is still no published literation showing the function of DMTF1 in thyroid cancer. Here, we proofed the same function of DMTF1 in cell proliferation and identified a novel mechanism, RNA m6A methylation regulating DMTF1 expression in cancer cell. It gives a new clue for cancer drug design targeting on DMTF1. Together with our cell-based perturbations—where METTL3 enhances proliferation via DMTF1 and DMTF1 overexpression rescues METTL3 knockdown—these results indicate that METTL3 is a candidate driver of PTC progression. Our cross-sectional microbiome data further associate reduced intratumoral D. lacustris with this METTL3-high state, but they do not establish that D. lacustris reduction causes METTL3 upregulation.

We identify a correlation between reduced intratumoral Delftia lacustris abundance and increased METTL3 expression in PTC, and we show that METTL3 promotes PTC cell proliferation through DMTF1, with DMTF1 overexpression rescuing METTL3 knockdown phenotypes. These results support a METTL3–DMTF1 proliferative axis in PTC and nominate D. lacustris as a potential biomarker linked to this axis. Given the study’s cross-sectional design, our findings should be interpreted as associative, not causal. They outline a clear, testable hypothesis for future longitudinal and spatially resolved studies to determine whether microbiome shifts are a consequence of tumor evolution or a contributor to METTL3-dependent tumor biology.

Author contributions

TL and KQ designed the project and wrote the manuscript. KQ did almost experiments. QX, DZ and YH did data analysis and helped to write the manuscript.

Funding

The research is supported by Fujian provincial natural science foundation project (No. 2018J01297); Joint fund project for scientific and technological innovation, Fujian Province (No. 2018Y9061).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The research was approved by the Institutional Ethics Committee Fujian Medical University Union Hospital (Approval No.: [2023-PTC-056]). Given the retrospective design, use of archived de-identified tissues, and minimal risk to participants, the requirement for written informed consent was waived by the Committee. All procedures complied with the Declaration of Helsinki and relevant institutional guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.