The tumor microbiome in cancer progression: mechanisms and therapeutic potential

Abstract

The tumor microbiome (TM) comprises diverse microbial communities, such as bacteria, fungi, and viruses. Recent advancements in microbial sequencing technologies have improved our understanding of the distribution and functional roles of microbes in solid tumors. The TM is formed through several mechanisms, such as direct invasion of mucosal barriers, diffusion from adjacent normal tissues, metastasis of tumor cells, and dissemination via blood and lymphatic circulation. Microbes play a critical role in the tumor microenvironment (TME), and the TM has a heterogeneous composition in different types of cancer. This heterogeneity affects tumor development, progression, and response to treatment. The TM modulates tumor cell physiology and immune responses via several signaling pathways, such as WNT/β-catenin, NF-κB, toll-like receptors (TLRs), ERK, and stimulator of interferon genes (STING). Extensive studies have characterized the role of TM in tumor progression, revealing the importance of genetic abnormalities, epigenetic changes, metabolic regulation, invasion and metastasis, and chronic inflammatory responses. The role of TM in cancer treatment, especially in immunotherapy, has received increasing attention, demonstrating significant regulatory potential. This review provides an in-depth overview of the development of TM detection technologies, explores its potential origins and heterogeneity, and elucidates the mechanisms by which TM contributes to tumorigenesis or tumor suppression. Furthermore, this review explored how TM can be used in cancer treatment, offering a comprehensive perspective on targeted and personalized approaches.

Introduction

Cancer is a highly heterogeneous disease caused by genetic mutations, cellular dysfunction, and changes in the microenvironment. Its biology encompasses complex molecular mechanisms across multiple levels. Accumulating evidence highlights the pivotal role of the human microbiome in cancer development and progression [1, 2]. The TM, a critical subset of the human microbiome, encompasses bacteria, viruses, and fungi within and around tumor tissues. It mediates complex host-tumor interactions and affects all stages of cancer biology, from its development to metastasis [3]. The human microbiome map has been refined through the synergy of cross-disciplinary technologies, including the optimization of in vitro culture systems, innovations in high-throughput sequencing, spatial omics analysis, and machine learning algorithms, breakthroughs in gene editing, and germ-free animal models. Multiple organs previously deemed sterile, such as the lungs, breast, liver, pancreas, prostate, and kidneys, have been confirmed to harbor low-abundance microbial colonization [4–7]. In a disease context, changes in microbiome composition, referred to as “dysbiosis”, can drive community-level interactions within microbial niches, thereby contributing to cancer progression [8]. TM diversely affects its host by changing signaling pathways and epigenetic landscapes, inducing DNA damage, promoting chronic inflammation, mediating epithelial–mesenchymal transition (EMT), and reshaping local immune responses through microbial metabolites [9–12]. Recently, innovative anti-tumor treatment strategies based on the TM have become an important research area. Most studies have focused on eliminating oncogenic microbes in the tumor tissues to enhance the efficacy of cancer treatment, while others have emphasized the oncolytic properties of intratumoral bacteria [13, 14]. This study systematically reviews the advancements in TM detection technologies and their evolution. Moreover, the study comprehensively analyzes the potential biological origins and heterogeneous composition of the TM. Additionally, it explores the functional mechanisms of the TM associated with malignant evolution and immune regulation in cancer. Therefore, we also explored the potential application of microbial-targeted intervention strategies in cancer therapy.

Development and progress of microbiological detection technology

The composition of microbial communities correlates with tumor type, progression, and prognosis. Hence, investigating the presence and distribution of microbes within tumors offers insights into their spatial distribution and functional roles within the TME [15]. Consequently, developing effective detection technologies for studying the TM is a critical research area. Over the past decades, significant advancements in microbiological science, driven by next generation sequencing (NGS), microbiome-specific computational pipelines, and wet-lab techniques, have enabled high- and low-throughput hypothesis testing [16, 17]. A timeline of microbial detection technology highlights the rapid advancement and evolution in this field?. These technological advances have unlocked unprecedented opportunities to investigate the diversity and its functional attributes.

Early techniques: traditional culture and molecular biological methods

The origin of microbial culture dates back to 1884, when Robert Koch introduced Koch’s postulates and developed various solid culture media visible under a microscope, such as agar media, which facilitated the cultivation and isolation of bacteria and established a framework for assessing microbial pathogenicity using culture-based methods [18, 19]. However, these methods have notable limitations: they are time-consuming, complex, require skilled personnel, and cannot detect viable but non-culturable microbial species.

Since the 1980s, microbial detection has transitioned from basic tissue culture and microscopic observation to the incorporation of molecular biology techniques. In 1985, Mullis et al. introduced the Polymerase Chain Reaction (PCR), which allows for the efficient amplification of specific DNA sequences, leading to the development of real-time polymerase chain reaction (RT-PCR) and quantitative real-time polymerase chain reaction (qRT-PCR). In a study characterizing the airway microbiome, variations in the number of PCR cycles and selection of target marker gene regions influenced the final bacterial community profile and detection of low biomass samples [20]. Compared to traditional methods, molecular approaches generally offer enhanced reliability and advantages in sensitivity, specificity, speed, and cost-effectiveness, thereby providing transformative tools for microbial detection [21, 22].

High-throughput sequencing technologies: microarrays and NGS

The advent of high-throughput sequencing ushered in unprecedented opportunities and challenges for microbial detection, as vast datasets of microbiomes continue to emerge [23]. Microarrays were initially used to study gene expression profiles in the mid-1990s and were first publicly reported for use in large cohorts of soft tissue tumor samples in 2001 [24]. DNA microarrays, comprising chips or slides with thousands of DNA elements, facilitate the simultaneous measurement of genome-wide mRNA levels to produce “gene expression profiles.” These arrays enable the parallel, rapid, and high-throughput detection of microbial sequences from diverse samples, exemplified by PhyloChip and GeoChip. Such DNA microarrays have transformed the study of complex microbial communities. However, the rapid advancement of NGS technologies has introduced significant challenges to the continued application of DNA microarrays in microbial community research [24, 25].

Characterizing the TM is challenging due to the unique properties and low biomass of microbes. Since 2005, the development and progress of NGS have facilitated a more comprehensive understanding of the local diversity and functional relevance of the TM in solid malignant tumors [16, 26]. 16 S rRNA gene sequencing, a widely used and cost-effective NGS platform, enables high-throughput characterization of microbial communities to elucidate their structural composition [11]. For instance, 16 S rRNA gene amplicon sequencing combined with culturomics was employed to assess and compare the structure and diversity of oral and pulmonary microbiome associated with lung cancer. This approach revealed significantly increased levels of Streptococcus and Prevotella in pneumonia and oral samples from patients with lung cancer; these microbes may serve as potential bacterial biomarkers and new targets for the diagnosis and treatment of lung cancer [27]. As an alternative, shotgun metagenome sequencing profiles the taxonomic composition of samples and assesses genes to characterize the genomic functional potential of the microbiome. Compared to 16 S rRNA gene sequencing, shotgun metagenome sequencing offers increased sequencing depth at higher costs, providing species- and even strain-level resolution [11].

Since most studies have focused on easily cultured microbiomes, the diversity of many microbiomes remains undercharacterized. Metagenomic sequencing, which can calculate and reconstruct the composition of microbial communities from a pool of sequence reads and identify new microbial genomes from unnamed species. Candidatus Chibobacter qucibialis was discovered as an unknown species. Taxonomically, it belongs to the order Clostridiales, which further increases the diversity of microbial genomes [28]. Metagenomics employs NGS to delineate the transcriptional landscapes of individual bacteria or entire microbiomes, thereby providing a more direct measure of microbial activity and function. This approach can be integrated with downstream proteomics and metabolomics [29].

High-throughput sequencing technologies, such as microarrays and NGS, have greatly advanced microbiome research, allowing us to dissect the structure and function of microbial communities with unprecedented precision. Previously, microarray technology played a critical role in gene expression profiling and microbial detection in the early stage, but due to its limitations in terms of resolution and throughput, it has gradually been replaced by NGS technology [30]. Thanks to its cost-effectiveness and efficiency, 16 S rRNA sequencing significantly improved the ability to resolve microbial community structure and has become one of the gold standards for microbiome studies [31]. While Shotgun metagenomics provides deeper species and functional analyses [32]. However, current research still faces limitations in terms of sequencing depth, standardization of computational analysis methods, and in-depth elucidation of the mechanisms underlying microbiome-host interaction.

Emerging technologies: microbial single cell sequencing and spatial analysis

The forthcoming wave of innovation is concentrating on the single-cell level. Developing mammalian single-cell RNA sequencing in 2009 heralded the emergence of the single cell sequencing (SCS) era [15]. Sequencing technology has continuously been updated and iterated over the past decade (Table 1). This section provides a comprehensive overview of the advantages and limitations of various microbial SCS technologies. In 2021, the microbial split-pool ligation transcriptomics technique (MicroSPLiT) facilitated single-cell transcriptome sequencing of Bacillus subtilis, revealing the heterogeneity in behavior, metabolism, and stress responses among a limited number of bacterial species present in the human body, which allows for effective clustering of bacteria and the identification of distinct cellular subpopulations that demonstrate differential gene expression across selected metabolic pathways, stress responses, or developmental processes. Conversely, bulk RNA-seq and low-throughput SCS methods cannot detect physiologically relevant rare bacterial cell states [33, 34]. In 2022, micro-seq technology combined various droplet microfluidic techniques with custom-developed bioinformatics analysis tools, enabling the acquisition of genomic information from thousands of single microbial cells within complex microbial communities. This advancement offers high-throughput and culture-independent capabilities for investigating the genomic blueprint of such communities at the level of individual microbes [35]. Recently, a high-throughput high-resolution single-microbe RNA sequencing (smRNA-seq) technology, known as smRandom-seq2, was developed based on droplet-based smRNA-seq methods, which accurately verified 373 reliable host-phage interactions within the human gut microbiome and revealed functional heterogeneity and complex interactions between bacteria and phages [36].

Table 1.

Landscape and comparison of Microbiome scRNA-seq methods

Time	Microbiome detection methods	Technical Features	Advantages	Limitations	Species	Sample multiplexing	Refs
2020	MATQ-seq	Plate-based, non-combinatorial	Lower dropout rate, higher throughput, high gene detection per bacterium	High cost, limited scale	Salmonella enterica P.aeruginosa	No	[60]
2020	PETRI-seq	Plate-based	Simple pipeline, low cost, high throughput	Limited sample number, low capture rate	Staphylococcus aureus E.coli	Yes	[61]
2021	MicroSPLiT	Plate-based	High scalability, captures full transcriptional diversity	Limited starting material, not validated for diverse bacterial communities	B.subtilis E.coli	Yes	[33, 34]
2023	ProBac-seq	Probe-based	Accurate quantification of intracellular transcripts, high capture rate	Requires prior genome knowledge, large number of probes, limited to known transcripts	B.subtilis E.coli C.perfringens	No	[62, 63]
2023	BacDrop	Droplet-based	Identifies population heterogeneity, reveals heterogeneous stress responses	Small sample size, high impurity levels, low detection efficiency, high sequencing cost	Klebsiella pneumoniae Enterococcus faecium Pseudomonas aeruginosa E.coli	Yes	[64]
2023	M3-seq	Plate-based, in situ indexing with droplet-based index- ing	Massively parallel gene expression profiling, sensitive mRNA capture	Low bacterial capture rates due to growth stage and sequencing depth differences	B.subtilis E.coli	Yes	[65]
2024	smRandom-seq2	Droplet-based	Improved reverse transcription efficiency, reduces cross-contamination, multi-omics integration	Throughput gap with natural microbial communities, increased sequencing cost due to non-depleted rRNA	Prevotella genus Roseburia genus	No	[36]

Recently, researchers developed imaging-based spatial transcriptomic technologies that measure the copy number and spatial distribution of RNA species in single cells, enabling gene expression profiling across diverse biological samples [37]. The limitations of single-cell technologies in capturing microbial heterogeneity at spatial scales have been addressed by spatial transcriptomics, which allows for in situ sequencing of genes within tissues, thereby rectifying the deficiencies inherent in single-cell techniques. The merging of these methodologies enhances the visualization of host-microbe interactions within the TME at spatial, cellular, and molecular levels. For example, a study exploring the spatial heterogeneity of tumors within the TME used 10x Visium spatial transcriptomics to identify and map tumor cell locations in patient tissues. GeoMx digital spatial analysis further revealed that bacterial communities were distributed across a highly immunosuppressive microecological landscape. Additionally, the study introduced the invasion adhesion directional expression sequencing (INV ADEseq) scRNA-seq method, which targets conserved regions of intracellular bacterial 16 S rRNA. This approach demonstrated that changes in pro-tumor gene expression within host cells may be linked to the enrichment of specific microbial communities exhibiting distinct spatial structures [38]. Recent advances have led to developing spatial host-microbiome sequencing (SHM-seq) technology, which integrates spatial transcriptomics to obtain detailed information on the spatial distribution of host gene expression and microbial communities. This innovative approach combines histology, spatial RNA-seq, and spatial 16 S sequencing, effectively demonstrating spatial gene expression programs associated with host-bacterial interactions [39]. Additionally, the gene expression profiles of individual bacteria and their physical distribution within defined spatial environments, such as tumors, can be investigated using fluorescence in situ hybridization (HiPRFISH) and parallel sequential fluorescence in situ hybridization (parseqFISH), which offer high phylogenetic resolution [40, 41]. In contrast, sequential error-robust fluorescence in situ hybridization (SEER-FISH) primarily investigates the spatial metagenomics of multi-species microbial communities, providing a robust methodology for analyzing the spatial ecology of complex microbial ecosystems [42]. Furthermore, these novel techniques can be combined with traditional non-genetic detection methods such as immunohistochemistry (IHC), which are generally cost-effective and relatively easy to implement, although they are associated with a high false-positive rate [43].

Future studies should optimize SCS and multi-omics integration, such as metatranscriptome, metabolome, and proteome, to more comprehensively investigate the role of microbiome in cancer. Combined with experimental validation and clinical translation, microbiome research can offernew breakthroughs in the field of precision medicine and personalized treatment.

The major pitfalls of TM detection and solutions

First, due to the nature of microorganisms, contamination in the environment or during experiments may mask microbial signals in tumor samples [44, 45]. Second, microorganisms are not a major component of tumors. The extremely low microbial DNA content in tumors, and the lack of poly-A tails in the RNA of prokaryotic organisms (most bacteria, archaea) make it difficult to capture effective microbial signals based on library preparation methods that rely on poly-A capture. In addition, low-complexity regions of the genome and contamination of the database genome itself may lead to mismatches [46].

Currently, macrogenome is the commonly used microbiome, but the lack of poly-A makes it only a partial solution despite switching to strand-specific library preparation [47]. At the same time, the sheer size of the data makes the results unsatisfactory after normalization. Fletcher et al. failed to reproduce the results of the original article, even though they chose a higher precision analysis process [48, 49], and Gihawi et al. pointed out a failing in one study [50, 51]. In that study, the normalization process could leak sample label information and amplify the technical noise; therefore, the machine learning models misused these false signals. Despite debates by Sepich-Poore et al. [52] and Austin et al. [53]in microbiome analysis, tissue samples with low microbial mass are believed to be susceptible to contamination interference and poor reproducibility. Subsequent studies should adopt stricter contamination control and more robust standardization strategies. In the future, an accurate standardization method should be uniformly applied by researchers to reduce controversy. Efforts in microbial single-cell transcriptomics still largely focus on cultured strains and investigate the diversity of cell wall structures in complex communities, the low abundance of bacterial RNA. The vast majority of bacterial RNA is rRNA and needs to be removed to increase the depth of mRNA sequencing. mRNAs have poor stability, which makes them prone to degradation and leads to their short half-life. At the same time, similar to macrogenomes, there is an urgent need for good standardized processes for noise reduction [54]. The design and use of 16 S rRNA bacterial-specific primers can overcome poly-A dependence. The designed primers are now abundant and can be systematically obtained from some databases [55, 56]. Accumulating evidence suggests that the internal organization of the bacterial transcriptome has functional consequences [40]. However, spatially relevant information is not available in the single-cell microbiome. With the improvement of FISH-related technologies, bacterial-MERFISH performs high-throughput, spatial analysis of thousands of longitudinal subunits in a single bacterium and effectively avoids errors caused by poly-A deficiency [57]. Future microbiomics analyses under space technology can compensate for the lack of nucleic acid data by relying on novel FISH techniques or through Raman spectroscopy.

In conclusion, the field of TM research is booming with the increasing application of technology; however, many challenges remain, including sample collection [58], small sample size [52, 59], low reproducibility [16], pollution interference [51], and high cost [59]. In modern TM detection technology, a combination of multiple technologies is often used for detection and validation. In the future, multi-omics and multi-modal analysis will be the focus of the development of TM detection technology, which will lead to a technological revolution in TM sequencing and bring TM research into a new era.

Overview of the microbiome in cancer

After decades of research, the characteristics of cancer have been deeply analyzed through host factors, such as somatic mutations, metabolic alterations, and immune escape [66, 67], but new ideas suggest that the microbiome is the first marker of cancer, and plays a key role in the development, diagnosis, and treatment of cancer [68]. There are nearly 1 trillion microbial species on Earth, but only 11 are recognized as carcinogens by the International Association of Cancer Registries. These “oncomicrobes” are responsible for approximately 2.2 million cancer cases each year, accounting for 13% of the total global cases, with well-studied pathogenic and epidemiological characteristics [69, 70]. In view of the complex role of microbes in host-associated cancer characteristics, there are still core scientific questions that need to be explored. For instance, the drivers of microbiome divergence, evolutionary pressure, and functional classification must be investigated. By integrating existing evidence, we comprehensively reviewed the functional properties of the microbiome in TME from the mechanistic pathway to the clinical application and discussed its translational implications.

Niche and origin of the TM

The potential effects of the microbiome on cancer development, progression, and treatment response are being delineated. In pathological conditions, each microbial niche can mediate community-level interactions that promote carcinogenesis through microbial dysbiosis [71, 72]. These interactions may occur directly among microorganisms or indirectly via the secretion of metabolites and effector proteins that affect tumor progression [73, 74]. Microbial members can circulate among different organ niches (Fig. 1). For instance, oral-native Fusobacterium species can migrate from the mouth, through the digestive tract or bloodstream, proliferate, and colonize in the large intestine, thereby leading to the progression of colorectal cancer (CRC) [75].

Fig. 1.

Niche and origin of the TM. (A) Multi-directional niche sharing of the microbiome among different organs. (B) Mucosal barrier invasion. The microbiome invades and colonizes the tumor through the damaged mucosal barrier. (C) Normal adjacent tissue (NAT). The microbial composition of tumor tissue is highly similar to that of NAT. (D) Concomitant tumor co-metastasis. The microbiome can migrate with the primary tumor to the site of the metastatic tumor. (E) Circulatory system transmission. The microbiome potentially migrates and colonize tumors via blood or lymphatic pathways. CRC, Colorectal cancer; TM, Tumor microbiome; TME, Tumor microenvironment; PMN, Premetastatic niche; ICT, Immune checkpoint blockade therapy; MLNs, Mesenteric lymph nodes

The origins and migration pathways of microorganisms within the tumor tissue can be classified into four categories (Fig. 1). Firstly, during tumorigenesis, the mucosal barrier of organs with exposed cavities, such as the gastrointestinal tract, respiratory tract, oral cavity, and urogenital tract, may be damaged, which allows microorganisms that originally colonize the mucosal surface to invade the tumor [76, 77]. Tjalsma et al. proposed a bacterial driver-passenger model, suggesting that species from the genus Bacteroides and the family Enterobacteriaceae colonize the colonic mucosa and “drive” tumor progression. During the carcinogenic process, changes in the microenvironment lead to the gradual replacement of driving bacteria by “passenger” bacteria, including probiotics, opportunistic pathogens, and symbiotes, which further facilitate tumor advancement [78]. Whether these microorganisms establish a predetermined ecological niche or if they represent a transient random colonization remains unclear. As an example, the pancreas—traditionally regarded as sterile—has been found to be colonized by TM, which migrate into the pancreas via the pancreatic duct from the compromised intestinal mucosal barrier, thereby reshaping the TME and increasing susceptibility to microbial translocation [79–81]. The second category involves normal adjacent tissue (NAT) that is recognized as a potential source of TM. The microbial composition in tumor tissue closely resembles that of NAT; however, notable differences exist in bacterial prevalence and the abundance of metabolic-related enzymes. For example, breast cancer (BC) is associated with increased bacterial diversity and increased levels of enzymes linked to anaerobic respiration [16]. This indicates that certain specific microorganisms may be crucial in tumor development. On the contrary, some researchers propose that the striking similarity in microbial communities between tumor sites and NAT may result from the migration of microorganisms from the tumor microenvironment into NAT [10, 12]. Further research is required to substantiate this notion as specific microenvironments within tumors may facilitate microbial colonization, characterized by factors such as immunosuppression, hypoxia, and nutrient-rich metabolic conditions, which should be verified through comprehensive metagenomic sequencing and gene identification [10]. The third category concerns co-metastasis associated with tumors. A study employing 16 S rRNA gene sequencing identified a correlation between the relative abundance of Fusobacterium and its associated microbiome in both primary CRC tumors and liver metastases. The dominant microbial genera in liver metastases closely mirrored those present in primary tumors, suggesting a stable pairing of microbial communities between Fusobacterium-positive primary and metastatic tumors. Furthermore. Nearly identical active Fusobacterium strains were detected in matched primary and metastatic CRC samples, thereby confirming the persistence of active Fusobacterium during the metastatic process, indicating that Fusobacterium may migrate alongside CRC cells to metastatic sites [82]. Additionally,F. nucleatum is associated with the microbiome in distant liver metastases of CRC, suggesting that it may also co-metastasize with tumors [43, 83]. However, the co-metastasis of bacteria with tumor cells also exists as metachronous metastasis. Intestinal bacterial migration occurs before the spread of tumor cells and forms favorable conditions for the colonization of tumor cells through premetastatic niche (PMN) [84]. The fourth category focuses on microbial dissemination through the circulatory system, which presents a promising avenue for microbiome-based cancer modulation. In the context of CRC, Escherichia coli disrupts the intestinal vascular barrier, enabling its entry into the bloodstream and facilitating the migration of microorganisms along the gut-liver axis. This subsequent colonization of the liver induces the formation of PMN [84]. In a mouse model of lung cancer, intestinal Akkermansia muciniphila (Akk) has been shown to regulate the TM and metabolic structure. Akk migrates into the bloodstream and colonizing lung cancer tissues, thereby influencing the tumor-associated microbiome and reprogramming tumor metabolism to inhibit tumor progression [85]. Furthermore, oral bacteria, such as F. nucleatum, colonize tumor tissues via hematogenous pathways, where they interact with Gal-GalNAc expressed in CRC through a Fap2-dependent mechanism [75]. Similar findings have been reported in mouse models of BC, where this interaction facilitated tumor growth and metastatic progression [7]. In addition, bacteria can migrate from the intestine to the mesenteric lymph nodes, either independently or via CX3CR1hi mononuclear phagocytes, which can subsequently reach the pancreatic region [80, 86]. The lymphatic pathway of microbial translocation is pivotal for augmenting extraintestinal anti-tumor immune responses. In preclinical melanoma models, immune checkpoint blockade therapy (ICT) causes dilatation of tall endothelial venules and expansion of lymphatic sinusoid in mesenteric lymph nodes(MLNs), disrupting the “mucosal firewall” function, allowing gut bacteria to travel through the blood to reach distant tumor sites and activate immune responses [87]. Moreover, tumor microenvironments characterized by immunosuppression, hypoxia, and nutrient-rich metabolism may potentiate microbial colonization, a phenomenon that awaits confirmation through whole-genome sequencing and genetic identification [10].

In conclusion, the primary sources of TM include mucosal invasion, NAT, the circulatory system, and tumor-associated co-metastasis. Nevertheless, the precise mechanisms by which these microbes infiltrate tumors, evade immune clearance, and establish stable microbial communities characterized by specific tumor subtypes remain a compelling research area. Scientists are actively working to accurately trace the migration pathways and transmission modes of tumor-associated microbiomes, indicating that there is significant potential for valuable research in this field.

Heterogeneity of the TM across cancer types

The TME exhibits high heterogeneity with significant variations in the composition and abundance of TM across different cancer types, owing to the diverse origins of the microbes present within tumors [88]. Recent advances have conducted comprehensive large-scale analyses of TM encompassing seven cancer types–brain cancer, pancreatic cancer, lung cancer, ovarian cancer, BC, melanoma, and bone cancer–along with their associated NAT. This analysis revealed that each tumor type is characterized by a unique microbial composition [16]. Recent years have seen remarkable progress in the understanding of TM composition, highlighting the necessity of investigating TM in various cancers to elucidate their roles in cancer progression. In this context, we present a summary of the primary microbial compositions associated with different cancer types in recent large-scale cohort studies (Table 2).

Table 2.

Brief summary of TM heterogeneity in various cancer types

Microorganisms	Cancer type	Methods	Status	Effect	Refs
F. nucleatum	Colorectal cancer	Whole-genome shotgun sequencing	Increase	Promoted inflammation and damage to DNA	[156]
		FISH assay	Increase	Improved the efficacy of anti-PD-1 therapy	[93]
		qPCR 16 S rRNA gene sequencing Microbial culture FISH assay	Increase	activated an oncogenic PI3K-AKT-NF-κB cascade and promoted tumorigenesis	[157]
		qPCR FISH assay	Increase	Promoted the polarization of M2-like macrophages	[158]
		qPCR 16 S rRNA gene sequencing Microbial culture	Increase	Promoted tumor metastasis	[82]
	Breast cancer	16 S rRNA gene sequencing FISH assay	Increase	Remodeled ITME and hindered ICB	[134]
	Esophageal squamous cell carcinoma	MeRIP-qPCR IHC	Increase	Mediated m6A methylation	[147]
	Oral squamous cell carcinoma	16 S rRNA gene sequencing FISH assay	Increase	Enhanced the bidirectional communication between OSCC cells and macrophages	[159]
F. nucleatum subsp. animalis clade 2	Colorectal cancer	PacBio single-molecule real-time sequencing	Increase	Induced pro-cancer metabolic changes	[90]
H. pylori	Gastric cancer	scRNA-seq Immunofluorescence IHC qRT-PCR	Increase	Mediated intercellular communication and regulated cellular microenvironment	[160]
		qRT‑PCR Immunofluorescence IHC	Increase	Disrupted DNA mismatch repair mechanisms	[161]
		qPCR IHC	Increase	Affected genetic instability	[95]
	Gastric MALT lymphoma	Histological examination of H. pylori	Increase	Promoted chronic inflammation	[97]
	Liver cancer	16 S rRNA amplicon sequencing	Increase	Related to carcinogenic processes	[101]
Streptococcus	Liver cancer	16 S rRNA amplicon sequencing Shotgun metagenomic sequencing IHC FISH assay	Increase	Related to tumor progression	[105]
	Gastric cancer	16 S rRNA gene sequencing	Increase		[162]
	Oral cancer	16 S rRNA V3V4 amplicon sequencing	Decrease		[125]
	Lung cancer	Nanopore sequencing technology	Increase	Associated with nicotine exposure	[126]
S. anginosus	Gastric cancer	Metagenomic sequencing scRNA-seq	Increase	Reshaped the TME and enriched pro-inflammatory Th17 cells and immunosuppressive macrophages	[96]
Malassezia	Pancreatic cancer	Fungal DNA sequencing IHC FISH assay	Increase	Activatived complement pathway and induced tumorigenesis	[48]
Human Papillomavirus16	Oral cancer	Digene Hybrid Capture 2(HC2) assay IHC	Increase	Related to disease diagnosis and prognosis	[116]
Epstein–Barr virus		Microarray analysis	Increase	Induced mitochondrial stress	[118]
Herpes Simplex Virus Type 1		qRT-PCR	Increase	Mediated STING signaling pathway	[119]
S. mutans		FISH assay Bacteria culture IHC	Increase	Increased production of tumor metabolites, reprogramming highly ITME	[120]
Modestobacter	Lung cancer	16 S rRNA gene sequencing	Increase	Related to tumorigenesis	[127]
Acidovorax		16 S rRNA gene sequencing FISH assay	Increase	Related to TP53 mutations, impaired epithelial function	[128]
Proteus	Lung cancer	Multi-omics analysis	Increase	Related to anti-tumor immunity	[131]
	Breast cancer	16 S rRNA gene sequencing	Increase	Related immune responses	[133]
Pseudomonas	Pancreatic cancer	Metagenomic sequencing	Increase	Associated with tumorigenesis and induction of inflammation	[113]
	Breast cancer	16 S rRNA gene sequencing	Increase	Related immune responses	[133]
Porphyromonas	Oral cancer	16 S rRNA V3V4 amplicon sequencing	Decrease	Related to tumor progression	[125]
	Endometrial Cancer	16 S rRNA V3-V5 amplicon sequencing	Increase	Related to high vaginal PH	[140]
L. crispatus	Endometrial Cancer	qPCR 16 S rRNA V1-V2 amplicon sequencing	Decrease	Related to endometrial organoid proliferation and inflammation	[141]
Cutibacterium acnes	Prostate cancer	Microarray analysis	Increase	Related to inflammatory response	[144]
Gammaproteobacteria	Pancreatic cancer	qPCR 16 S rRNA gene sequencing IHC FISH assay	Increase	Related to anti-tumor immunity and chemotherapy	[110]
B. fragilis	Breast cancer	16 S rRNA gene sequencing IHC FISH assay	Increase	Promoted cancer cell stemness and chemoresistance	[136]
Lactobacillus	Gastric cancer	Ultrahigh performance liquid chromatography 16 S rRNA gene sequencing	Increase	Related to tumor progression; metabolic regulation	[163]
Streptococcaceae	Breast cancer	16 S rRNA gene sequencing	Increase	Related to tumor subtype and development	[137]

Digestive system cancers

Colorectal cancer

The intratumoral microbial community exhibited significant heterogeneity, with varying abundances of tumor-associated bacteria, such as Fusobacterium, Parvimonas, Prevotella, and Bacteroides, at different stages of CRC. Throughout the adenoma-carcinoma pathology, the relative abundance of F. nucleatum notably increased from mucosal carcinoma to more advanced stages, potentially reflecting microbial metabolic activity. This suggests that changes in the microbiome participate in CRC progression [89]. Additionally, in tumor cells located at distant metastatic sites, Bacteroides, Selenomonas, Fusobacterium, and Prevotella species have been identified, implying a possible link between these bacteria and CRC metastasis [82]. Through genomics and functional experiments, Zepeda-Rivera et al. identified an evolutionary branch of F. nucleatum, Fna C2 (F. nucleatum subsp. animalis clade 2). Animal models confirmed that Fna C2 can significantly increase the burden of intestinal adenomas and induce pro-cancer metabolic changes. Its unique metabolic and virulence genes can enhance its survival and carcinogenicity in the acidic environment of the intestine [90]. Moreover, distinct microbial profiles were observed in CRC based on the tumor location. Specifically, Romboutsia and Bifidobacterium exhibited a higher abundance in left-sided CRC, whereas Veillonella and Haemophilus showed a greater abundance in right-sided CRC. This microbial difference may reflect the difference in biological subtypes between right-sided and left-sided CRC, but it is not necessarily the decisive factor [91]. This finding explains the biological subtype differences between right-sided and left-sided CRC, with increased levels of the “microsatellite instability (MSI)” immune consensus molecular subtype 1 and “metabolic” CMS3 observed in right-sided CRC [92]. A recent study demonstrated that high levels of F. nucleatum enhanced the sensitivity of microsatellite stable (MSS) CRC to programmed death 1 (PD-1) blockade. The underlying mechanism involves the production of substantial amounts of butyrate by F. nucleatum, which epigenetically activates transcription factor TBX21 (T-bet). This activation inhibits PD-1 expression in CD8 + tumor-infiltrating lymphocytes (TILs), thereby increasing their cytotoxicity of CD8 + TILs and their capacity to eliminate tumor cells. This enhances the efficacy of anti-PD-1 therapies [93].

Gastric cancer

Helicobacter pylori (H. pylori) is a major risk factor for Gastric cancer (GC) and has been classified as a type I carcinogen [94]. For instance, Costa et al. demonstrated through in vivo and in vitro experiments that H. pylori can inhibit the expression of the transcriptional regulatory factor USF1, and the absence of USF1 affects the progression of GC by promoting the degradation of the key tumor suppressor protein p53 mediated by MDM2 (the main E3 ubiquitin ligase of p53) [95]. Studies have found that the non- H. pylori pathogen Streptococcus anginosus (S.anginosus) can induce gastritis and drive precancerous lesions and tumorigenesis of the gastric mucosa. Mechanism speaking, S.anginosus binds to the ANXA2 receptor of gastric epithelial cells through the surface protein TMPC, mediates colonization, and thereby activates the MAPK signaling pathway, promoting cell proliferation and inhibiting apoptosis. Furthermore, S.anginosus reshapes the TME, enriches pro-inflammatory T helper cells 17(Th17) and immunosuppressive macrophages, and synergistically promotes tumor progression [96]. Gastric mucosa-associated lymphoid tissue (MALT) lymphoma (gastric MALT lymphoma) is a low-grade malignant lymphoma that originates from MALT. After H. pylori infection, Chronic active inflammation can form on the surface of the gastric mucosa. This long-term inflammatory stimulation can lead to the proliferation of mucocutaneous lymphoid tissue, thereby triggering a series of immune responses and promoting the clonal proliferation of B lymphocytes. Eventually, it may develop into gastric MALT lymphoma [97]. In a C57BL/6J mouse model, Matsui et al. verified that the lactic acid bacteria strain Lactobacillus gasseri SBT2055 (LG2055) could inhibit the colonization of Helicobacter suis by competing for binding sites or secreting lactic acid. LG2055 induced TGF-β and secretory IgA to enhance mucosal immunity. So as to effectively prevent the development of gastric MALT lymphoma [98].

Liver cancer

The liver is intricately connected to the intestine and serves as the primary target of metabolites derived from gut microbiota and microbe-associated molecular patterns (MAMPs). This connection is especially significant in cases of intestinal permeability, where bacteria and their metabolites can easily translocate to the liver [99]. H. pylori and similar species have been identified in the liver of patients with hepatocellular carcinoma(HCC). Although in vitro studies have indicated a potential relationship between H. pylori infection and liver cancer development, there is currently no evidence that this species directly induces tumor formation [100–102]. Chronic liver disease is frequently associated with developing liver cancer, primarily due to viral hepatitis. Notably, studies have revealed that approximately 56% of liver cancers are related to Hepatitis B Virus (HBV), whereas 20% are associated with Hepatitis C Virus (HCV) [103, 104]. Recently, there has been a growing focus on the oral-gut-liver axis. A prospective cohort study involving patients with HBV-HCC identified the presence of several genera, including Stenotrophomonas maltophilia, Parabacteroides distasonis, Streptococcus salivarius, and Streptococcus mitis, within tumor tissues. The synergy of the oral gut microbiome has been translated into new microbial characteristics that serve as early detection signals for the onset of HCC [105]. Interestingly, despite the bile ducts being directly connected to the intestine, research on the role of the gut microbiome in cholangiocarcinoma is scarce. A study analyzed tumor samples from patients with cholangiocarcinoma and identified a tissue-specific microbiome primarily consisting of Dietziaceae, Pseudomonadaceae, and Oxalobacteraceae. Notably, the species Stenotrophomonas exhibited a substantial increase in tumor tissues compared to paired normal tissues [106].

Pancreatic cancer

The presence of this TM in pancreatic cancers, particularly pancreatic ductal adenocarcinoma(PDAC), has been demonstrated in several studies [80, 107, 108].Gammaproteobacteria, the predominant taxon associated with PDAC, may influence anti-tumor immunity and chemotherapy resistance [109, 110]. Additionally, H. pylori and HBV have been identified as potential risk factors for pancreatic tumorigenesis [111, 112]. The “basal-like” subtype of PDAC is characterized by a highly invasive nature and a unique intratumoral microbiome, where increases in the abundance of Sphingopyxis, Pseudomonas, and Acinetobacter are closely linked to cancer development and inflammation [113]. Moreover, the microbial diversity in PDAC tumors surpasses that in normal pancreatic tissues. Research indicates that tumors in long-term survivors exhibit significant enrichment of Sacharopolyspora, Streptomyces, and Pseudoxanthomonas, along with higher densities of granzyme B + cells and CD3 + and CD8 + T cells, potentially facilitating tumorigenesis by suppressing innate and adaptive immune responses [114]. The use of antibiotics may disrupt the balance of microbiota, reduce commensal bacteria with immunomodulatory effects, and weaken chemotherapy-induced antitumor immunity. This study reveals for the first time that antibiotic use is significantly associated with shortened survival in patients with metastatic PDAC, suggesting that gut microbiota may be a new target for optimizing the efficacy of chemotherapy. The need for rigorous evaluation of antibiotic use in clinical Settings is called for, although further mechanistic validation is needed [115]. In addition to bacterial changes, the fungal population in PDAC tissues has been found to increase by approximately 3,000 times compared to that in normal pancreatic tissues, with a notable enrichment of Malassezia identified in both mouse and human fungal populations. This phenomenon may arise from the activation of the MBL-C3 cascade via the interaction of Mannan-Binding Lectin (MBL) with polysaccharides on the fungal cell wall, thereby promoting tumor progression [48].

Non-digestive system cancers

Oral cancer

Oral squamous cell carcinoma (OSCC) is the most prevalent malignant tumor of the oral cavity, and it has long been recognized that viral infections may influence the risk of developing this disease. HPV has emerged as a significant potential etiological factor, with evidence suggesting that HPV infection increases the risk of OSCC three-fold, particularly for HPV type 16, the most critical subtype [116, 117]. Other oncogenic viruses, including Herpes Simplex Virus Type 1 (HSV-1) and Epstein-Barr Virus (EBV), have also been implicated in promoting OSCC progression [118, 119]. Using a rat model, a recent study revealed that tumor-colonizing Streptococcus mutans (S.mutans) stimulates the overproduction of kynurenic acid (KYNA) by OSCC cells via its protein antigen c (PAc). KYNA overproduction promotes the expansion and infiltration of highly immunosuppressive neutrophils into the TME. These neutrophils further exacerbate immune suppression by secreting IL-1β, which leads to CD8 + T cell exhaustion and accelerates the progression of OSCC [120]. F. nucleatum and Porphyromonas gingivalis (P. gingivalis) serve as critical risk factors in OSCC and oral precancerous lesions, exhibiting potent tumor-promoting activities [121]. Both F. nucleatum and P. gingivalis are significantly more abundant in OSCC tissues than in normal tissues and facilitate epithelial cell proliferation while suppressing apoptosis, altering the inflammatory microenvironment, and generating carcinogenic metabolites [122–124]. The composition of the oral microbiome evolves with the progression of oral cancer. Yang et al. conducted 16 S rRNA sequencing on samples from patients with different stages of OSCC. With the progression of OSCC, the abundance of Fusobacterium increased at the genus level, but the abundance of Streptococcus, Haemophilus, Porphyromonas and Actinomyces decreased [125].

Lung cancer

The pulmonary mucosa is directly connected to the external environment, making it susceptible to microbial exposure and environmental stimuli while harboring diverse microbial communities [126]. Although the microbial composition in lung cancer tissues was highly similar to that in NAT, alpha diversity was significantly lower in cancer tissues than that of NAT. A recent study using nanopore sequencing to analyze full-length 16 S of bacteria revealed significant enrichment of Streptococcaceae, Lactobacillaceae, and Neisseriales in tumor tissues. Additionally, in nicotine-exposed groups, the abundance of Haemophilus paralinfluenzae and Streptococcus gordonii was significantly higher, thereby demonstrating a relationship between the microbiome and smoking [126]. The prevalence of Propionibacterium and Enterobacteriaceae is lower in cancer tissues than that in NAT tissues, whereas Modestobacter is more abundant in lung tumor tissues [127]. Histological subtypes and stages of lung cancer are also associated with the heterogeneity. Compared to lung adenocarcinoma (LUAD), lung squamous cell carcinoma (LSCC) exhibits a more diverse microbiome with interactions between microbial exposure and genetic mutations. The relative abundance of Acidovorax, Klebsiella, and Anaerococcus increases, especially in tumors with TP53 mutations, which can impair epithelial function and are associated with increased levels of Acidovorax [128]. Lung cancer did not appear in germ-free or antibiotic-treated mice, even in the presence of KRAS mutation and p53 deletion. Commensal lung microbiota promotes the proliferation and activation of γδT cells, and these cells produce effector molecules, such as IL-17, which trigger a pro-tumor inflammatory response [129]. It is worth noting that Aspergillus sydowii can also be enriched in LUAD and trigger the secretion of IL-1β through the β-glucan /Dectin-1/CARD9 signaling axis, thereby inducing the expansion of myeloid-derived inhibitory cells (MDSCs). Furthermore, it inhibits the function of CD8 + T cells and increases the infiltration of PD-1 + T cells, thereby accelerating tumor progression and worsening the prognosis of patients [130]. The interaction between the host and microbiome is crucial for establishing the immune landscape of the lungs. In LUAD, the MS2 molecular subtype shows increased neutrophils and γδT cells, and Desulfococcus, Terrabacter, Bacteroides, Proteus, and Neisseria are closely correlated with γδT cells, potentially influencing tumor progression and leading to poorer prognosis [131].

Breast cancer

BC is the leading cause of cancer-related mortality in women. The microbiome is widely recognized as a significant factor associated with the risk and development of BC [132]. Among all tumor types, BC exhibit the highest levels of microbial diversity and abundance [16]. Tumor-resident microbiota plays a critical role in the development, progression, and immune microenvironment of BC, and its diversity and functional characteristics are closely associated with the pathogenesis of breast cancer [3, 10]. The microbial composition of breast tumor tissues is markedly distinct from that of normal tissues, with Proteus and Pseudomonas being highly enriched in BC tissues but virtually absent in non-tumor tissues [133]. The abundance of F. nucleatum in BC is negatively correlated with the immunotherapy response and promotes a large influx of myeloid cells into the immunosuppressive tumor microenvironment (ITME), thereby impairing the efficacy of immune checkpoint blockade (ICB) therapy [134]. F. nucleatum can also colonize triple-negative breast cancer (TNBC) tissues, contributing to ITME modulation and tumor metastasis [135]. Enterotoxigenic Bacteroides fragilis (ETBF) enhances stemness and chemoresistance of BC cells [136]. Moreover, detailed analyses revealed that TM was correlated with race, BC subtype, and disease stage. The family Streptococcaceae is more abundant in TNBC, and the genus Bosea load increases with BC progression [137].

Reproductive system neoplasms

Ovarian and endometrial cancers

The microbial community in ovarian cancer is likely to be complex, and local microbial dysbiosis may influence the occurrence or progression of ovarian cancer. Infection with Chlamydia trachomatis, Neisseria gonorrhoeae, or Mycoplasma genitalium is clinically correlated with ovarian cancer susceptibility and progression. Specifically, Chlamydia trachomatis infection increases the overall risk of ovarian cancer through persistent inflammation and chromosomal aberrations [138, 139]. Compared with that of NAT, the proportion of HPV in ovarian cancer is higher, and high-risk human papillomavirus (HR-HPV) types 16, 18, and 45 are significantly associated with advanced tumor stages [10]. Atopobium and Porphyromonas are particularly enriched and associated with endometrial cancer [140]. In a study using endometrial organoid models, Porphyromonas, Prevotella, Peptoniphilus, and Anaerococcus were significantly enriched in patients with endometrial cancer compared to benign uteri, whereas the relative abundance of Lactobacillus (including L. crispatus) was reduced. L. crispatus exerts anti-inflammatory and anti-mitotic effects on endometrial cells [141].

Prostate cancer

Viral infections, including HPV, EBV, and Human Herpesvirus, have been implicated in developing prostate cancer (PCa). These viruses may contribute to PCa pathogenesis through various mechanisms such as direct disruption of cellular genomic stability, modulation of cell survival signaling pathways, and interaction with the host immune system, thereby significantly influencing the pathophysiology of PCa [142]. Davidsson et al. found that co-culturing Propionibacterium acnes (P. acnes) with prostate cell line PNT1A changed cell proliferation and increased the secretion of cytokines and chemokines. This result supports the hypothesis that P. acnes infection promotes the development of PCa [143]. In addition, elevated levels of IL-6 and IL-8 were reported to be associated with the development and metastasis of advanced PCa, suggesting an inflammatory response in infected P. acnes [144].

Other cancers

Numerous studies have identified the presence of microbes in other tumor types in addition to tumors of the digestive, non-digestive, and reproductive systems. In melanoma, the microbiome modulates chemokine levels and CD8 + T-cell infiltration, thereby influencing patient survival. Specifically, Algibacter and Epilithonimonas are negatively correlated with CD8 + T cells, whereas certain genera, including Lachnoclostridium, Gelidibacter, and Flammeovirga, are positively correlated with CD8 + T cells [145]. Nasopharyngeal carcinoma is associated not only with EBV infection but also with the presence of certain bacteria in tumor tissues, with Corynebacterium and Staphylococcus being predominant. The total bacterial load within the tumor is negatively correlated with prognosis [146]. F. nucleatum also promotes the development and metastasis of esophageal squamous cell carcinoma (ESCC). This mechanism involves methyltransferase-like protein 3(METTL3)-mediated m6A methylation during F. nucleatum infection, which stabilizes mRNA in a YTHDF1-dependent manner, thereby promoting ESCC proliferation and metastasis [147]. Currently, research on microbes in intracranial tumors remains limited, but microbes have been detected in pituitary neuroendocrine tumors (PitNETs) with varying abundances across different subtypes. Corynebacteriaceae, Aerococcaceae, and Clostridiales are more enriched in growth hormone-secreting PitNET tissues, whereas Fusobacteriaceae, Tissierellaceae, and Aerococcaceae are substantially enriched in adrenocorticotropic hormone-secreting PitNET tissues [148].

These findings indicate that there are significant microbiome differences among different types of cancer. From the perspective of host genetic factors, somatic mutations of key genes can lead to the aggregation of unique bacterial populations. In LSCC, TP53 gene mutations were associated with an increased abundance of Acidovorax [128]. Microorganisms capable of degrading cigarette smoke metabolites (such as toluene) were found to be enriched in lung cancer and upregulated the related metabolic pathways, suggesting that these metabolites formed suitable niches for specific microorganisms [16]. However, many reports have only discovered differences in the microbiome among different types of cancer or different subtypes of the same type of cancer [16, 127, 149], but the specific molecular mechanisms remain unexplored. The interaction patterns between microorganisms and human cancers and their dynamic changes with cancer progression are mainly driven by TME. The unique characteristics of TME, such as hypoxia and necrosis, form a highly selective niche. The multi-dimensional differences have selective effects on the colonization of microorganisms, thereby driving specific evolutionary adaptation strategies [150, 151].

The “role assignment” of TM is also an interesting direction worth exploring. Some carcinogenic microorganisms can contribute to cancer development through mutations induced by genetic toxins. In particular, bacteroides fragilis toxin (BFT), reactive oxygen species (ROS) producers [152], and colibactin (a DNA alkylating agent) [153] are closely related to the mutant characteristics observed in CRC, head and neck cancer, and cancers of the urinary system. The “complicit” microbes are not directly carcinogenic. These microbial communities and their metabolites are involved in the development of tumors mainly by regulating the host immune response. For example, secondary bile acids produced by microbial metabolism can downregulate the expression of CXCL16, a specific CXCR6 ligand, in hepatic sinusoids, thereby inhibiting the recruitment of CXCR6 + NKT cells, impairing immune surveillance in the liver and promoting tumorigenesis, which was reversed by vancomycin [154]. Certain “passenger” bacteria, such as Clostridium and Streptococcus, cannot colonize healthy intestines, but they gain competitive advantages in the TME of CRC due to barrier permeability and metabolic changes, gradually replacing the main pathogenic driver bacteria. Symbiotic bacteria (such as Colibacter) may delay disease progression, but their role remains unclear [78, 155].

In summary, TM is present across diverse types of cancer and is characterized by significant diversity and heterogeneity. The structure and abundance of TM differ markedly among various types, subtypes, and stages of cancer. Despite these findings, the relationship between certain types of cancers and TM remains underexplored, highlighting the need for further scientific investigation.

TM affects signaling pathways involved in cancer progression and immune response

Signaling pathways, including those of cell movement, growth, survival, and metabolism, are integral to the fundamental activities of normal and tumor tissues. TM is crucial in tumor-associated signaling cascades. By modulating diverse signal transduction pathways, TM facilitates cancer initiation and progression while also participating in immune regulation. These pathways are frequently implicated in cell survival, growth, proliferation, inflammation, invasion, and stem cell regulation. (Fig. 2).

Fig. 2.

TM coordinates multiple signal transduction cascades affecting cancer progression and participates in immune regulation. WNT/β-catenin signaling: H. pylori, Salmonella, and B. fragilis elicit β-catenin signaling activation through E-cadherin-mediated phosphorylation events, which indirectly or directly stimulate downstream gene transcription, driving cellular proliferation, migration, and invasion; TLR signaling: F. nucleatum LPS is recognized by TLR4, triggering the TLR4/M MyD88 cascade and activating the NF-κB pathway to promote chronic inflammation; MERK signaling: P. gingivalis derived gingipains stimulate the RAS-RAF-MEK-ERK signaling cascade; PI3K signaling: E6/E7 oncoproteins of HR-HPV contribute to changes in the PI3K/AKT/mTOR signaling pathway, facilitating cell growth and proliferation; STING signaling: LGG induces IFN-I production via the cGAS-STING-TBK1-IRF7 cascade in DCs, while Akk activates the STING/IRF3/IFN-I pathway with c-di-AMP to polarize anti-tumor macrophages. CagA, Cytotoxin-associated gene A; BFT, Bacteroides fragilis toxin; TCF, T-cell factor; c-Myc, Cellular Myelocytomatosis oncogene; MMPs, Matrix metalloproteinases; RMI2, RecQ-mediated genome instability protein 2; BXD, Banxia Xiexin Decoction; AvrA, Avirulence protein A; TLR, Toll-like receptor; LPS, Lipopolysaccharide; MYD88, Myeloid differentiation primary response protein 88; EpOME, Epoxy-Octadecenoic Acid; ETBF, Enterotoxigenic Bacteroides fragilis; NFAT5, Nuclear Factor of Activated T-cells 5; JMJD2B, JmjC-domain-containing histone demethylase 2 B; NANOG, Nanog homeobox; SOX2, Sex-determining region Y-box 2; DCs, Dendritic cells; STING, Stimulator of interferon genes

WNT/β-catenin signaling pathway

The WNT/β-catenin signaling pathway is pivotal in embryonic development, adult tissue homeostasis, and regeneration. It also governs the self-renewal and growth properties of cells, and cancer initiation and progression [164]. Changes in the β-catenin signaling pathway are frequently observed in various malignancies, including CRC, GC, and primary liver cancer, and may be influenced by certain TM [165, 166]. H. pylori infection promotes developing GC through multiple mechanisms. Some strains express cytotoxin-associated gene A (CagA); when these strains colonize the stomach, the risk of peptic ulcer disease increases and is associated with a higher risk of gastric adenocarcinoma [167]. Direct injection of CagA protein into the cytoplasm of host cells modulates β-catenin, thereby promoting cell proliferation and tumor progression. CagA binds to the cytoplasmic domain of E-cadherin, disrupts the E-cadherin-β-catenin complex, induces nuclear translocation of β-catenin, and activates the Wnt-β-catenin pathway, which is essential for the self-renewal of cancer stem cells [168, 169]. Salmonella releases the virulence factor AvrA via a type III secretion system, inducing various post-translational modifications of β-catenin, including phosphorylation and acetylation, which collectively regulate its activity. Phosphorylation of Thr41 and Ser-33/37 enhances the ubiquitin-mediated degradation of β-catenin, while phosphorylation of Ser-552 promotes nuclear β-catenin signaling, a key event in stem cell activation [170, 171]. AvrA upregulates the phosphorylation of Ser-552 while downregulating the phosphorylation of Ser-33/37 and Thr41, thereby enhancing nuclear β-catenin signaling and acetylation while inhibiting ubiquitin-mediated degradation of β-catenin [172, 173]. Reports have indicated that the N-terminal phosphorylation and C-terminal acetylation of β-catenin robustly activate TCF [174, 175]. β-catenin binds to the transcriptional complex TCF and translocates to the nucleus, where it stimulates transcription or downstream target genes, such as c-Myc, cyclin D1, MMPs, or survivin. When the complex binds to the TCF-binding site in the RMI2 promoter, it induces the migration and invasion of EMT markers and HCC [176]. Bacteroides fragilis (B. fragilis) expresses BFT, which induces the rapid cleavage of E-cadherin. Initially, BFT stimulates shedding of the extracellular domain of E-cadherin, followed by activation of host cell γ-secretase to cleave intracellular E-cadherin. This proteolysis triggers TCF-dependent nuclear β-catenin signaling [12, 152]. In addition to its adhesin, Fap2 directly interacts with the host polysaccharide Gal-GalNAc [75]. another F. nucleatum adhesin, FadA, is essential in the invasion of epithelial and CRC cells. FadA interacts with E-cadherin, leading to β-catenin translocation and expression of downstream Wnt-β-catenin target genes [177]. Specifically, FadA adhesion causes phosphorylation of E-cadherin on the cell membrane, followed by clathrin-mediated internalization. This reduces phosphorylated β-catenin levels, allowing β-catenin to accumulate in the cytoplasm and translocate to the nucleus, activating transcription factors such as TCF and NF-κB and oncogenes such as Myc and cyclin D1 [178]. Concurrently, inflammatory genes, such as interleukin-6 (IL-6), IL-8, and IL-18, are upregulated [177]. Recent studies demonstrate that Banxia Xiexin Decoction (BXD) serum inhibits FadA binding to E-cadherin, reducing F. nucleatum adhesion and invasion in CRC cells and downregulating the E-cadherin/β-catenin signaling pathway [179]. Conversely, Bifidobacterium adolescentis (B.a) inhibits CRC via the Wnt/β-catenin pathway. Specifically, B.a activates and induces high expression of GAS1 in CD143 + cancer-associated fibroblasts (CAFs) via the Wnt/β-catenin pathway. GAS1 exhibits tumor-suppressive effects, offering a novel therapeutic target for probiotic-based TME modulation [180].

Toll like receptor signaling pathway

TLRs function as pattern recognition receptors (PRRs) that detect microbe-associated molecular patterns (MAMPs) such as lipopolysaccharide (LPS), flagellin, and peptidoglycan. Disruptions in TLR-microbiome interactions are associated with inflammatory disease states [181, 182]. TLRs share a conserved domain structure comprising an extracellular leucine-rich repeat domain for pathogen-associated molecular pattern (PAMP) recognition, a transmembrane domain, and a cytoplasmic Toll/IL-1 receptor domain that initiates downstream signaling [183]. Upon ligand binding, TLRs recruit signaling molecules via the myeloid differentiation primary response protein 88 (MyD88) pathway or MyD88-independent mechanisms, triggering downstream signaling cascades. This activation stimulates pathways, such as the canonical and MAPK signaling, ultimately leading to the upregulation of inflammatory cytokines [184]. TLR signaling plays a complex role in TM-induced effects, with microbes capable of triggering TLR signaling in tumor cells to promote tumor progression [185]. For instance, F. nucleatum binding to TLR4 on CRC cells activates AKT signaling, mediating an increase in the downstream metabolite 12,13-EpOME, thereby promoting CRC development and EMT-mediated metastasis [186]. The LPS of F. nucleatum is recognized by TLR4, activating the TLR4/MyD88 cascade to stimulate the NF-κB pathway, enhancing tumor cell proliferation and chronic inflammation [187]. In a TLR4-dependent manner, F. nucleatum LPS induces miRNA-21 expression, activates autophagy in CRC cells, and confers chemoresistance [187, 188]. A recent study demonstrated that specific miRNA mimics are TLR antagonists, inhibiting their expression, thereby limiting the release of pro-inflammatory and pro-tumorigenic cytokines, leading to tumor cell apoptosis. This suggests that the integration of TLRs and miRNAs can affect CRC pathogenesis [189]. Additionally, ETBF.

upregulates JmjC-domain-containing histone demethylase 2 B (JMJD2B) via a TLR4-NFAT5-dependent pathway, increasing the expression of stem cell transcription factors such as sex determining region Y-box 2 (SOX2) and Nanog homeobox (NANOG). This indicated that ETBF LPS may be crucial in promoting stemness regulation associated with CRC development [190].

NF-κB signaling pathway

Inflammation in tumor development is the key determinant of cancer progression [191]. The NF-κB signaling pathway is crucial for chronic inflammation induced by microbes. As previously reported, microbes can activate NF-κB via β-catenin and TLR signaling pathways, leading to the release of inflammatory factors and inducing a chronic inflammatory state. This pathological process is bidirectional: immune cells recruited during chronic inflammation release inflammatory mediators that further activate NF-κB signaling. These interactions form a positive feedback loop that drives tumor development and progression [192, 193]. In addition to Fap2 and FadA, recent studies have identified another F. nucleatum adhesin, RadD, which directly binds to the overexpressed receptor CD147 on CRC cells, initiating the PI3K-AKT-NF-κB-MMP9 cascade and enhancing CRC development; targeting the RadD-CD147 axis is a potential therapeutic strategy for CRC [157]. Furthermore, F. nucleatum can upregulate CD11b + myeloid cells and inflammatory markers in CRC through NF-κB-driven modulation of the immune microenvironment, thereby promoting tumor development [178, 194]. Multiple studies have shown that BFT not only cleaves E-cadherin but also activates multiple MAPKs in the intestinal epithelium, leading to NF-κB pathway activation, increased secretion of inflammatory factors, recruitment of polymorphonuclear cells, and induction of mucosal inflammatory responses [152, 195–198].

MAPK/ERK signaling pathway

Microbes can activate the ERK signaling pathway in tumor cells through direct or indirect mechanisms. The ERK pathway, also known as the Ras-RAF-MEK-ERK cascade, is highly conserved. In this pathway, MAPKK (i.e., RAF) phosphorylation activates MEK, which in turn activates ERK. Activated ERK is translocated to the nucleus, where it regulates transcription factors and gene expression associated with cell growth and proliferation [199, 200]. H. pylori infection is associated with GC and is detected in 90% of MALT lymphoma cases, a rare non-Hodgkin lymphoma originating from gastric B cells [201]. CagA, a key virulence factor, is phosphorylated by the Src family of kinases upon entry into host cells. It interacts with SHP-2 to form a CagA-SHP-2 complex that activates the ERK-MAPK pathway, promotes cell proliferation, and inhibits apoptosis. This complex also upregulates the expression of the anti-apoptotic proteins Bcl-2 and Bcl-XL [201, 202]. Additionally, intracellular P. gingivalis can adhere to and invade host cells and stimulate the RAS/RAF/MEK/ERK pathway through gingipains, thereby enhancing the phosphorylation of MEK1/2 and ERK1/2 and promoting CRC cell proliferation [203].

PI3K signaling pathway

Phosphatidylinositol 3-kinase (PI3K) is a critical lipid kinase that regulates cellular functions, such as proliferation, survival, metabolism, and cancer cell invasion. Aberrant activation of the PI3K/Akt/mTOR signaling pathway is implicated in oncogenesis. PI3K serves as a key coordinator of intracellular and extracellular signals by integrating growth factors, cytokines, and external stimuli into intracellular signals that modulate diverse signaling pathways [204]. Comparative analyses of lower airway transcriptomes revealed significant differences between patients with lung cancer and controls, with notable upregulation of the PI3K signaling pathway in patients with cancer. Specifically, enrichment of oral microbial communities in the lower airways of patients with lung cancer, including Streptococcus and Veillonella, is closely associated with PI3K upregulation. Consistent with these findings, in vitro studies have demonstrated that airway epithelial cells exposed to Veillonella, Prevotella, and Streptococcus exhibit enhanced PI3K signaling [205]. Furthermore, genomic analyses of bronchial airway lesions in smoking patients with lung cancer showed a significant increase in PI3K pathway activation, suggesting that dysregulation of this pathway is an early reversible event in lung cancer development [206]. Viral infections that modulate the PI3K pathway are frequently associated with cancer progression. For example, HR-HPV infection is strongly associated with head and neck carcinomas. The oncogenic mechanism of HR-HPV involves the overexpression of E6/E7 oncoproteins, which drive changes in the PI3K/AKT/mTOR signaling pathway, thereby promoting cell proliferation, migration, and invasion and accelerating tumor progression [207]. Similarly, oncogenic proteins from other viruses, such as the envelope protein gp120 of HIV [208], X protein encoded by HBV (HBx) [209], and oncoprotein LMP1 encoded by EBV [210], modulate this pathway to facilitate tumor formation.

cGAS-STING signaling pathway

STING is a PRR that binds to cyclic dinucleotides (CDN) and exerts its effects on cancer by inducing the production of type I interferon (IFN-I) and inflammatory cytokines, as well as promoting regulated cell death. STING has emerged as a research focus in the fields of antiviral and anti-tumor immunity [211]. cGAS is recognized as the primary DNA sensor, detecting double-stranded DNA and catalyzing the synthesis of the second messenger 2’3’-cyclic GMP-AMP (cGAMP) from ATP and GTP. In addition to cGAMP, STING can also bind to other released CDN during bacterial infections, such as cyclic di-GMP (c-di-GMP) and cyclic di-AMP (c-di-AMP), thereby inducing IFN-β synthesis independent of cGAS [212]. In several preclinical cancer models, microbiota from a high-fiber diet, such as Akk, produce STING agonists (e.g., c-di-AMP). These agonists induce IFN-I production by intratumoral monocytes and polarize them toward an anti-tumor macrophage phenotype. They also modulate the crosstalk between natural killer (NK) cells and dendritic cells (DCs) [213]. In a mouse cancer model, Lactobacillus rhamnosus GG (LGG) synergized with ICB to enhance anti-tumor responses. Mechanistically, LGG activates the cGAS-STING-TBK1-IRF7 signaling cascade in DCs, leading to IFN-I production and enhanced cross-priming of anti-tumor CD8 + T cells [214]. During infection, bacteria release DNA, and cytosolic bacteria trigger the release of host mitochondrial DNA (mtDNA) to regulate the STING pathway. For example, Salmonella typhimurium infection induces the release of both host mtDNA and bacterial DNA, which accumulate in the host cytoplasm and trigger an interferon gamma response [215]. Studies have shown that the bacterial DNA that activates the cGAS-STING pathway may originate from bacterial outer membrane vesicles (OMVs). For instance, P. gingivalis OMVs upregulate the NF-κB signaling pathway, which downregulates STING expression [216]. However, the mechanisms underlying STING activation by many microbes remain unclear, specifically, whether they are mediated by cGAS recognition of bacterial DNA or other bacterial products. Thus, investigating the interactions between bacteria and their hosts may uncover new mechanisms related to STING activation, offering novel insights into cancer development, immune regulation, and the design of therapeutic strategies.

Other signaling pathways

Beyond the signaling pathways previously discussed, TM may activate additional, more specific signaling pathways. Research indicates that bacteria residing within BC cells can inhibit the RhoAROCK signaling pathway, enabling tumor cells to withstand mechanical stress in the vascular environment. This inhibition protects tumor cells from damage, facilitates cytoskeletal remodeling, and promotes distal colonization, thereby accelerating tumor metastasis [217]. In TNBC, the TM-related metabolite trimethylamine N-oxide (TMAO) induces ferroptosis in tumor cells by activating the endoplasmic reticulum stress kinase PERK, thereby enhancing CD8 + T cell-mediated anti-tumor immune response against TNBC [218]. Additionally, TM activates the β-catenin signaling pathway, leading to increased c-Myc expression. Increased levels of c-Myc significantly enhance both transcription and translation within the endoplasmic reticulum, resulting in the accumulation of misfolded proteins and subsequent activation of the PERK signaling pathway, thus providing further evidence that microorganisms can induce endoplasmic reticulum stress [219]. Moreover, the α5-nicotinic acetylcholine receptors (α5-nAChRs)-Notch signaling pathway promotes the proliferation, migration, and invasion of melanoma cells [220].

In conclusion, the TM plays a complex and multidimensional role in regulating cancer biology, involving a variety of cellular signaling pathways, immune regulatory networks, and metabolic interactions. Our current understanding of the signaling pathways influenced by these microbial mechanisms constitutes “a corner of the land” in this expansive knowledge domain. No study has definitively determined whether and how different microbial species specifically affect subsets of these signaling pathways. There remains a significant knowledge gap concerning the mechanisms by which these microorganisms mediate signaling pathway regulation and the specific roles these pathways play in cellular physiological and pathological states.

Molecular mechanisms how TM influence carcinogenesis or cancer prevention

Microbial infections exhibit both promoting and inhibiting effects during cancer development, thus playing a dual role that is crucial in oncology research. A comprehensive understanding of how tumor-associated microorganisms influence the mechanisms underlying cancer progression is essential for accurate prognosis assessment and formulation of effective treatment strategies (Fig. 3).

Fig. 3.

Mechanisms of the TM affecting cancer progression. (A) Gene structural abnormalities and epigenetic modifications. (B) Secretion of TM metabolites affects TME. (C) Promotion of tumor invasion and metastasis. (D) TM dysbiosis triggers carcinogenesis associated with inflammation. SCFAs, Short-chain fatty acids; LCA, Lithocholic acid; AhR, Aryl hydrocarbon receptor; EMT, Epithelial–mesenchymal transition; ICAM-1, Intercellular Adhesion Molecule-1; ALPK1, Alpha Kinase 1; ROS, Reactive oxygen species; NICD, Notch intracellular domain; TLRs, Toll-like receptors

Gene structural abnormalities and epigenetic modifications

The TM facilitates the initiation and progression of tumors by inducing genomic instability and structural abnormalities [9]. Within the TME, specific microorganisms and their secreted bacterial toxins are essential for inducing DNA damage. These microorganisms directly interact with the genomes of the colonized tissues, resulting in an increased frequency of DNA mutations. As these mutations accumulate and reach a critical threshold, they disrupt the regulation of cell growth and ultimately trigger tumorigenesis [221, 222]. A prime example are E. coli strains harboring the polyketide synthase (pks) gene (designated pks + E.coli), which has been implicated in somatic mutations and DNA damage in CRC [223, 224]. Pleguezuelos-Manzano et al. Co-cultured human intestinal organoids (derived from healthy intestinal stem cells) with pks⁺ E. coli to simulate long-term bacterial exposure. They found that these bacteria can induce specific single base substitution (SBS-pks) and small fragment deletion (ID-pks). It has been proven that bacterial exposure can lead to specific mutations, thereby increasing the risk of CRC. Interestingly, in addition to CRC, SBS-pks and ID-pks mutations have also been detected in head and neck cancer and tumors of the urinary tract system [153]. E.coli can penetrate the colonic mucosal layer and colonize intestinal polyps, which are potential precursors of cancer, producing an unstable DNA alkylating agent known as colibactin [225]. Colibactin is a toxin produced by E. coli, which can lead to DNA damage and characteristic mutations, namely SBS88 and ID18. The latest findings showed that SBS88 and ID18 are more common in early-onset CRC, and the mutational burden in patients under 40 years old is 3.3 times higher than that in patients over 70 years old. This indicates that early exposure to colibactin plays an important role in the development of CRC [226]. Colibactin targets adenine-rich DNA regions, leading to the formation of DNA interstrand crosslinks and double-strand breaks (DSBs) in human cells. This type of DNA damage activates the phosphorylation of related proteins, resulting in temporary cell cycle arrest and cell swelling and contributing to cytotoxicity and mutations that promote tumorigenesis [153, 227]. Other carcinogenic bacteria, such as F. nucleatum, activate the E-cadherin/β-catenin pathway through a FadA-dependent mechanism, upregulating checkpoint kinase 2 (CHK2) in mouse models of CRC. This process not only facilitates DNA damage but also promotes tumor cell growth [228]. Moreover, F. nucleatum infection can activate the Ku70/p53 signaling pathway, leading to dependency-related DSBs and further supporting developing OSCC [229]. Additionally, small-molecule metabolites produced by microorganisms can influence tumor development by promoting DNA damage. For instance, a study identified that small molecules from various human gut microbiota directly impaired DNA in acellular assays, inducing the expression of DSB markers (γ-H2AX) and causing epithelial cell cycle arrest. Specifically, indolimine derivatives from Morganella morganii (M. morganii) exacerbate colon tumorigenesis in germ-free mice [230]. Another study revealed that enteropathogenic strains of E. coli and H. pylori expressing an effector protein may disrupt the DNA mismatch repair system, thereby further destabilizing the genome and promoting tumorigenesis [231, 232]. The activity of these microorganisms is closely linked to the production of ROS, hydrogen sulfide, and nitric oxide [233–235]. For example, BFT produced by ETBF increases ROS levels in HT29/c1 and T84 intestinal epithelial cells, resulting in oxidative DNA damage and malignant transformation. The underlying mechanism involves BFT enhancing the expression of spermine oxidase (SMO), triggering SMO-dependent ROS production and upregulating DNA damage markers (γ-H2A) [166, 236].

Numerous studies have demonstrated that viruses can induce DNA damage, manipulate host systems, cause uncontrolled cell transformation and division, and ultimately promote cancer progression. For example, Merkel cell polyomavirus (MCPyV) can integrate DNA clones into tumor genomes and continuously express the viral T antigen, contributing to at least 60% of Merkel cell carcinoma cases [237, 238]. Human T-cell leukemia virus type 1 (HTLV-1), a retrovirus associated with lymphoproliferative diseases, such as adult T-cell leukemia/lymphoma, produces a viral protein known as Tax, which not only hinders DNA repair but also correlates with the accumulation of DNA strand break foci and increased phosphorylation of H2AX [239].

Epigenetic modifications, including DNA methylation, noncoding RNA regulation, and histone modifications, are critical mechanisms by which microorganisms modulate gene expression during cancer progression. H. pylori (especially CagA-positive strains) induces abnormal hypermethylation of the promoter region of tumor suppressor genes in gastric mucosa by activating inflammatory signaling pathways, such as NF-κB/STAT3, and upregulating DNA methyltransferases, which is a key factor affecting the pathogenesis of gastric adenocarcinoma [240]. Additionally, the intricate interplay between changes in DNA methylation and microbial characteristics may facilitate distal metastasis in patients with gastric adenocarcinoma and affect prognosis. Experimental evidence has confirmed that S. saccharolyticus enhances tumor cell metastasis by modulating the expression of ZNF215 [241]. Bacterial species, such as Hungatella hathewayi and F. nucleatum have been identified as key players in the regulation of DNA methylation. These bacteria induce epigenetic changes in host cell tumor suppressor genes through the action of DNA methyltransferases, thereby promoting CRC cell proliferation [242]. Conversely, some studies have suggested that microorganisms regulate host cell DNA methylation through enzyme-independent mechanisms. This highlights the distinct patterns of methyl donor-related microbial pathways in CRC tumor tissues compared to normal tissues, suggesting complex influences of the microbiome on host DNA methylation patterns [243]. Importantly, microbial metabolites often mediate complex interactions between tumor-associated microbiota and epigenetic variations. For example, short-chain fatty acids (SCFAs) derived from microbes induce genomic epigenetic changes by modulating the activities of histone acetyltransferases and deacetylases [244]. SCFAs enhance the expression of hypoxia-inducible factor 1 (HIF-1) by inhibiting histone deacetylase activity, which increases the production of antimicrobial effectors and improves their macrophage-mediated clearance from the microbiome [245]. Additionally, butyrate produced by tumor-associated bacteria increases histone H3K27 acetylation at the H19 promoter, leading to M2 macrophage polarization, increased H19 expression in tumor cells, and enhanced lung cancer metastasis [246]. Furthermore, Mycoplasma hominis infection increases m6A modification levels in RNA molecules, facilitating the degradation of mitochondrial fusion protein 1 mRNA. This process exacerbates mitochondrial fission, contributing to monocyte polyploidization and the acquisition of cancer stem cell characteristics [247]. HBV also exerts significant effects on viral replication, immune evasion, and cancer induction through the induction of m6A modifications in RNA, underscoring the complex molecular interactions between the host and virus [248].

Microbial-derived metabolites reshape the TME

Microbial metabolites accumulate in the TME and function as ligands for specific receptors and regulatory factors that influence protein activity, thereby acting as significant modulators of immune response [249]. SCFAs, such as butyrate, propionate, and acetate, are produced through microbial fermentation of dietary fibers and have shown significant anti-tumor effects in the prevention and treatment of CRC [249–251]. Butyrate regulates CD4 + T cell differentiation by binding to DCs and the GPR109a receptor on the surface of macrophages. In the APCMin/+ mouse model, this mechanism can increase the proportion of Tregs and IL-10 + CD4 + T cells while reducing the abundance of pro-inflammatory Th17 [252]. IL-10 enhances the immunosuppressive function of Tregs by promoting STAT3 phosphorylation, thereby inhibiting Th17-mediated inflammatory response and blocking chronic inflammation-driven progression of colon cancer [253]. In addition, studies have indicated that butyrate can induce the apoptosis of CRC cells by downregulating the genes associated with histone deacetylase activity. Furthermore, butyrate enhances the expression levels of granzyme B and IFN-γ in CD8 + T cells while also regulating the glycolytic pathway, tricarboxylic acid (TCA) cycle, and fatty acid oxidation in anti-tumor immune cells, thereby enhancing the efficacy of immune checkpoint inhibitors (ICIs) [43, 254]. Despite their potent anti-cancer potency, SCFAs may promote cancer development. When the F. nucleatum strain F7-1 and the altered Schaedler’s flora (ASF) colonize the intestinal tract of germ-free mice, SCFAs produced by F. nucleatum bind to the free fatty acid receptor 2 (FFAR2). This binding not only increases the risk of intestinal cancer but also promotes the expression of IL-17, inducing an inflammatory environment suitable for tumorigenesis [255]. In addition, the cross-regulation of fungi-bacterial metabolism-immunity is critical in cancer. Commensal bacteria maintain intestinal metabolic homeostasis by producing SCFAs, activate anti-tumor CD8 + T cells and maintain Th1/Th17 immune balance. Fungal β-glucan induces Th2 polarization through the Dectin-1 signaling pathway, triggering M2 macrophage expansion and T cell depletion, which together shape ITME and accelerate tumor progression. The symbiotic imbalance of the two factors directly drives treatment resistance by enhancing Th2 response [58, 256].

Although previous studies suggested that inosine possesses immunosuppressive properties [257],recent investigations have revealed that it can reprogram the TME and enhance the effectiveness of immunotherapy [258, 259]. In the presence of exogenous IFN-γ and co-stimulatory signals from DCs, inosine promotes the differentiation of Th1 by binding to the adenosine A2A receptor (A2AR) on T cells, significantly increasing Th1 cell anti-tumor activity in CRC, melanoma, and bladder cancer [73]. Conversely, in the absence of exogenous stimuli, the inosine produced by Lactobacillus reuteri(Lr) inhibits the differentiation of Th1 and Th2 cells through A2AR signaling [260]. Primary bile acids can be metabolized by gut microbiota into secondary bile acids, with lithocholic acid (LCA) being a notable product [261]. LCA has a dual role in the TME: it facilitates carcinogenesis in CRC by upregulating Micro-RNA-21 expression through Erk1/2 signaling, AP-1 transcription factor, and STAT3 [262]. Conversely, two derivatives of LCA, 3-oxoLCA and isoalloLCA, exhibit anti-cancer properties. Specifically, 3-oxoLCA inhibits the differentiation of Th17 cells by directly binding to the retinoic acid receptor-related orphan receptor-γt (RORγt), while IsoalloLCA enhances Treg cell differentiation through the generation of mitochondrial reactive oxygen species (mitoROS), thus suppressing colitis [263]. Studies have shown that specific p53 mutations have a dual role in the gut environment: when gallic acid produced by microbial metabolism is lacking, these mutations exhibit protective properties. However, in the presence of gallic acid, the oncogenic phenotype was transformed. This phenomenon was confirmed in in vivo models and in organoid experiments, revealing the complex interplay between microbial metabolites and host gene function [264]. Indole, a metabolite of tryptophan, has been demonstrated to activate aryl hydrocarbon receptor (AhR), thereby promoting macrophage polarization, inhibiting inflammatory T cell infiltration, and facilitating cancer cell growth. The absence of AhR or its pharmacological inhibition in myeloid cells leads to reduced growth of PDAC, enhances the efficacy of immune checkpoint blockade, and promotes an increased accumulation of tumor necrosis factor-α(TNFα) + IFNγ + CD8 + T cells within tumors [265]. Another study demonstrated that Lr can migrate to the melanoma tissue and release indole-3-acetaldehyde (I3A), a tryptophan metabolite, in the TME, thereby activating AhR in CD8 T cells, promoting the production of IFN-γ, and enhancing the anti-tumor immune response. Although Lr colonizes the gut, its main anti-tumor effect relies on I3A release from the TME rather than on the overall alteration of the gut microbiota [266].

Invasion and metastasis

Tumor cells undergo intrinsic changes to adapt to unfavorable conditions during invasion and metastasis. TM can modulate these adaptive programs. For example, activation of the EMT program enhances cancer cell invasion and metastasis [267–270]; increased stemness facilitates tumor growth at secondary sites following metastasis [271, 272]; induction of tumor cells adhesion prevents apoptosis [273–275]; and modulation of the mechanical stress response offers resistance to cellular damage [217, 276]. Collectively, these processes enhance the invasion and metastasis of tumor cells.

The EMT program confers migratory mesenchymal traits to tumor cells while disrupting epithelial cell-cell adhesion, thereby facilitating their invasion and dissemination. This transformation is driven by the activation of the Wnt/β-catenin, TGFβ, and MAPK signaling pathways, along with coordinated transcriptional programs involving Zeb, Twist, and Snail, leading to developing a metastatic mesenchymal phenotype [270]. Significant associations between microorganisms and EMT have been documented in numerous studies. For instance, flagellin from Desulfovibrio vulgaris (DSV) increases the expression of EMT-related genes, such as N-cadherin and Vimentin, inducing a malignant phenotype in CRC cells via the TRAF6/TAK1 signaling pathway [268]. In CRC liver metastasis models, infection with F. nucleatum increased IL-8 expression by downregulating miR-5692a and inducing EMT via the ERK/ZEB1 axis [267]. Furthermore, F. nucleatum secretes a novel virulence protein, Fn-Dps, which promotes EMT through the upregulation of the chemokines CCL2 and CCL7, thereby enhancing CRC cell invasion and metastasis in murine models [269]. Maintenance of plasticity and stemness in tumor cells is crucial for the initiation of metastasis. Studies have indicated that in human BC cell lines, colonization enhances distal organ metastasis in human breast cancer cell lines. This bacterium cleaves E-cadherin, promoting the nuclear translocation of β-catenin and the accumulation of the Notch effector molecule NICD in the nucleus. Consequently, activation of the Wnt and Notch signaling pathways in mouse transplant tumor models results in enhanced stem cell characteristics, heightened tumor growth, and accelerated metastatic progression [271]. In a mouse GC model infected with CagA-positive H. pylori, there was a noticeable increase in the EMT and mesenchymal phenotypes of GES-1 cells, leading to enhanced efficiency of tumor sphere formation and increased CD44 expression, a known marker of gastric cancer stem cells (CSCs) [272]. Although TM has been shown to enhance the stemness characteristics of tumor cells post-metastasis, it is also possible that tumor cells enter a dormant state following metastasis only to reactivate when conditions become conducive. The influence of TM on the post-metastatic dormant state of tumor cells remains an underexplored domain, presenting a potential focal point for future research.

Cell adhesion molecules on the surface of tumor cells can significantly enhance their survival and help circumvent challenges during metastasis [273, 275]. For example, in human CRC cell lines, F. nucleatum mediates a unique PRR and ALPK1 activity, which enhances the NF-κB pathway and leads to the upregulation of the adhesion molecule ICAM1. Increased ICAM1 expression markedly fosters tumor cell adhesion to endothelial cells, facilitating tumor cell infiltration and the establishment of new metastatic foci in nude mouse tail vein experiments [274]. Circulating tumor cells are subjected to various mechanical stresses in the bloodstream that can result in cellular damage [276]. In an MMTV-PyMT mouse spontaneous breast tumor model, the invasion of Streptococcus, Staphylococcus, and Lactobacillus into host tumor cells inhibited the RhoAROCK signaling pathway, remodeled the actin cytoskeleton, and increased the resistance of tumor cells to mechanical stress, thus promoting their survival and distant organ metastasis [217]. Additionally, TM shape the surrounding microenvironment of tumor cells, indirectly facilitating metastatic processes. In CRC cells infected by F. nucleatum, hnRNPA 2B1 mediates the release of exosomes containing miR-122-5p and activates the FUT 8/TGF-β1/Smads axis, which promotes EMT and accelerates tumor metastasis [277]. This underscores that neighboring tumor cells do not have to be invaded by bacteria to instigate the metastatic process; instead, they can engage in paracrine signaling via exosomes, facilitating the initiation of metastasis.

Expression of inflammatory factors and chronic inflammation

The production of inflammatory factors and sustained chronic inflammation can create a microenvironment conducive to tumor development, thereby serving as a significant contributor to cancer induction [278]. The core of the association between inflammation and cancer lies in the abnormal transcription of genes encoding inflammatory mediators, growth factors, transport proteins, and angiogenic factors, as well as genomic instability and damage, and epigenetic regulatory dysfunction [279]. TM can interact with PRRs, such as TLRs, triggering the release of diverse cytokines and activating the NF-κB signaling pathway. This cascade forms a positive feedback loop that induces proinflammatory responses and promotes tumor progression [193]. Microbial agents can bind to TLRs on immune cells, thereby influencing the tumor immune microenvironment. Dysbiosis of the pulmonary microbiota, which is characterized by an imbalance between symbionts and pathogens, results in the production of microbial products (LPS and peptidoglycans) that activate TLRs. This activation leads to the stimulation of alveolar macrophages and neutrophils, increasing the local levels of IL-1β, IL-23, and activated lung γδ T cells that secrete IL-17, ultimately exacerbating local inflammation and further facilitating lung cancer progression [129]. Moreover, bacterial-derived LPS acts as a trigger that activates TLR4, mediating the CCL2-dependent recruitment of monocyte-like macrophages, which in turn promotes IL-1β production and enhances Th17 cell activation, thereby intensifying inflammation and establishing a precancerous inflammatory environment favorable for tumor genesis [280]. Interestingly, in gastrointestinal tumors, Candida spp. also drives the transformation of pancreatic intraepithelial neoplasia to adenocarcinoma by promoting the release of IL-1β. Consistently, the abundance of Candida spp. was found to be significantly and positively correlated with the risk of metastasis [281]. Additionally, components derived from Pseudomonas aeruginosa, including its flagella and cytotoxin ExoU, exhibit potent inflammatory activity by recruiting neutrophils while simultaneously activating the NF-κB signaling pathway, consequently accelerating the progression of oral cancer [282]. Deterioration of the intestinal epithelial barrier allows microbial products to invade and activate tumor-infiltrating inflammatory DCs, thereby inducing the polarization of γδ T17 cells, which release a considerable array of pro-inflammatory cytokines, including IL-17, IL-8, granulocyte-macrophage colony-stimulating factor (GM-CSF), and TNF-α. These cytokines further attract polymorphonuclear myeloid-derived suppressor cells (PMN-MDSCs), transforming the inflammatory microenvironment into an immunosuppressive state and promoting CRC progression [283]. Additionally, F. nucleatum-induced autophagy inhibition in Caco-2 colorectal adenocarcinoma cells leads to ROS accumulation, which triggers the production of pro-inflammatory cytokines such as IL-8, IL-1β, and TNF-α [284].

In addition to direct inflammatory stimulation, microbial components may affect tumor immune responses through molecular mimicry. For example, Fluckiger et al. found that the tail length measurement protein (TMP) of Enterococcus hirae (E. hirae) bacteriophage contains MHC class I (H-2Kb) limiting epitopes TSLARFANI (TMP1). The E. hirae strain with TMP1 can induce the anti-tumor immune response. In the mouse model, TMP1-specific CD8 + T cells can cross-recognize the PSMB4 peptide segment and inhibit the growth of MCA205 sarcoma and TC1 lung cancer. In anti-PD-1 therapy, cancer patients with enterococcus carrying TMP phages exhibited a longer survival [285]. Forty-one types of HLA-related bacterial peptides derived from bacteria were discovered by analyzing melanoma based on HLA peptidomics. Melanoma cells can present these peptides. In vitro experiments have shown that they can activate TILs, promote the secretion of IFN-γ, and trigger T-cell immune responses. This finding confirms that bacterial peptides are immunogenic and can serve as targets for immunotherapy [286].

In conclusion, TM plays a complex and dual role in the development of cancer. It may promote the development of cancer through multiple mechanisms. In some cases, it may also prevent tumorigenesis. In-depth research on the mechanisms by which TM is involved in cancer development is of great significance for the early diagnosis, prognosis assessment and precise treatment of cancer.

Microbiome-based therapy: an emerging and unexpected aid in cancer treatment

Recent advancements in the study of tumor-associated microbiomes have elucidated the intricate relationship between microorganisms and cancer treatment, spanning multiple types of therapeutic, including immunotherapy, chemotherapy, and radiotherapy, which offer a novel perspective for broadening cancer treatment strategies.

Immunotherapy

Increasing research on microorganisms and immunotherapy, particularly ICIs, has underscored their potential to alter the immunosuppressive landscape of the TME. However, heterogeneity in patient responses leads to efficacy only in selected patient cohorts [287]. Investigation of the underlying mechanisms revealed that the microbiome plays an unexpectedly pivotal role in modulating local and systemic immune responses to tumors. As a result, numerous studies are actively exploring therapeutic strategies aimed at microbiome modulation; notable examples include fecal microbiota transplantation (FMT) [288] and probiotics [289]. These investigations seek to restore gut and TM composition to enhance immunotherapy effectiveness and extend its benefits to broader patient populations, presenting a promising new therapeutic avenue [288, 289]. FMT has demonstrated the potential to enhance the suppressed immune function by reconstructing the microbial balance. A phase II clinical trial showed that the combination of FMT and the anti-PD-1 pembrolizumab achieved a disease control rate of 40% in the treatment of patients with PD-1-resistant advanced melanoma [290]. Studies have confirmed that there are significant differences in the composition of the intestinal microbiota between responders and non-responders to anti-PD-1 agents. Animal experiments also verified that after introducing FMT of respondents into germ-free mice, it can enhance CD8⁺ T cell infiltration, upregulate the expression of programmed death ligand 1 (PD-L1), and reduce the abundance of immunosuppressive cells, thereby enhancing anti-tumor immunity and inhibiting tumor growth. However, transplantation of fecal microbiota from non-responders did not have such effects [291]. In solid gastrointestinal tumors, FMT containing beneficial microbiota yielded an overall response and disease control rates of 7.7 and 46.2%, respectively, confirming that FMT promotes T-cell activity by enhancing cytotoxic T-cell infiltration, which may help overcome resistance to PD-1 inhibitors [292]. Studies have shown that gut microbiota regulates the PD-L2-RGMb pathway through downregulation. In particular, Coprobacillus cateniformis can inhibit PD-L2 expression on DCs, and conditional knockdown of RGMb on CD8 + T cells can enhance anti-tumor immune response and overcome resistance to PD-1/PD-L1 inhibitors. This study provides a new strategy for cancer immunotherapy combined with PD-L2-RGMb [293]. Moreover, FMT can alleviate small intestinal colitis induced by ICIs in patients with cancer, indicating its potential to address gastrointestinal dysfunction while enhancing immunotherapy outcomes [294]. However, variability in the safety profile, quantification methods, and practical implementation of FMT raises concerns regarding reproducibility and quality control. Despite the overall good safety profile of FMT, two cases of extended-spectrum β-lactamase-producing E.coli bacteremia with FMT were reported, one of which resulted in death, revealing the potential risk of spreading drug-resistant bacteria [295]. Recently, approaches that use automatic washing and related delivery methods for FMT, termed washed microbiota transplantation (WMT), have emerged to enhance the safety, quantification, and delivery of microbial suspensions during transplantation [296, 297]. The combination of immunotherapy with probiotics is another promising research topic [298]. Probiotics, which are live microorganisms, have been shown to enhance the efficacy of αPD-1 immunotherapy through the combined action of Lactobacillus johnsonii and Clostridium sporogenes, producing indole-3-propionic acid and thus improving ICB responsiveness across various cancers, including melanoma, BC, and CRC. This finding highlights the potential of this approach [299]. However, the indiscriminate use of commercial probiotics may not only fail to enhance immunotherapy effectiveness but could also trigger autoimmunity-related events [300]. For example, it has been shown that patients using commercially available probiotic supplements, such as Bifidobacterium and lactobacilli, exhibit poorer response to ICB. Animal models have shown that probiotics suppress IFN-γ + CD8 + T cell responses in TME and impair the efficacy of ICB, whereas a high-fiber diet can improve progression-free survival (PFS) in ICB-treated patients. A high-fiber diet is associated with greater microbial diversity and enrichment of beneficial bacteria. These microbiota may enhance the efficacy of ICB by activating immune checkpoints [301]. Additionally, the diversity of microbiome-derived immunoenhancers indicates that collaborative formulations incorporating multiple probiotics or prebiotics may optimize therapeutic effects. This underscores the prospective research trajectory that focuses on synthetic microbial communities as a sustained strategy for post-FMT microbiota reconstruction [302]. Therefore, it is essential to elucidate the potential mechanisms underlying these microbial interventions and develop tailored probiotic strategies suited to diverse populations based on their lifestyle and dietary habits. Furthermore, to reduce the risk of infection among immunocompromised patients with cancer, employing microbial-derived metabolites or prebiotics as alternatives to direct probiotic administration may be warranted [303, 304]. We have listed clinical trials on the microbiome and cancer treatment outcomes (Table 3).

Table 3.

Overview of clinical trials on the Microbiome and cancer treatment outcomes

Intervention	Cancer therapy	Cancer entity	Study phase	Endpoint	NCT numbers
Fecal Microbiota Transplantation(FMT)	Immune Checkpoint Blockade	Melanoma	Phase 2	Therapy response	NCT06623461
	Immune Checkpoint Inhibitors	Renal Cell Carcinoma	Phase 2	Therapy response	NCT04758507
	Immune Checkpoint Inhibitors	Non–small-cell lung cance, Melanoma	Phase 2	R​​esponse	NCT04951583
	Immune Checkpoint Inhibitors	Lung Cancer	Phase 2	R​​esponse	NCT05502913
	Immune Checkpoint Inhibitors	Melanoma	Phase 2	R​​esponse	NCT05251389
	Immune Checkpoint Inhibitors	Melanoma, Head and Neck Squamous Cell Carcinoma	Phase 2	R​​esponse	NCT05286294
	Radiochemotherapy	Colorectal cancer	Phase 4	Therapy response	NCT06931808
	Chemotherapy	Non–small-cell lung cance	Phase 2	R​​esponse	NCT06403111
Probiotics	Immune Checkpoint Inhibition	Periampullary Neoplasia	Phase 3	R​​esponse	NCT05736315
	Immune Checkpoint Inhibition	Colorectal cancer	Phase 2	Toxicity	NCT00936572
	Radiotherapy	Anal Cancer Squamous Cell	Phase 2	R​​esponse	NCT03870607
	Chemotherapy	Lung Cancer	Phase 1	R​​esponse	NCT02771470
Prebiotics	Immune Checkpoint Blockade	Melanoma	Phase 2	Therapy response	NCT01706393
	Immune Checkpoint Blockade	Unresectable Melanoma	Phase 2	R​​esponse	NCT06466434
	Immune Checkpoint Inhibition	Non–small-cell lung cance, Melanoma	Phase 1	R​​esponse	NCT05303493
Antibiotic	Immune Checkpoint Inhibition	Lung Cancer	Phase 2	R​​esponse	NCT06483347
	Immune Checkpoint Inhibitors	Metastatic Neuroendocrine Tumors	Phase 2	R​​esponse	NCT00027638
	Immune Checkpoint Inhibition	Non–small-cell lung cancer	Phase 2	Response​	NCT07001618
	Immune Checkpoint Inhibition	Prostate Cancer	Phase 2	Response​	NCT00108732
	Immune Checkpoint Inhibition	Breast Cancer	Phase 2	Therapy response	NCT06112379
	Chemotherapy	Neoplasms	Phase 2	Therapy response	NCT02366884
Diet	Immune Checkpoint Inhibitors	Solid Tumor	Phase 1	Response​	NCT04552418
	Immune Checkpoint Inhibitors	Melanoma	Phase 2	Response​	NCT06250335
	Radiotherapy	Malignant Neoplasm of Rectum, Malignant Neoplasm of Anus	Phase 3	Response​	NCT04534075

Derived from clinical trials registry of the NIH Library of Medicine

Although considerable research has been directed toward the gut microbiota, the role of TM in immunotherapy remains underexplored, particularly in contexts such as adoptive cell transfer (ACT) and chimeric antigen receptor (CAR) T-cell therapy. Moreover, interactions between the gut and TM, along with the potential impact of gut microbiome changes on TM composition and the characteristics of the host immune microenvironment, represent promising yet underdeveloped areas of scientific inquiry.

Chemotherapy

The TM significantly influences chemotherapeutic responses and numerous studies have identified microorganisms as pivotal factors in chemotherapy resistance. These findings suggest that microorganisms are promising therapeutic targets for enhancing the efficacy of chemotherapy [109, 305, 306]. For instance, F. nucleatum has been shown to fosters chemoresistance in CRC by modulating oxaliplatin resistance through the coordination of miRNAs (primarily miR-4802 and miR-18a), TLRs, and autophagy pathways [188].

ETBF was notably enriched in patients with BC who did not respond to taxane-based neoadjuvant chemotherapy. ETBF’s toxic protein BFT-1 of ETBF interacts with NOD1 to enhance stemness and drug resistance in BC cells. The co-expression of ETBF and NOD1 in tumor tissues signifies adverse chemotherapy responses in patients with BC. Additionally, inhibition of NOD1 via Nodinitib-1 and elimination of ETBF with metronidazole have been shown to enhance chemotherapy sensitivity by targeting breast cancer stem cells, suggesting a potential new avenue to combat chemotherapy resistance in BC [136].

Microorganisms also modulate bioactive forms of chemotherapeutic agents through enzymatic metabolic pathways. In a CRC mouse model, intratumoral Gammaproteobacteria can metabolize the active drug gemcitabine (2’,2’-difluorodeoxycytidine) into its inactive form (2’,2’-difluorodeoxyuridine) by expressing a long isoform of the bacterial enzyme cytidine deaminase (CDD_L), leading to gemcitabine resistance. This chemoresistance can be reversed with the antibiotic ciprofloxacin. In an extension of this research, 86 of 113 human PDAC cases were found to harbor intratumoral Gammaproteobacteria, specifically from the Enterobacteriaceae and Pseudomonadaceae families, suggesting that these intratumoral bacteria may contribute to gemcitabine resistance in PDAC [110]. Interestingly, fungal communities have a detrimental effect on chemotherapy, and studies indicate that the elimination of fungi can enhance the effectiveness of gemcitabine [48]. Additionally, the microbially derived tryptophan metabolite indole-3-acetic acid (3-IAA) has emerged as a key enhancer of chemotherapy response in humanized mouse models of PDAC [307]. Certain members of the TM family, including E. coli, B. fragilis, Bifidobacterium breve, and Prevotella micra, exhibit resistance to 5-fluorouracil (5-FU), a first-line chemotherapeutic agent. These microorganisms may convert 5-FU into its non-toxic form, thereby diminishing its efficacy and bioavailability, and undermining its ability to inhibit F. nucleatum [308]. Microorganisms can modulate the efficacy of chemotherapy through inflammatory responses and regulation of anti-tumor immunity. Cyclophosphamide promotes the migration of specific gram-positive bacteria to sub-lymphatic organs and activates a memory Th1 cell immune response through the action of pathogenic Th17. However, the pathogenic Th17 response was diminished in germ-free mice or in those with gram-positive bacteria eradicated by antibiotics, leading to tumor resistance to cyclophosphamide [309]. In conclusion, a comprehensive understanding of how TM influence resistance mechanisms can enhance the effectiveness of existing therapies. The interactions between TM and chemotherapeutic agents exhibit considerable plasticity, allowing for the potential customization of treatment strategies for individual patients with cancer.

Radiotherapy

Radiotherapy uses intense radiation to destroy tumor cells and reduce the tumor size. However, while effective, it also inflicts collateral damage on the surrounding healthy tissues. Emerging evidence suggests that microbiota may significantly influence the efficacy of radiotherapy [310]. Research has demonstrated that specific gut microbiota such as Bacteroidetes and Firmicutes can modulate the immune response to radiation, potentially enhancing therapeutic outcomes [311]. Prebiotics, which are selectively used by microbiota, have been shown to improve radiotherapy outcomes. For instance, a clinical trial in patients with gynecological cancer revealed that prebiotic intake from 1 week before to 3 weeks after radiotherapy improved post-treatment stool consistency [312]. Similarly, probiotics, which are live microorganisms that confer health benefits when consumed in adequate amounts, can alter the gut microbiota to induce immune responses to radiation, thereby improving the efficacy of radiotherapy [313]. SCFAs, particularly butyrate, have been shown to control the transition of the cell cycle from a radioresistant to a radiosensitive state. This effect is primarily mediated by butyrate, which enhances the transcriptional function of Forkhead Box O3A (FOXO3A) and induces cell cycle arrest by regulating p57, p21, and GADD45, significantly increasing the radiosensitivity of CRC patient-derived organoids (PDOs) [314]. Other microbial metabolites such as indole compounds have demonstrated anti-tumor properties in glioblastoma and may enhance the effects of radiotherapy [315]. FMT protects patients from radiation-induced damage. In a study involving five patients with chronic radiation enteritis, three showed significant improvements, including reduced diarrhea, rectal bleeding, abdominal/rectal discomfort, and fecal incontinence [316]. Additionally, the TME is a critical factor influencing the efficacy of radiotherapy. The TME differs from the normal internal microenvironment in several physical and chemical properties, including hypoxia, acidity, redox imbalance, and increased ATP levels [317]. Under hypoxic conditions, increased levels of signaling factors such as HIF-1 and vascular endothelial growth factor-A (VEGF-A), along with the upregulation of various gene signaling pathways, promote radioresistance in tumors, which is one of the most significant challenges in clinical oncology [318]. Therefore, innovative approaches are required to improve the hypoxic tumor microenvironment. Engineered microbes can deliver therapeutic agents directly to tumor sites, thereby destroying hypoxic tumor cells or releasing oxygen to enhance tumor circulation and improve drug delivery systems for cancer treatment. For example, several bacterial strains, including E. coli, can serve as effective delivery vehicles to target hypoxic cells in tumors, thereby enhancing the efficacy of radiotherapy [319]. In a Lewis lung carcinoma xenograft mouse model, Bifidobacterium infantis coupled with specific monoclonal antibodies effectively targeted and destroyed hypoxic tumor regions, reduced HIF-1α expression, enhanced radiation-induced DSBs, and induced tumor cell death [320]. Another strategy to increase the oxygen supply at the tumor site involves local treatment with botulinum neurotoxin type A within a safe dosage range. This treatment blocks the binding of the presynaptic membrane to vesicles, inhibits norepinephrine production at the neuromuscular junction, and increases tumor perfusion and oxygenation [321]. Given the limitations of radiotherapy for human tumors, monotherapy often fails to meet the clinical needs. Therefore, combining radiotherapy and microbiota-based therapies has emerged as a promising research direction. For instance, Listeria has been shown to be an effective carrier for delivering radionuclides to tumor sites. By incorporating ³²P into a phosphate-free medium, Listeria-³²P was developed. This strain kills tumor cells through ionizing radiation induced by ³²P and ROS generated by Listeria and has shown promise in pancreatic cancer treatment [322]. In a mouse model of melanoma, combining lipopolysaccharide-mutant Salmonella with X-ray therapy produced a super-additive effect compared with radiotherapy alone [323]. Similarly, combined treatment with γ-radiation enhances the killing effect in melanoma, thereby reducing its inherent radioresistance [324]. In summary, the emerging understanding of the interplay between the microbiota and radiotherapy offers intriguing possibilities for enhancing tumor control and mitigating the side effects of radiotherapy.

TM-related applications in cancer therapy: advancing towards precision oncology

Engineered bacteria

Advances in synthetic biology have led to significant preclinical progress in the utilization of engineered bacteria for cancer therapy [325–327]. These bacteria are tailored to produce cytotoxic molecules (Fig. 4) and disrupt the metabolic stability and structural integrity of tumor cells, ultimately triggering apoptosis. For instance, bacteria modified to secrete cytolysin A have demonstrated the ability to induce apoptosis in lung cancer, pancreatic cancer, and CRC mouse models by forming pores in the membranes of tumor cells, thereby effectively inhibiting tumor growth [328–330]. Similarly, the pore-forming toxin listeriolysin O (LLO), produced by Listeria monocytogenes, facilitates pore formation in tumor cell membranes, leading to cell lysis [331]. When engineered with tumor-targeting mechanisms, the non-pathogenic E. coli strain MG1655 effectively stimulated localized TNF-α production in murine tumors, resulting in significant anti-tumor activity [332]. Engineered bacteria that target tumors can produce prodrug-converting enzymes, thereby enhancing therapeutic efficacy (Fig. 4). A study found that intratumoral injections of attenuated Salmonella typhimurium (S. Typhimurium) strain VNP20009, which expresses E. coli cytosine deaminase, converted the prodrug 5-fluorocytosine (5-FC) into its active form, 5-FU, significantly increasing the levels of 5-FU within tumors in patients with cancer [333]. In another CRC mouse model, the natural compound baicalin, used as a glucuronide prodrug, exhibited enhanced anti-tumor effects when administered in combination with E. coli DH5α strains carrying the pRSETB-lux/βG plasmid [334]. Additionally, the genetically modified S. Typhimurium strain VNP20009 can deliver carboxypeptidase G2, demonstrating anti-tumor activity when combined with various substrate prodrugs [335]. Furthermore, bacteria can serve as effective vectors for delivering genetic material to tumor cells (Fig. 4). A study utilizing a bacterial genetic material delivery system based on Listeria monocytogenes highlighted the potential application of this strain in gene therapy for malignant tumors [336]. The researchers took advantage of the natural ability of Acinetobacter baylyi to design a biosensor capable of detecting specific DNA sequences. They engineered the bacteria to induce a specific phenotype (e.g., antibiotic resistance) upon the ingestion of DNA containing a specific tumor-associated mutation [337]. Additionally, a genetically engineered leucine and arginine auxotrophic strain of S. Typhimurium A1-R can induce the transition of tumor cells from G0/G1 to S/G2/M cell cycle phases, enhancing their susceptibility to chemotherapy [338].

Fig. 4.

TM-involved applications in cancer therapy. (A) Engineered bacteria exert a significant anti-tumor effect by producing cytotoxic molecules, converting prodrugs with specialized enzymes, delivering therapeutic payloads, and activating the immune response. Role of C. novyi-NT in targeted therapy within the hypoxic and necrotic regions of tumors. (B) OVs trigger ICD, releasing tumor-associated antigens and initiating an anti-tumor immune response. OVs equipped with cytokines can remodel the TME. The combined strategy of OVs with ICI and CAR-T cells can enhance therapeutic activity. (C) Antibiotics and bacteriophages manipulate the TM to achieve the targeted elimination of the carcinogenic microbiome.LLO, Listeriolysin O; TNFα,Tumor necrosis factor-α; TAAs, Tumor-associated antigens; PD-L1, Programmed death ligand 1, CTLA4, Cytotoxic T-lymphocyte-associated protein 4; PLC, Phospholipase C; BGNP, Branched gold nanoparticles; ITME, Immunosuppressive tumor microenvironment; ICD, Immunogenic cell death; ICI, Immune checkpoint inhibitor; TCR, T cell receptor; MHC, Major Histocompatibility Complex; PAMPs, Pathogen-associated molecular patterns; DAMPs, Damage-associated molecular patterns; DCs, Dendritic cells; OVs, Oncolytic viruses; CAR-T, Chimeric antigen receptor T-cell

The immune response stimulated by these engineered bacteria primarily depends on the activation of CD8 + T cells (Fig. 4),Genetically modified S. Typhimurium strains can secrete cytokines such as LIGHT130 or CCL21 [339], IL-18 [340], and IL-15 [341], which activate CD8 + T cells and promote the production of pro-inflammatory cytokines including TNF, IFN-γ, and IL-1β, thereby significantly enhancing anti-tumor immune responses across various mouse models. Furthermore, gene engineering allows the design of bacteria capable of delivering tumor-associated antigens (TAAs), which increase the presentation of TAA epitopes on MHC class I molecules, thus inducing tumor-specific CD8 + T cell activation [6, 342]. Notably, in the context of ICIs, an engineered probiotic system has been successfully developed for the precise production and targeted release of nanobodies against PD-L1 and cytotoxic T-lymphocyte-associated protein 4 (CTLA4), which have prompted sustained tumor regression and systemic anti-tumor immunity in preclinical models. This engineered probiotic system effectively integrated immunology and synthetic biology to enhance checkpoint blockade delivery [343]. Vincent et al. used engineered E. coli Nissle 1917 to achieve direct tumor-labeling by releasing a synthetic CAR target (“Tag”) within the tumor via a synchronous cleavage circuit. CAR-T cells are designed to recognize these synthetic tags and achieve specific tumor killing. Combining the tumor targeting of probiotics and the killing ability of CAR-T cells can overcome the antigen limitation of traditional CAR-T therapy, which provides a new idea for the immunotherapy of solid tumors [344].

However, the application of genetically engineered bacteria in cancer therapy has limitations. Establishing the optimal initial dosage and treatment regimen for treatment with live bacteria is challenging because the dynamic behavior and dose-response relationships of live bacteria differ significantly from those of conventional therapeutics. Therefore, it is imperative to implement measures to prevent contamination of the medical environment by live bacteria and ensure safety in clinical contexts.

Intratumoral injection of Clostridium novyi-NT

Resistance to conventional anti-cancer therapies and hypoxia in patients with advanced solid tumors are the two primary factors that contribute to the failure of standard treatments. Clostridium novyi-NT(C.novyi-NT)is an attenuated anaerobe that selectively germinates in the hypoxic and necrotic regions of tumors, secreting phospholipase C (PLC) after its spores germinate within the tumor tissues. This bacterium uses a dual mechanism that results in the lysis of tumor cells, while simultaneously activating the host anti-tumor immune response [345] (Fig. 4). In a study involving 20 dogs with advanced refractory sarcoma inoculated with C.novyi-NT spores, notable changes in tumor characteristics and immune markers were observed. The results indicated that C.novyi-NT disrupted the tumor vasculature, exacerbated hypoxia, and consequently facilitated gradual tumor regression. The treatment recruits neutrophils and macrophages to the tumor site, which in turn release IL-6 and IL-8 and initiate an inflammatory response. This response leads to a significant increase in CD8 + T cells and effector memory T cells, resulting in an “immune hot” phenotype in the tumor tissue post-treatment [346]. Moreover, another study illustrated the efficacy of combining C. novyi-NT treatment with Anti-Ly6G antibodies, enabling the specific depletion of neutrophils. This combination weakens the capacity of neutrophils to release ROS and arginase-1 (Arg1), which otherwise inhibits T cell function. Consequently, ITME was improved, significantly enhancing therapeutic outcomes [347]. This collaborative approach offers promising application potential (Fig. 4). A novel strategy involves developing branched gold nanoparticles (BGNP) to coat C.novyi-NT spores. This modification facilitates the real-time monitoring of tumor targeting and efficacy using computed tomography (CT) imaging. Compared to unmodified spores, BGNP-coated spores resulted in a tumor volume reduction beyond 40% in the CT-guided treatment groups. This study marks the first integration of nanomaterial engineering with oncolytic bacterial therapy to overcome the technical challenges associated with inaccurate tumor injection localization [348]. Given that glioblastomas are characterized by a high ITME, conventional therapies often yield limited success. Researchers have attempted to address this challenge by coating C.novyi-NT spores with the bee venom peptide-RADA 32 hybrid nanofibers incorporating metformin, an immunomodulator that enhances both innate and adaptive immune responses. This approach stimulates sustained CD8 + T cell activity, promotes the polarization of M1 macrophages, and upregulates the expression of HIF-1-α, PD-L1, and CXCL9 within the TME [349]. Additionally, several clinical trials involving C.novyi-NT have been initiated, with one study reporting a reduction in tumor burden and enhanced tumor-specific T-cell responses in patients with refractory advanced solid tumors following localized C.novyi-NT treatment [350]. In another study, C.novyi-NT was effective in reducing the tumor burden in a patient with advanced leiomyosarcoma. Following the injection of C.novyi-NT spores, CT was performed to assess the clinical characteristics pertinent to tumor cell eradication. Tissue biopsies have revealed widespread cytolysis and necrosis of tumor cells [351]. Nonetheless, reliance solely on oncolytic bacterial therapy may not fully eradicate all tumor cells, resulting in the risk of tumor progression. These findings underline the necessity for further in-depth clinical trials involving patients with cancer to enhance our understanding and treatment approaches.

Oncolytic viruses

Oncolytic viruses (OVs) represent a promising class of emerging immunotherapies for cancer that use the inherent ability of certain replicating viruses to preferentially infect and lyse tumor cells while preserving the integrity of non-tumor cells. OVs facilitate selective replication within tumor cells and deliver various eukaryotic transgene payloads, triggering immunogenic cell death (ICD), releasing TAAs, and initiating anti-tumor immune responses, all of which exhibit a tolerable safety profile that largely minimizes overlap with existing cancer therapies [352]. Upon lysis of tumor cells, viral progeny is released along with signals from TAAs, PAMPs, and damage-associated molecular patterns (DAMPs), which are coupled with tumor ICD (Fig. 4). PAMPs and DAMPs activate the innate immune response by binding to receptors such as TLRs. Moreover, mature DCs and NK cells are stimulated, thereby increasing the capacity of OVs to mediate tumor clearance [353]. Various pro-inflammatory cytokines incorporated as accompanying transgenes in OVs have been extensively studied (Fig. 4). For instance, a novel replication-competent recombinant oncolytic herpes simplex virus type 1 (VG161) enhances tumor infiltration of T cells and NK cells by encoding IL-12 and IL-15. Its efficacy has been successfully demonstrated in conjunction with ICIs through the modulation of tumor immunity and restructuring of TME metabolism, resulting in lasting anti-tumor effects [354, 355]. Sustained expression of IL-23 and the oncolytic effects of the virus increase the levels of Th1 chemokines and anti-tumor factors, including IFN-γ, TNF-α, perforin, IL-2, granzyme B, and activated T cells, thus creating a TME more conducive to anti-tumor immunity [356].

Additionally, combined strategies that integrate OVs with ICIs can enhance therapeutic efficacy (Fig. 4), For example, ICIs often demonstrate limited effectiveness in tumors characterized by low immune cell infiltration, known as the “immune desert” phenotype. In contrast, OVs can promote the recruitment of immune cells to the TME [357] and enhance the efficacy of ICIs, which depends on local PD-L1 expression. Variability exists between tumors and hosts. Many OVs induce PD-1 and PD-L1 expression, thereby establishing more sensitive targets for anti-PD-L1 immunotherapy. Therefore, OVs may partially reverse certain aspects of ICI resistance, with potential synergistic effects of OVs and ICIs corroborated in animal models [358, 359]. Preclinical investigations have evaluated combinatorial therapies that combine OVs and CAR-T cells and demonstrated improved anti-tumor activity (Fig. 4) [360]. In a glioblastoma mouse model, the use of an IL-7-encoding oncolytic adenovirus significantly enhanced the proliferation and persistence of tumor-infiltrating B7 H3-CAR-T cells, and in vitro studies further validated its synergistic anti-cancer effects [361]. Similarly, in a TNBC mouse xenograft model, the combination of mesothelin-targeted CAR -T cells and oncolytic adenoviral therapy was investigated, promoting the expression of a soluble transforming growth factor β receptor II fused to a human IgG Fc fragment [362]. Currently, four OVs and one non-oncolytic virus have received global approval for cancer treatment. Talimogene laherparepvec (T-VEC), an engineered oncolytic HSV1, is indicated for the treatment of patients with unresectable melanoma following initial surgery, and was the only OV approved by the FDA in 2015 [363]. Furthermore, in December 2022, the FDA approved a non-oncolytic adenovirus encoding IFNα-2b for the treatment of Bacillus Calmette-Guerin (BCG)-unresponsive non-muscle invasive bladder cancer (NMIBC) [364]. BCG has been the standard treatment for NMIBC for decades, which can exert anti-tumor effects by activating the local immune system. It is still the only anti-bacterial treatment for cancer approved by the FDA. The latest CREST global phase 3 trial brings a breakthrough: the new PD-1 antibody sasanlimab combined with traditional BCG offers excellent efficacy in high-risk patients with NMIBC, providing a new option for treatment [365].

In addition to these combinatorial strategies, OVs can be safely integrated with other cancer treatment modalities to mediate tumor regression via alternative mechanisms. Such strategies include cytotoxic chemotherapy, targeted molecular therapy, adoptive T-cell therapy, and radiotherapy [357]. Although OVs generally exhibit tolerable safety profiles, their replication potential necessitates specific logistic and biosafety considerations. Therefore, it is crucial to evaluate dosing, volume, timing, and administration routes in clinical contexts; this data is expected to facilitate the handling, management, and administration of OVs [366].

Antibiotics and bacteriophages

The TM is a promising target for activating anti-tumor immune responses and has significant implications for the treatment of solid tumors. Recent studies have shown that the use of antibiotics to eradicate bacteria within tumors can alter the immune tolerance of the TME in PDAC, enhance immune surveillance, and augment the efficacy of ICB therapies [367]. Specifically, oral or aerosolized antibiotics have been found to improve the composition of T cells and macrophage infiltration while simultaneously inhibiting the proliferation of MDSCs, thereby reducing mammary tumor growth in mouse models. Additionally, the combination of paclitaxel and ampicillin has been shown to enhance chemotherapeutic efficacy against BC [368]. In a separate study involving a mouse model of BC, antibiotics counteracted the metastatic effects of F. nucleatum [7]. However, the potential systemic effects of oral antibiotics must be carefully considered, including changes in microbial composition and metabolism that affect the host intestinal barriers and modulation of immune function, often associated with reduced responses to immunotherapy [369]. Consequently, the development of more targeted strategies for bacterial elimination within tumors is of utmost importance (Fig. 4). To address this need, drug-loaded nanoparticles have been proposed to specifically eradicate bacteria that colonize tumor cells. Specifically, a liposome-encapsulated antibiotic, the silver-tinidazole complex (LipoAgTNZ), targets bacteria within primary CRC tumors and liver metastases. This approach stimulates the production of novel antigens by anti-tumor CD8 + T cells with both heterologous and homologous bacterial epitopes that enhance immunogenicity and stimulate T cell recognition in both infected and uninfected tumor cells. The results indicated that the long-term survival rate of CRC models infected with F. nucleatum surpassed 70% [370]. Furthermore, the nanosizing of antibiotics and the construction of metronidazole-fluorouracil dual-targeting nanoparticles that accumulate in the TME and release these compounds represent a dual-strike strategy against both bacteria and cancer cells, exemplifying the effectiveness of this “best of both worlds” approach [371]. Nonetheless, antibiotics can also cause adverse effects. For example, patients with lung cancer undergoing PD-1/PD-L1 therapy who received antibiotics exhibited a significantly increased risk of immune-related adverse events, likely due to the diminished microbial diversity resulting from antibiotic administration [372]. Similarly, the application of broad-spectrum antibiotics prior to CD19-targeted CAR-T cell therapy can lead to an increased tumor burden and complicated systemic inflammatory responses [373]. In a retrospective study of 291 patients with advanced cancer treated with ICI, peritreatment antibiotic use significantly affected treatment response. The median PFS and OS of patients who did not receive antibiotics were 6.3 months and 21.7 months, respectively. In contrast, OS was 17.7 months in patients with a single use of antibiotics. The OS and PFS of those who used multiple antibiotics or those who used antibiotics in the long term were 6.3 months and 2.8 months, respectively. Multivariate analysis showed that antibiotic use was an independent predictor of worse PFS and OS, suggesting that antibiotics should be used with caution during treatment with ICIs [374].

Bacteriophage (phage) therapy is emerging as a promising alternative to antibiotic therapy. Bacteriophages are among the most abundant biological entities and constitute a diverse group of viruses that infect bacteria. Owing to their specificity for bacterial infections, phages have attracted increasing attention as alternatives to traditional antibacterial treatments, particularly in cases of antibiotic-resistant infections [375]. Recent advances have involved the modification of phages through synthetic biology to create programmable bacterial killers that effectively target oncogenic bacteria with increased abundance in tumor-associated microbiomes (Fig. 4) [376]. For instance, phages specific to F. nucleatum have been engineered to carry irinotecan nanoparticles, facilitating covalent delivery to murine colon tumors and significantly enhancing first-line chemotherapy outcomes in CRC [377]. However, systemic phage therapy presents challenges such as the potential development of bacterial resistance from inappropriate or prolonged use, immune responses against phages, and rapid clearance of phages by complement activation and macrophages in the spleen and liver [11, 378]. Oral phage therapy offers a precise method for targeting pathogenic bacteria while avoiding systemic immune responses, thereby supporting the overall stability of the microbiome. Studies have indicated that high-density viral-like particles present in FMT, comprising up to 90% of the phages, can effectively modulate dysbiosis associated with small intestinal bacterial overgrowth and eliminate pathogenic bacteria in mouse models [379]. Employing phage cocktails that target bacterial receptors can enhance specificity against pathogenic bacteria. For example, researchers have developed an effective mixture of Staphylococcus aureus phages and antibiotics that demonstrated substantial efficacy in vivo in wound infection models [380]. Future research should focus on identifying bacterial species associated with carcinogenic risks and characterizing anti-cancer strains, as well as on thoroughly elucidating the tripartite interactions among phages, bacteria, and tumor cells. Breakthroughs in addressing these pivotal scientific questions will provide a fundamental basis for developing innovative cancer treatment strategies based on the dynamic properties of phages.

Conclusion and future perspectives

The rapid advancement in our understanding of the role of TM in tumor progression has ushered in the development of promising cancer treatment strategies. A growing body of research has illustrated the relationship between tumor-associated microbiota and cancer. However, from a single-cell perspective, high-resolution techniques must be developed to delineate the cellular components in the tumor microenvironment, thereby enhancing our understanding of the intrinsic microbiome associated with tumors. NGS and advanced computational tools, combined with machine learning methods, can facilitate in-depth studies on the interactions between tumors and the microbiome in individual neoplasms. This integrated approach can substantially advance cancer treatment. Furthermore, we highlighted the considerable heterogeneity in the origin and composition of TM across different types of cancer, emphasizing that TM can activate several signaling pathways in tumor cells. It is necessary to clarify whether specific microbial species can affect distinct subsets of signaling pathways. In terms of mechanism, we delineated the effects of genetic changes, epigenetic modifications, metabolic regulation, invasion, metastasis, and chronic inflammatory responses on tumor progression. In recent years, significant progress has been made in studying the relationship between TM and the efficacy of chemotherapy, radiotherapy, and immunotherapy. However, the mechanism through which TM affects anti-tumor immunity and therapeutic effect still remains unclear, which seriously restricts the clinical application of microbial-based therapies. In the future, animal models are needed to investigate the relationship between microorganisms and tumors across different disciplines. In clinical settings, assessment of microbial characteristics and treatment response can help find new targets, explore the interaction between TM and inflammation, metabolism, etc., and establish new strategies for cancer prevention and treatment. These studies can provide a basis for the application of microbial therapy and promote innovation in cancer treatment.

Abbreviations

5-FU

5-fluorouracil

ACT

Adoptive cell transfer

AhR

Aryl hydrocarbon receptor

BC

Breast cancer

BFT

Bacteroides fragilis toxin

BGNP

Branched gold nanoparticle

CAFs

Cancer-associated fibroblasts

CagA

Cytotoxin-associated gene A

CAR

Chimeric antigen receptor

CDN

Cyclic dinucleotides

CRC

Colorectal cancer

CSCs

Cancer stem cells

CTLA4

Cytotoxic T-lymphocyte-associated protein 4

DAMPs

Damage-associated molecular patterns

DCs

Dendritic cells

DSBs

Double-strand breaks

EMT

Epithelial-mesenchymal transition

ESCC

Esophageal squamous cell carcinoma

FISH

Fluorescence in situ hybridization

FMT

Fecal microbiota transplantation

GC

Gastric cancer

GM-CSF

Granulocyte-macrophage colony-stimulating factor

HCC

Hepatocellular carcinoma

HIF-1

Hypoxia-inducible factor 1

ICB

Immune checkpoint blocker

ICD

Immunogenic cell death

ICI

Immune checkpoint inhibitor

ICT

Immune checkpoint blockade therapy

IFN-I

Type I interferon

IHC

Immunohistochemistry

IL-6

Interlukin-6

ITME

Immunosuppressive tumor microenvironment

LCA

Lithocholic acid

LSCC

Lung squamous cell carcinoma

LUAD

Lung adenocarcinoma

MALT

Mucosa-associated lymphoid tissue

MAMP

Microbial-associated molecular pattern

MBL

Mannan-binding lectin

MDSC

Myeloid-derived inhibitory cells

METTL3

Methyltransferase-like protein 3

MicroSPLiT

Microbial split-pool ligation transcriptomics

MSI

Microsatellite instability

MyD88

Myeloid differentiation primary-response protein 88

NAT

Normal adjacent tissue

NGS

Next generation sequencing

NK

Natural killer

NMIBC

Non-muscle invasive bladder cancer

OMVs

Outer membrane vesicles

OS

Overall survival

OSCC

Oral squamous cell carcinoma

Ovs

Oncolytic viruses

PAMP

Pathogen-associated molecular pattern

PCa

Prostate cancer

PCR

Polymerase chain reaction

PD-1

Programmed death 1

PDAC

Pancreatic ductal adenocarcinoma

PD-L1

Programmed death ligand 1

PFS

Progression-free survival

PitNET

Pituitary neuroendocrine tumor

PI3K

Phosphatidylinositol 3-kinase

PMN

Premetastatic niche

PMN-MDSCs

Polymorphonuclear myeloid-derived suppressor cells

PRR

Pattern recognition receptor

ROS

Reactive oxygen species

SCFAs

Short-chain fatty acids

SCS

Single cell sequencing

SMO

Spermine oxidase

STING

Stimulator of interferon genes

TAAs

Tumor-associated antigens

Th17

T helper cells 17

TIL

Tumor-infiltrating lymphocyte

TLR

Toll like receptor

TM

Tumor microbiome

TME

Tumor microenvironment

TNBC

Triple-negative breast cancer

TNF-α

Tumor necrosis factor-α

WMT

Washed microbiota transplantation

Author contributions

WTZ preliminarily constructed the conceptual framework and fnished the manuscript, YHX, HR, YLL, QW and MDR revised the manuscript, WTZ, LT, XHZ, CQ and YFL performed the literature search, MSY advised and supervised the work. All authors have read and approved the fnal manuscript.

Funding

This study was supported by the National Natural Science Foundation of China (No.82472867);The Outstanding Young Scholar Project of the Natural Science Foundation of Heilongjiang Province (YQ2024H003);Postgraduate Research & Practice Innovation Program of Harbin Medical University (YJSCX2024-78HYD).

Data availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All of the authors read and approved the fnal manuscript for publication.

Competing interests

The authors declare no competing interests.