Multi-omics analysis reveals intratumoral microbiota as modulators of the immune environment within thyroid cancer

Introduction

Thyroid cancer (TC) ranks as the most prevalent malignancy of the endocrine system, encompassing several histological types: papillary thyroid carcinoma (PTC), follicular thyroid carcinoma (FTC), Hurthle cell carcinoma (HCTC), medullary thyroid carcinoma (MTC), and anaplastic thyroid carcinoma (ATC). These types constitute 80.2%, 11.4%, 3.1%, 3.5%, and 1.7% of cases respectively (1). Previous studies have shown that the development of TC is influenced by a number of risk factors, some of which are not modifiable, such as age, gender, race, and family history of TC being the most important risk factors. The older the age, the higher the incidence and the lower the survival rate. TC is three times more common in women than in men (2). In addition, radiation and hormones are also considered to be important influencing factors (3).

Current research on the mechanisms of TC development focuses on genetic factors. BRAF mutations, which cause activation of the mitogen-activated protein kinase (MAPK) signaling pathway, are detected in more than 50% of TC patients (4). The activation of the MAPK signaling pathway promotes tumor progression. Mutations in RAS, phosphatidylinositol-4,5-bisphosphate 3-kinase catalytic subunit alpha (PIK3CA) and AKT1 are also thought to be associated with thyroid development, causing phosphatase and tensin homolog deleted on chromosome ten (PTEN) inactivation and activation of the downstream Wnt/β-catenin signaling pathway to promote tumor development (5).

Recent studies highlight the significant role of intratumoral microbiota within the tumor microenvironment in influencing tumor development. Intratumoral microbiota can influence cell signaling pathways by regulating signaling molecules that influence cell signaling pathways. Disrupting the homeostatic balance between bacteria in the tumor and the host immune system as well as affecting DNA stability can cause cancer development (6). The relationship between tumor-associated microbes and TC development remains unclear. Research suggests that the presence of Aspergillus in thyroid tissue might stem from thyroid tumor cells, leading to a distinct microbial community (7). Studies indicate that intratumoral microbiota, including species like Pseudomonas aeruginosa, exhibit varied abundance profiles across different stages of PTC. Higher alpha diversity has been observed in women, suggesting gender-specific microbial dynamics within thyroid tumors. Moreover, interactions between tumor bacteria and antibodies associated with autoimmune thyroid disease (AITD) have been identified, with specific microbial taxa showing associations with anti-thyroid peroxidase (TPO) levels. Notably, bacteria such as members of the Prevotella family, Anaplasma phylum, and Bifidobacteria demonstrate negative correlations with TPO levels, underscoring their potential immunomodulatory roles in TC microenvironments (8). A study by Dai et al. characterized both tumor and matched peritumor tissue from 55 patients with early-stage TC (stage I and II) who underwent thyroidectomy. They found that tumor tissues exhibited significantly lower alpha diversity and abundance compared to peritumor tissues. Notably, patients with stage N1 had higher alpha diversity in their thyroid microbiomes compared to stage N0 patients, while no significant differences were observed between female and male patients (9).

This study aims to analyze the bacterial composition of thyroid tumors and adjacent paracancerous tissues by mapping sequencing data to bacterial. Additionally, it explores the associations between tumor microbiome composition, gene expression, and immune infiltration profiles. The goal is to elucidate the impact of these microbial interactions on thyroid carcinogenesis and progression. By examining how microbiota influence thyroid genes and immune cells, this research seeks to enhance our understanding of TC development and provide insights for new clinical diagnostic and therapeutic strategies. We present this article in accordance with the STREGA reporting checklist (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-762/rc).

Methods

Data sources

Tumor microbiome data from the BIC database (http://bic.jhlab.tw/) (10). Transcriptome files were obtained from The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/). Clinical information was obtained from NCI Genomic Data Commons (GDC) searches. TCGA data and clinical information downloaded from Xena database (https://xenabrowser.net/datapages/) (11). This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Microbiological data processing and analysis

The R package “vegan” was used to calculate the Shannon index, Simpson’s index and chao1 index of the colony, which were used to characterise the alpha diversity of the colony. Beta diversity was analysed using principal coordinate analysis (PCoA) based on bray distance and partial least squares discriminant analysis (PLS-DA). Linear discriminant analysis effect size (LEfSe) was used to assess differential taxa between subgroups.

Gene transcriptome analysis

Differential gene expression analyses were performed using the R package limma using the criteria |log₂ fold change| >1 and false discovery rate (FDR) < 0.05 as differential expression gene screening thresholds (12). Immune cell and stromal cell scores were assessed using ESTIMATE (13), and common 22 immune cell infiltration scores were calculated by (14). Phenotype-related gene modules were calculated using the R package “WGCNA” (weighted gene co-expression network analysis), and relevant differential genes were obtained by intersecting with differentially expressed genes. Phenotype-related gene modules were calculated using the R package “WGCNA”, and relevant differential genes were obtained by intersecting with differentially expressed genes. The downstream pathways were further enriched using enrichr (https://maayanlab.cloud/Enrichr/) (15) and Metascape (https://metascape.org/gp/index.html) (16).

Statistical analyses

The unpaired t-test was used to analyse normally distributed variables and the Mann-Whitney test was used to assess non-normally distributed variables. The Wilcoxon rank sum test was used to compare differences between groups. For more than two groups we used analysis of variance (ANOVA) test. Correlation was evaluated using Spearman correlation assessment. All statistical P values were two-sided, where P<0.05 was considered statistically significant. All data processing was done in R 4.1.0 software.

Results

Sample information

The clinical characteristics of the patients are summarised in Table 1. As total of 565 sample data from 506 patients were included in the study, of which 506 were tumor tissue samples and 59 were normal solid tissue samples.

Differences in bacterial alpha diversity between normal and tumor tissue

We compute the alpha diversity of TC. Significant differences were found between tumor and paracancerous tissue on the Shannon, Simpson, and chao1 indices. Normal tissues’ alpha diversity was higher than tumor tissues’ (Figure 1A). However, further analysis of TC tumor tissues between different tumor, node, metastasis (TNM) stages revealed only a difference in the chao1 index between II and IV (Figure 1B). Also, PCoA showed that in TC, there was no difference in the alpha diversity of the microbiota between different TNM stages (Figure 1C,1D). However, considering that only chao1 was significantly different in the diversity index for different stages, we could not conclude that microbial alpha diversity in tumors correlates with tumor stage. These results suggest an association between the tumorigenic process and tissue microbiome diversity in TC patients.

Figure 1 Microbiological diversity of paraneoplastic and tumor tissues of thyroid cancer. (A) Comparison of alpha diversity of paraneoplastic and tumor tissues. (B) Comparison of alpha diversity of tumor tissues between different stages. (C) Comparison of β-diversity between paraneoplastic and tumor tissues. (D) Comparison of β-diversity of tumor tissues between different stages. The Wilcoxon test was used to measure the difference in microbial composition between the two groups. *, P<0.05; **, P<0.01; ***, P<0.001. The chao1 index represents the richness of microbial species; Simpson’s index and Shannon’s index reflect the diversity of microbial species. PCoA, principal coordinate analysis.

Biological microbiological characterisation of different tissues

A total of 47 phyla were identified in the data analysis, and taxonomic composition analysis showed that five phyla were predominant in all samples (paracancerous and tumor tissues): Proteobacteria (43.07%), Ascomycetes (53.36%), Fungi (35.70%), Bacillariophyta (3.31%), Actinomycetes (13.91%) and Cyanobacteria (1.36%) (Figure 2A). There were a total of 11 phylum-level abundances that differed, with an increase in the abundance of Firmicutes in tumor tissues relative to paracancerous tissues among those that differed in core microbiota (top 20 in relative abundance), Actinobacteria, Bacteroidetes, Planctomycetes, Tenericutes and the Spirochaetes decreased in abundance. In addition, none of the paracancerous tissues of different genders showed significant differences in the composition of bacterial species at the phylum level. However, it is interesting to note that in tumor tissues, the abundance of Actinobacteria, Chloroflexi in the core microbiota of female tumor tissues was significantly higher than that of male tissues (Figure S1A). In addition, we used different age intervals to categorize the population into young (<40 years), middle (40–65 years), and old (>65 years) groups. We found in core microbiota, abundance of Gemmatimonadetess was significantly higher in the old group than in the other groups. and the abundance of Planctomycetes and Tenericutes was also higher in the elderly group than in the young group (Figure S1B, Table S1).

Figure 2 Differences in the microbial composition of paraneoplastic and tumors. (A) Composition of the top 15 species at the phylum level. (B) Composition of the top 30 species at the genus level. (C) Cladogram showing taxonomic trees of taxa with significant differences in abundance between different species of tumor tissue and paracancerous tissue. Taxonomic associations between microbiome communities from normal and tumor tissues are described. Each node represents a specific taxonomic type. Yellow nodes indicate taxonomic features that are not significantly different between paracancerous tissue and tumor tissue. Red nodes indicate taxa that are more abundant in paracancerous tissue than in tumor tissue, while green nodes indicate taxa that are more abundant in tumor tissue. (D) LDA scores for bacterial taxa were shown to differ significantly in abundance between paraneoplastic and tumor tissues. LDA, linear discriminant analysis.

A total of 886 genera were identified at the genus level, comparing the relative abundance of tumor tissue and paracancerous tissue microbiota, a total of 12 genera differed, of which seven were core microbiota (top 20 in relative abundance). Among the significant differential groups, tumor tissue relative to paraneoplastic tissue Leptospira, Corynebacterium, and Actinoplanes decreased in relative abundance, and Brevibacillus, Peptoclostridium, and Bacillus increased in relative abundance (Figure 2B). At the genus level, the core microbiota of paracancerous tissues did not show gender differences. The abundance of Paenibacillus and Burkholderia in tumor tissues was significantly higher in females than in male tissues (Figure S1C). Subgroup analyses for different age intervals revealed that the abundance of tumor tissues Corynebacterium and Burkholderia was significantly higher in the elderly population than in the normal population, and Azotobacter and Paenibacillus were higher than in the middle-aged population (Figure S1D, Table S2).

Further LEfse analysis of the tumor tissues revealed that the core strains Spirochaetes, Fusobacteria, Firmicutes, and Actinobacteria could be used as markers at the portal level. Based on linear discriminant analysis (LDA) analysis, 16 genera had potential as markers at the genus level, which encompassed six genera of core differential bacterial groups (Figure 2C,2D).

Identification of relevant differentially expressed genes between tumor tissue and paracancerous tissue

In order to understand the main effects of the tumor microbiome on the tumor genome, we analyzed differential genes in paraneoplastic and tumor tissues, as well as gene modules by WGCNA. We found that 570 genes were up-regulated and 1,309 genes were down-regulated in the tumor tissues compared to the paraneoplastic tissues (Figure 3A). We analyzed the relationship between the gene modules and the phenotype using WGCNA to identify further the differential genes associated with the phenotype. We screened for the differential genes among them. In total, 180 differential genes were screened (Figure 3B,3C).

Figure 3 Identification of relevant differentially expressed genes. (A) Volcano diagram showing differentially expressed genes in paraneoplastic and tumor tissues. Red is up-regulated and green is down-regulated. (B) WGCNA analysis of relevant gene expression modules in tumors and paracancerous tissues. (C) Veen plot showing differentially expressed genes associated with tumorigenesis. WGCNA, weighted gene co-expression network analysis.

Differential expressed gene enrichment analysis

We performed downstream enrichment analyses of differentially expressed genes to explore the changes during thyroid carcinogenesis further. Gene Ontology (GO) analysis revealed that the differentially expressed gene products were mainly located in cellular vesicles, affecting calcium ion binding and lipid binding, thereby regulating the response to stimuli (Figure 4A-4C). GO analysis revealed that the differentially expressed gene products were mainly located in cellular vesicles, affecting calcium ion binding and lipid binding, thereby regulating the response to stimuli. Enrichment analysis of signaling pathways and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis showed that it mainly affects fatty acid degradation, beta-alanine metabolism, and circadian rhythm. WikiPathway showed that it mainly affects prostaglandin synthesis and regulation of ethanol effects on histone modifications and fatty acid omega oxidation. Reactome showed that it mainly affects mitochondrial fatty acid beta-oxidation. BioPlanet showed significant regulation of prostaglandin biosynthesis and regulation, fatty acid metabolism, and histidine metabolism (Figure 4D-4G). Further enrichment analysis network maps were constructed by metascape and revealed that fatty acid metabolism may be an essential pathway connecting cellular functions (Figure 5).

Figure 4 Enrichment analysis of associated differentially expressed genes. (A-C) GO enrichment analysis: (A) biological process; (B) cellular component; (C) molecular function. (D-G) Signaling pathway enrichment analysis: (D) KEGG enrichment analysis; (E) WikiPathway enrichment analysis; (F) Reactome enrichment analysis; (G) BioPlanet enrichment analysis. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 5 Network of enriched terms.

Intratumoral microbial composition correlates with immune cell composition

To further understand the impact of intratumoral microbes on tumor tissue, we used ESTIMATE to score tumor tissue and paracancerous tissue. Paracancerous tissues were found to have lower immune cell scores than tumor tissues, while there was no difference in stromal cell scores (Figure 6A). Suggesting that intratumoral microbes contributing to tumorigenesis may be associated with immune cells, further analysis of immune cells by Cibersort revealed that M2-type macrophages, resting mast cells, and activated dendritic cell (DC) were significantly up-regulated in tumor tissues compared to paraneoplastic tissues. In contrast, CD8 cells, M1-type macrophages, activated mast cells, and eosinophils scores were decreased (Figure 6B).

Figure 6 Immune cell infiltration analysis. (A) ESTIMATE calculates the immune cell infiltration score. (B) CIBERSORT calculates the 22 immune cell scores. *, P<0.05. NK, natural killer; Tregs, regulatory T cells.

To investigate the effect of intratumoral microorganisms on immune cells, we performed a correlation analysis between differential microbiota and immune cells. We found that there did not appear to be a correlation between differential microbiota and immune cells (|r|<0.2). However, we further analyzed the microbiota with differential genes and found that the differential microbiota correlated with ABCA9, ADH1B, APOD, ARL4A, ART4, C16orf86, C1QTNF2, C1QTNF7, and POLR2J4. Moreover, there is a correlation between these differential genes related to microbiota and immune cells, suggesting that our differential microbiota may regulate the gene and signaling pathways in tumor and paracancerous tissues, thus regulating immune cells to cause tumorigenesis (Figure 7).

Figure 7 Microbe-gene-immune cell association analysis. (A) Heatmap of the association between the top 15 microbes and DEG. (B) Heatmap of the association between the top 15 immune cells and DEG. (C) “Microbe-Gene-Immune Cell” interconnection network diagram. DEG, differentially expressed gene; NK, natural killer; Tregs, regulatory T cells.

Discussion

The role of intra-tumor microorganisms in tumors has received increasing attention, and as an essential component of the tumor microenvironment, intratumoral microbiota are closely associated with tumor development. We found that the microbiota α diversity was significantly higher in paracancerous tissues compared to tumor tissues using data from the TCGA database. This is the same as Dai’s findings (9). However, interestingly, Yuan came to a different conclusion. Their study found no difference in microbial diversity in peripheral tissues compared to diseased tissues in benign thyroid nodules or TC (8). Our results support Dai’s conclusion more. In addition, we found a direct significant difference in the chao1 index of tumor tissues only between stages II and IV, which is inconsistent with the negative correlation found by Yuan between the diversity index and TC progression. This may be due to Yuan’s small sample size.

Analyzing the microbiota composition, we found significant changes in the relative abundance of the main genera Leptospira, Corynebacterium, Actinoplanes, Brevibacillus, Peptoclostridium, and Bacillus. Leptospira, Corynebacterium and Actinoplanes were significantly down-regulated in tumor tissue compared to paraneoplastic tissue. Among them, extracts of Actinoplanes proved to have a wide range of anti-tumoral effects, inhibiting breast cancer, glioblastoma, lung cancer, and kidney cancer. Among them, the extract of Actinoplanes proved to have good antitumor activity and inhibited breast cancer, glioblastoma, lung cancer, and kidney cancer (17). Some researchers have studied the metabolites of Actinoplanes and found that chlorinated genistein and 7,8-dihydroxy-1-methylnaphtho[2,3-c]furan-4,9-dione in the metabolites of Actinoplanes showed good antitumor activity, and the antitumor activity of 7,8-dihydroxy-1-methylnaphtho[2,3-c]furan-4,9-dione in the metabolites of Actinoplanes was also found to be good. The antitumor activity of the metabolites was found to be good. This may be an essential factor in its inhibition of tumorigenesis (18,19). Corynebacterium has been reported to be positively associated with the progression of esophageal squamous cell carcinoma. This is similar to our findings, but further studies on the mechanism of action of Corynebacterium are needed (20). The role of Bacillus in tumors is controversial. Patients with tumors are more susceptible to Bacillus bacteremia. It has been demonstrated that Bacillus abundance rises in patients with TC and that Bacillus concentrations increase with thyroid-stimulating hormone, which may contribute to the development of TC (21). However, metabolites of Bacillus usually have antitumor effects. Furthermore, in a study of pancreatic ductal adenocarcinoma (PDAC), Bacillus sphaericus in paraneoplastic tissues was found to impede the development of PDAC and to increase the infiltration of inflammatory neutrophils in the tumor (22). Our results favor that Bacillus cause tumorigeneses.

It is also interesting to note that in the analysis of the transcriptome of tumor tissues, we found that the primary metabolic pathway regulated by the different genes in paraneoplastic and tumor tissues was not glycolysis or oxidative phosphorylation but fatty acid metabolism. Obesity has also been observed to be an essential risk factor for TC (23), and obesity is strongly associated with fatty acid metabolism. Enhanced fatty acid metabolism in TC triggers metastasis and immunosuppression (24,25). In addition, our study revealed that the differential gene also affects mitochondria fatty acid β-oxidation (FAO). Targeting aberrant FAO targets considered a potential strategy for tumor therapy (26).

Immune cells as an important component of the tumor microenvironment play an important role in tumor development. Our study revealed that intratumoral microbes modulate DC cells, T cells, macrophages, B cells, eosinophils, natural killer (NK) cells and mast cells. Tumor antigen presentation by DCs can trigger antigen-specific cytotoxic CD8 T lymphocytes to exert tumor suppressive effects (27). CD4 T cell infiltration correlates with the conventional dendritic cell type 2 (cDC2) ratio. The higher the frequency of DC, the higher the rate of CD4 T cell tumor infiltration. The anti-tumor function of CD4 T cells is not only the ability to directly activate cytotoxic T lymphocytes (CTLs), but also to inhibit tumors through the secretion of INFγ-induced macrophage and NK cell activation (28). Mast cells have a role in tumors and are uncertain, being a favourable prognostic factor in some cancers and a marker of poor prognosis in others. The higher score of mast cells in paraneoplastic tissues than in tumor tissues in our conclusions may be related to the fact that mast cells activate NK cells to secrete IFN-γ in a contact-dependent manner and also mediate changes in monocytes thereby reducing the immunosuppressive signals of NK cells to exert anti-tumor effects (29). We also found significant differences in macrophage polarisation between tumor and paracancerous tissues, with macrophages polarising towards the M2 type in tumor tissue and towards M1 in paracancerous tissues. M2 cells are thought to exert an immunosuppressive effect to promote tumor growth. In addition, FAO provides a critical energy source for macrophage polarisation towards the M2 phenotype, and inhibition of FAO in tumor-associated macrophages promotes anti-tumor differentiation of tumor-associated macrophages and inhibits tumor growth (30). This is consistent with our previous enrichment pathway finding.

There are still some limitations to our study. As we were unable to obtain thyroid tissue from normal subjects, we had difficulty confirming the origin of the thyroid microbiota, and it also made it possible to make comparisons only from paraneoplastic and tumor tissue. Paraneoplastic tissue is likely to be influenced by the nearby tumor microenvironment and may not reflect the characteristics of truly healthy tissue. The bacterial composition of paracancerous tissues is susceptible to the influence of nearby tumor tissue. The use of normal tissue greatly reduces such deviations. Second, because most TCs have a good prognosis, it is difficult for us to identify precisely which microbiota affects death and prognosis in TC. Finally, we did not analyze the TC microbiota at different stages and subtypes.

Conclusions

We have revealed differences in microbial composition between TC tumor tissues and paracancerous tissues through a multi-omics approach. The correlation was found between some microorganisms and tumorigenesis and development. Correlation-based analyses also revealed a correlation between tumor microbes and gene expression, and may modulating the infiltration of immune cells by modulating fatty acid metabolism thereby influencing tumorigenesis.

Acknowledgments

The authors thank the TCGA and BIC databases for providing publicly available data.

Footnote

Reporting Checklist: The authors have completed the STREGA reporting checklist. Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-762/rc

Peer Review File: Available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-762/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tcr.amegroups.com/article/view/10.21037/tcr-2025-762/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cabanillas ME, McFadden DG, Durante C. Thyroid cancer. Lancet 2016;388:2783-95. [Crossref] [PubMed]
2. Miranda-Filho A, Lortet-Tieulent J, Bray F, et al. Thyroid cancer incidence trends by histology in 25 countries: a population-based study. Lancet Diabetes Endocrinol 2021;9:225-34. [Crossref] [PubMed]
3. Saenko V, Mitsutake N. Radiation-Related Thyroid Cancer. Endocr Rev 2024;45:1-29. [Crossref] [PubMed]
4. Busaidy NL, Konda B, Wei L, et al. Dabrafenib Versus Dabrafenib + Trametinib in BRAF-Mutated Radioactive Iodine Refractory Differentiated Thyroid Cancer: Results of a Randomized, Phase 2, Open-Label Multicenter Trial. Thyroid 2022;32:1184-92. [Crossref] [PubMed]
5. Prete A, Borges de Souza P, Censi S, et al. Update on Fundamental Mechanisms of Thyroid Cancer. Front Endocrinol (Lausanne) 2020;11:102. [Crossref] [PubMed]
6. Li J, Zhang Y, Cai Y, et al. Multi-omics analysis elucidates the relationship between intratumor microbiome and host immune heterogeneity in breast cancer. Microbiol Spectr 2024;12:e0410423. [Crossref] [PubMed]
7. Liu CJ, Chen SQ, Zhang SY, et al. The comparison of microbial communities in thyroid tissues from thyroid carcinoma patients. J Microbiol 2021;59:988-1001. [Crossref] [PubMed]
8. Yuan L, Yang P, Wei G, et al. Tumor microbiome diversity influences papillary thyroid cancer invasion. Commun Biol 2022;5:864. [Crossref] [PubMed]
9. Dai D, Yang Y, Yang Y, et al. Alterations of thyroid microbiota across different thyroid microhabitats in patients with thyroid carcinoma. J Transl Med 2021;19:488. [Crossref] [PubMed]
10. Chen DW, Lang BHH, McLeod DSA, et al. Thyroid cancer. Lancet 2023;401:1531-44. [Crossref] [PubMed]
11. Lee BT, Barber GP, Benet-Pagès A, et al. The UCSC Genome Browser database: 2022 update. Nucleic Acids Res 2022;50:D1115-22. [Crossref] [PubMed]
12. Ritchie ME, Phipson B, Wu D, et al. limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res 2015;43:e47. [Crossref] [PubMed]
13. Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nat Commun 2013;4:2612. [Crossref] [PubMed]
14. Thorsson V, Gibbs DL, Brown SD, et al. The Immune Landscape of Cancer. Immunity 2018;48:812-830.e14. [Crossref] [PubMed]
15. Kuleshov MV, Jones MR, Rouillard AD, et al. Enrichr: a comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res 2016;44:W90-7. [Crossref] [PubMed]
16. Zhou Y, Zhou B, Pache L, et al. Metascape provides a biologist-oriented resource for the analysis of systems-level datasets. Nat Commun 2019;10:1523. [Crossref] [PubMed]
17. Silva LJ, Crevelin EJ, Souza DT, et al. Actinobacteria from Antarctica as a source for anticancer discovery. Sci Rep 2020;10:13870. [Crossref] [PubMed]
18. Xiang WS, Zhang J, Wang JD, et al. Isolation and identification of chlorinated genistein from Actinoplanes sp. HBDN08 with antioxidant and antitumor activities. J Agric Food Chem 2010;58:1933-8. [Crossref] [PubMed]
19. Breyer S, Effenberger-Neidnicht K, Knauer S, et al. Synthesis, anticancer activity, and iron affinity of the Actinoplanes metabolite 7,8-dihydroxy-1-methylnaphtho[2,3-c]furan-4,9-dione. Bioorg Med Chem 2011;19:1264-7.
20. Zhang B, Xiao Q, Chen H, et al. Comparison of tumor-associated and nontumor-associated esophageal mucosa microbiota in patients with esophageal squamous cell carcinoma. Medicine (Baltimore) 2022;101:e30483. [Crossref] [PubMed]
21. An J, Kwon H, Kim YJ. The Role of Blood Microbiome in the Development of Thyroid Cancer in Breast Cancer Survivors. Cancers (Basel) 2023;15:4492. [Crossref] [PubMed]
22. Zhang Z, Wang T, Xu M, et al. Deciphering the pancreatic cancer microbiome in Mainland China: Impact of Exiguobacterium/Bacillus ratio on tumor progression and prognostic significance. Pharmacol Res 2024;204:107197. [Crossref] [PubMed]
23. Fussey JM, Beaumont RN, Wood AR, et al. Does Obesity Cause Thyroid Cancer? A Mendelian Randomization Study. J Clin Endocrinol Metab 2020;105:e2398-407. [Crossref] [PubMed]
24. Lu J, Zhang Y, Sun M, et al. Multi-Omics Analysis of Fatty Acid Metabolism in Thyroid Carcinoma. Front Oncol 2021;11:737127. [Crossref] [PubMed]
25. Liu C, Zhou X, Pan Y, et al. Pyruvate carboxylase promotes thyroid cancer aggressiveness through fatty acid synthesis. BMC Cancer 2021;21:722. [Crossref] [PubMed]
26. Chen H, Yang Z, Sun Y, et al. Targeting the Key Enzymes of Abnormal Fatty Acid β-oxidation as a Potential Strategy for Tumor Therapy. Front Biosci (Landmark Ed) 2022;27:95. [Crossref] [PubMed]
27. Böttcher JP, Reis e Sousa C. The Role of Type 1 Conventional Dendritic Cells in Cancer Immunity. Trends Cancer 2018;4:784-92. [Crossref] [PubMed]
28. Del Prete A, Salvi V, Soriani A, et al. Dendritic cell subsets in cancer immunity and tumor antigen sensing. Cell Mol Immunol 2023;20:432-47. [Crossref] [PubMed]
29. Lichterman JN, Reddy SM. Mast Cells: A New Frontier for Cancer Immunotherapy. Cells 2021;10:1270. [Crossref] [PubMed]
30. Mehla K, Singh PK. Metabolic Regulation of Macrophage Polarization in Cancer. Trends Cancer 2019;5:822-34. [Crossref] [PubMed]