Genetically Predicted Gut Microbiota in Correlation with the Risk of Head and Neck Cancer

Abstract

Background

Existing evidence supports observational associations between the gut microbiome composition and susceptibility to extraintestinal cancers; however, the causal relationship between gut microbiome composition and the risk of developing head and neck cancer (HNC) remains uncertain. Therefore, we conducted a two-sample Mendelian randomization (MR) analysis using publicly available genome-wide association study (GWAS) data to investigate the potential causal link between gut microbiome composition and HNC.

Methods and analysis

Relevant investigations were conducted to examine the casualties of exposures and outcomes, using inverse variance weighted (IVW), weighted median (WM) and the MR-Egger method. Sensitivity analyses, including the MR-Egger intercept test, Cochran’s Q test, and leave-one-out analysis, were used to assess pleiotropy and heterogeneity. In addition, the Steiger directionality test was applied to exclude studies with potential directional biases.

Results

Our study identified 16 causal relationships between the genetically predicted abundances of gut microbiota and the risk of developing HNC. After sensitivity analysis on these identified relationships, 10 [class (Actinobacteria, Betaproteobacteria), order Bacillales, family (Ruminococcaceae), genus (Eubacterium_ventriosum_group, Coprococcus2, Coprococcus3, Desulfovibrio, Howardella, Subdoligranulum, Veillonella, Victivallis)] out of the initially identified 16 causal relationships were ultimately validated as being associated with HNC tumors. Steiger directionality test indicated that the MR estimate of causal direction was accurate (All p < 0.05).

Conclusion

Our MR analysis revealed that the gut microbiome composition was causally associated with the risk of developing HNC and could serve as a new biomarker for preventive and therapeutic strategies for HNC.

Keywords

head and neck cancer gut microbiota causal relationship biomarker

1. Introduction

Head and neck cancer (HNC) is the sixth most common type of cancer, with approximately 930,000 new cases and more than 460,000 deaths reported worldwide each year, of which oral and oropharyngeal cancers are the most common subtypes.^1,2 Human papillomavirus (HPV) status, Epstein-Barr virus (EBV) status, tobacco use status, alcohol consumption status, and poor oral hygiene are major risk factors for the development of HNC.^3-7

To date, increasing evidence suggests a potential association between gut microbiota characteristics and cancer risk.^8,9 Previous studies based on animal models have reported that the gut microbiome is involved in tumor development through various signaling pathways.^10,11 One study revealed differences in overall microbiome characteristics between healthy people and HNC patients.¹² The results from a randomized, double-blind, placebo-controlled study revealed that probiotics reduced the severity of oral mucositis induced by chemoradiotherapy in nasopharyngeal carcinoma patients by regulating the gut microbiome.¹³ Recently, Hes et al investigated the link between the gut microbiome composition and the side effects of chemoradiation in head and neck squamous cell carcinoma patients and reported that patients with severe mucositis had reduced survival and a reduced abundance of bacteria that are specifically associated with mucosal inflammation.¹⁴

It is important to note that existing evidence from observational studies alone is insufficient to draw firm conclusions about a potential causal relationship between the gut microbiome composition and HNC risk. Although randomized controlled trials (RCTs) are the gold standard for determining causality, the long latency from exposure to certain microbiota to carcinogenesis makes it impractical in clinical settings.¹⁵ In this context, a new approach is necessary to investigate the causal relationship between the gut microbiome and cancer risk.

Mendelian randomization (MR) is a method that uses instrumental variables to infer causal relationships between exposures and outcomes and is widely used to explore potential causal relationships between environmental exposures and diseases.^16-18 With two-sample MR analysis, researchers can use single-nucleotide polymorphisms (SNPs) from independent genome-wide association studies (GWAS) that show strong associations with specific outcomes to infer causal relationships between exposure factors and outcomes.

Over the past few years, Mendelian randomization (MR) studies have substantially advanced our understanding of the causal role of gut microbiota in various cancers. A recent scoping review systematically summarized 12 MR studies linking gut microbial taxa to 14 cancer types, including colorectal, lung, breast, prostate, and gastric cancers, among others.¹⁹ Nevertheless, the evidence for HNC remains scarce and fragmented. To date, only a few MR investigations have focused on oral cavity or head and neck malignancies.^20-23

Despite these important contributions, several key gaps remain. First, the number of MR studies specifically targeting HNC is extremely limited, and the microbial taxa examined vary considerably across studies, leading to inconsistent findings. Second, most existing studies have relatively modest sample sizes and do not systematically evaluate a broad range of gut microbial taxa using a harmonized analytical framework. Third, no previous study has comprehensively integrated multiple MR methods (e.g., inverse variance weighted, MR-Egger, weighted median) with extensive sensitivity analyses to assess the causal effects of a wide panel of gut microbiota on HNC risk in a single, well-powered design.

Our study was designed to fill these gaps. Here, we systematically investigate the genetically predicted association between gut microbial taxa and the risk of HNC using a two-sample MR approach. We leverage the largest available genome-wide association study (GWAS) summary data for gut microbiota (MiBioGen consortium) and for HNC (from the UK Biobank and one large European-ancestry GWAS). By applying a rigorous analytical pipeline that includes multiple MR estimators, comprehensive sensitivity analyses, and correction for multiple testing, we provide robust evidence on the causal role of specific gut bacteria in HNC development. To the best of our knowledge, this is the first study to systematically evaluate such a wide spectrum of gut microbial taxa in relation to HNC using a unified MR framework with stringent control for pleiotropy and bias.

A growing body of evidence supports the “gut–head and neck axis” hypothesis, which posits that gut microbiota and their metabolites can enter the systemic circulation and remotely influence the immune microenvironment of the upper aerodigestive tract, including the head and neck mucosa.^24,25 This long-distance crosstalk is primarily mediated by microbial metabolites such as short-chain fatty acids (e.g., butyrate) and by the modulation of circulating immune cells,²⁶ a paradigm well established in other gut-distal cancers and increasingly recognized in head and neck oncology.

In this study, we conducted a two-sample MR analysis using GWAS data. The final objective of this MR analysis was to elucidate the direction and magnitude of the causal relationship between the gut microbiome composition and the risk of developing HNC.

2. Methods

2.1. Study Design

Utilizing a two-sample Mendelian randomization framework, we assessed the causal impact of gut microbiota on head and neck cancer (HNC). To comprehensively investigate the contribution of gut microbiota to cancer incidence, we conducted Mendelian randomization analyses at five distinct taxonomic levels, namely, phylum, class, order, family, and genus. The study design accompanied by the fundamental MR assumptions is presented in Figure 1. We followed the suggestions provided in the STROBE-MR guidelines to guarantee a thorough and open presentation of our MR analysis.

Figure 1.

Three core assumptions of mendelian randomization and the analytical workflow of the study. The diagram illustrates the instrumental variable (IV) assumptions: (1) genetic variants are strongly associated with the exposure (gut microbiota); (2) genetic variants are independent of potential confounders; and (3) genetic variants influence the outcome (head and neck cancer) only through the exposure. The flowchart also presents the key data sources (GWAS datasets for gut microbiota and head and neck cancer), the selection of genetic instruments, and the causal pathway from exposure to outcome under the MR framework

2.2. Data Sources

The genetic information of gut microbiota for the exposure variables was obtained from the largest GWAS conducted by the MiBioGen consortium (https://mibiogen.gcc.rug.nl/menu/main/home), comprising 18340 participants from 24 cohorts.²⁷ In total, 211 taxa were included in the GWAS (9 phyla, 16 classes, 20 orders, 35 families, and 131 genera). The GWAS summary data for head and neck cancer were obtained from the UK Biobank database. HNC cases in the UK Biobank database were classified in the GWAS catalog (GWAS ID: ieu-b-4912; https://opengwas.io/datasets/ieu-b-4912). The study included 373,122 participants, including 1106 HNC patients and 372,016 control participants. Detailed information on the GWAS data sources used in this study is shown in Supplemental Table S1.

2.3. Instrument Variable (IV) Selection

Considering the limited number of available SNPs, we employed a lenient p-value threshold of <1×10⁻⁵, which is commonly used when dealing with a restricted SNP pool. Subsequently, linkage disequilibrium (LD) clumping was performed to mitigate LD among single-nucleotide polymorphisms (SNPs) (r²< 0.001, distance=10,000 kb). Subsequently, we harmonized the exposure and outcome data to ensure consistency. In cases of palindromic SNPs, the forward stranded alleles were inferred based on allele frequency information.

2.4. MR Analysis

The random-effects inverse-variance weighting (IVW), weighted median (WM), and MR-Egger methods were employed for the Mendelian randomization analysis to assess the robustness of the MR results.²⁸ We performed a stratified Bonferroni correction according to the number of microbial features at each taxonomic level: phylum (n = 4, threshold = 0.05/4), class (n = 16, threshold = 0.05/16), order (n = 1, threshold = 0.05), family (n = 29, threshold = 0.05/29), and genus (n = 104, threshold = 0.05/104). A P value below the corresponding level specific threshold was considered statistically significant; a P value < 0.05 but above the level specific threshold was considered suggestive. To assess the robustness of our findings, we performed sensitivity analyses employing a range of statistical methods to examine potential violations of the MR assumptions. Heterogeneity was examined using the IVW method. The Cochran Q test was used to evaluate heterogeneity, and P values greater than 0.05 were considered to indicate a lack of heterogeneity.²⁹ The intercept term derived from the MR-Egger regression was used to detect horizontal pleiotropy. Leave-one-out (LOO) analysis was implemented to determine whether the results were affected by a single SNP. Moreover, we conducted the MR Steiger directionality test to validate the alignment of our findings with our hypothesis.

The analyses were conducted using the R program (version 4.2.0) with the “TwoSampleMR” package (version 0.5.8)³⁰ and the “Mendelian Randomization” package (version 0.9.0).³¹

3. Results

After rigorous instrument selection steps, at a significance level of p < 1×10⁻⁵, a total of 2001 SNPs were incorporated into the MR analysis. A total of 162 bacterial taxa were grouped into five taxonomic levels (phylum, class, order, family, and genus). Eight bacterial taxa were unknown and were thus excluded from our study, leaving 154 bacterial taxa (4 phyla, 16 classes, 1 order, 29 families, and 104 genera) were included in the MR analysis (Supplemental Table S2). The number of SNPs associated with each of the bacterial taxa ranged from 2 to 19.

3.1. Causal Association of Gut Bacteria With the Risk of Developing HNC

In the MR analysis with IVW, WM, and MR-Egger regression methods, a total of 16 causative relationship features were identified at P < 0.05 (Table 1), including class (Actinobacteria, Betaproteobacteria, Lentisphaeria), order Bacillales, family (Ruminococcaceae), genus (Eubacterium_ventriosum_group, Coprococcus2, Coprococcus3, Desulfovibrio, Erysipelatoclostridium, Howardella, Lachnospiraceae_FCS020_group, Paraprevotella, Subdoligranulum, Veillonella, Victivallis). Following Bonferroni correction, only the association with Actinobacteria remained significant; all other associations were no longer significant. Actinobacteria are rich sources of natural secondary metabolites with potent anticancer and cytotoxic properties.³² Moreover, Desulfovibrio plays an important role in cancer development.³³ Our research findings indicated that Actinobacteria have a beneficial effect on the development of HNC, whereas Desulfovibrio has a detrimental impact on the development of HNC. Scatter plots depicting microbiota constituents are presented in Supplemental Figure S1-16. In total, 4 microbiotas were associated with a decreased risk of developing HSC, and 12 microbiotas were associated with an increased risk of developing HSC.

Table 1.

Significant MR Results

Gut microbiota	SNPs (n)	IVW		WM		MR egger
Gut microbiota	SNPs (n)	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
class.Actinobacteria	15	0.998 (0.997-0.999)	0.003	0.998 (0.996-1.000)	0.012	0.997 (0.994-1.001)	0.131
class.Betaproteobacteria	12	1.000 (1.000-1.003)	0.018	1.002 (1.000-1.004)	0.076	1.000 (0.996-1.005)	0.850
class.Lentisphaeria	8	1.000 (0.998-1.001)	0.788	0.999 (0.998-1.000)	0.121	0.995 (0.992-0.999)	0.039
family.Ruminococcaceae	9	1.000 (0.998-1.001)	0.618	0.999 (0.997-1.001)	0.252	0.994 (0.991-0.998)	0.026
genus.Eubacterium_ventriosum_group	12	1.001 (1.000-1.003)	0.029	1.002 (1.000-1.003)	0.083	1.003 (0.997-1.008)	0.372
genus.Coprococcus2	8	0.999 (0.997-1.000)	0.115	0.998 (0.996-1.000)	0.022	1.004 (0.991-1.017)	0.608
genus.Coprococcus3	7	0.998 (0.997-1.000)	0.044	0.999 (0.996-1.001)	0.246	0.995 (0.989-1.001)	0.168
genus.Desulfovibrio	10	1.001 (1.000-1.002)	0.059	1.002 (1.000-1.003)	0.049	1.003 (0.999-1.006)	0.169
genus.Erysipelatoclostridium	14	1.001 (1.000-1.002)	0.298	1.001 (0.999-1.002)	0.380	1.005 (1.001-1.009)	0.024
genus.Howardella	9	1.001 (1.000-1.001)	0.141	1.001 (1.000-1.002)	0.037	1.003 (0.999-1.006)	0.164
genus.Lachnospiraceae_FCS020_group	12	0.999 (0.997-1.000)	0.169	0.999 (0.997-1.001)	0.197	0.995 (0.992-0.998)	0.015
genus.Paraprevotella	11	1.001 (0.999-1.002)	0.442	1.001 (1.000-1.003)	0.021	0.999 (0.995-1.004)	0.819
genus.Subdoligranulum	9	1.002 (1.001-1.004)	0.010	1.002 (1.000-1.005)	0.039	1.003 (0.998-1.009)	0.262
genus.Veillonella	7	1.001 (1.000-1.003)	0.033	1.002 (1.000-1.004)	0.037	1.003 (0.996-1.009)	0.450
genus.Victivallis	6	1.001 (1.000-1.003)	0.027	1.001 (1.000-1.002)	0.111	1.004 (0.992-1.017)	0.523
order.Bacillales	9	1.000 (1.000-1.001)	0.299	1.000 (0.999-1.001)	0.455	1.005 (1.001-1.008)	0.034

3.2. Sensitivity Analysis and Steiger

First, heterogeneity analysis was conducted. The results of Cochran’s Q test revealed that three microbiota P values were lower than 0.05, which indicated their heterogeneity, including those for class. Lentisphaeria (P=0.012), genus. Lachnospiraceae_FCS020_group (P=0.048) and genus. Paraprevotella (P=0.009, Table 2). There was little evidence of heterogeneity in the other 13 gut microbiota, with P > 0.05. After the three bacterial taxa with heterogeneity were excluded, the horizontal pleiotropy between the remained 13 instrumental variables and HNC was assessed via MR-Egger regression. Little evidence of horizontal pleiotropy was found except for family. Ruminococcaceae (P=0.026), genus. Erysipelatoclostridium (P=0.036) and order. Bacillales (P = 0.046, Table 3). After removing these unqualified relationships with heterogeneity and pleiotropy analysis, the leave-one-out analysis demonstrated that the remaining 10 causal relationships between gut microbiota and the risk of developing HNC remained robust (Figure 2).

Table 2.

Heterogeneity Analysis Using Cochran’s Q-Test in the IVW Method

Exposure	SNPs (n)	Method	Q	Q_df	Q_pval
class.Actinobacteria	15	IVW	9.28	14	0.812
class.Betaproteobacteria	12	IVW	6.38	11	0.847
class.Lentisphaeria	8	IVW	17.96	7	0.012
family.Ruminococcaceae	9	IVW	9.79	8	0.280
genus.Eubacterium_ventriosum_group	12	IVW	7.30	11	0.774
genus.Coprococcus2	8	IVW	8.92	7	0.259
genus.Coprococcus3	7	IVW	3.74	6	0.712
genus.Desulfovibrio	10	IVW	10.85	9	0.286
genus.Erysipelatoclostridium	14	IVW	14.65	13	0.329
genus.Howardella	9	IVW	6.14	8	0.632
genus.Lachnospiraceae_FCS020_group	12	IVW	19.82	11	0.048
genus.Paraprevotella	11	IVW	23.42	10	0.009
genus.Subdoligranulum	9	IVW	8.54	8	0.382
genus.Veillonella	7	IVW	6.15	6	0.406
genus.Victivallis	6	IVW	9.47	5	0.092
order.Bacillales	9	IVW	10.64	8	0.223

Table 3.

Pleiotropy Analysis Using MR-Egger Method

Exposure	SNPs (n)	egger_intercept	se	pval
class.Actinobacteria	15	7.01E-05	1.16E-04	0.554
class.Betaproteobacteria	12	8.26E-05	1.64E-04	0.626
family.Ruminococcaceae	9	4.43E-04	1.57E-04	0.026
genus.Eubacterium_ventriosum_group	12	-8.58E-05	2.05E-04	0.684
genus.Coprococcus2	8	-3.57E-04	4.82E-04	0.487
genus.Coprococcus3	7	2.06E-04	1.92E-04	0.332
genus.Desulfovibrio	10	-1.70E-04	1.83E-04	0.379
genus.Erysipelatoclostridium	14	-3.60E-04	1.52E-04	0.036
genus.Howardella	9	-3.15E-04	2.53E-04	0.253
genus.Subdoligranulum	9	-7.37E-05	1.77E-04	0.690
genus.Veillonella	7	-1.06E-04	2.72E-04	0.713
genus.Victivallis	6	-3.71E-04	7.78E-04	0.658
order.Bacillales	9	-6.14E-04	2.53E-04	0.046

Figure 2.

Mendelian randomization analysis of HNC and gut microbiota. Forest plot showing effect estimates and 95% confidence intervals for gut microbiota on HNC risk. Analysis employed three Mendelian randomization methods: inverse variance weighted (IVW), weighted median (WM), and MR Egger regression approaches. The 10 most significant microbial taxa were demonstrated to show potential protective or risk effects

Additionally, the MR Steiger directionality test was employed to examine the causal relationship between gut microbiota and the risk of developing HNC. Supplemental Table S3 shows the results. The variance in the outcome (snp_r2.outcome) was less than that for each exposure (snp_r2.exposure), confirming the causal direction.

4. Discussion

In patients with malignant tumors, abnormal gut microbiota abundance and dysbiosis are associated with tumor occurrence and development. For example, studies have shown an increase in Bacteroidetes in epithelial ovarian cancer patients compared to healthy control participants.³⁴ The pretreatment abundances of Faecalibacterium, Prevotella, and Phascolarctobacterium in the gut microbiome of HNC patients are negatively correlated with the risk of tumor recurrence.³⁵ In this study, MR and sensitivity analysis finally revealed 10 strong causal relationships between gut microbiota and the incidence of HNC. The Steiger directionality test further verified the causal relationship between gut microbiota and the risk of developing HNC. We also performed reverse MR to explore the causal relationships between HNC status and the gut microbiome. However, we obtained only 20 SNPs as instrumental variables (Supplemental Table S4). Furthermore, there were no available SNPs when harmonizing the exposure and outcome data.

Understanding the causal association between gut microbiota and oral and oropharyngeal cancers can provide valuable insights for making informed decisions regarding health management and disease prevention strategies. In our study, among the 10 gut microbiome species identified as causally linked to head and neck cancer, several of these microbial species have already been extensively investigated for their significant association with tumor development. In our study, Actinobacteria had a detrimental impact on the development of HNC (OR=0.998 [P=0.003, 95% CI: 0.619-0.991, IVW model]). Actinobacteria are widely distributed in nature and are the most economically and biologically valuable microorganisms. The secondary metabolites of actinobacteria possess extensive antibacterial and antitumor activities, making them valuable resources for the development of new antibacterial and antitumor drugs.^32,36 In our study, the presence of Desulfovibrio had a deleterious effect on HNC (OR = 1.002 [P=0.049, 95% CI: 1.000-1.003, WM method]).

Actinobacteria are highly prominent in the gut system, supplementing the enzymes involved in the degradation and biotransformation of substances introduced to the diet which are lack in human intestine. Actinobacteria in the gut system was reported to participate in the degradation of resistant starch³⁷ and play a role in the transformation of Linoleic acid (LA) into conjugated LA, which boosts the immune function.³⁸ B. psuedocatenulatum, a species of Actinobacteria, was reported to display the function in down-regulating the inflammatory cytokines and chemokines,³⁹ and improving the inflammatory response by boosting tumor-killing factors such as LPS (Lipopolysaccharide) and TNF (Tumor Necrosis Factor).⁴⁰ Actinobacteria can also modulate autoimmune and immune-inflammatory response by inducing regulatory T-cells.⁴¹ In this study, we identified the causal relationship between Actinobacteria and the HNC. The abovementioned evidence suggests that the potential mechanism might be the immune modulation function of the bacteria, though a solid conclusion should be drawn by further investigation.

Coprococcus are notable for the ability to ferment dietary fibers into short-chain fatty acids (SCFA), especially butyrate.⁴² Butyrate has been considered of significant importance, as involved in multiple important physiological functions, such as trans-epithelial transport, amelioration of mucosal inflammation, alleviation of oxidative stress, enforcement of the epithelial barrier, and protection against colorectal cancer (CRC).⁴³ It is reported that the ingestion of sodium butyrate could contribute to an increased expression of TGF-β, which possess anti-tumor function.⁴⁴ Microbial butyrate could differentially inhibit glucose transportation, glycolysis, and DNA synthesis via suppressing GLUT1 and G6PD through the GPR109a-AKT pathway.⁴⁵ As cancer cells are usually in demand of higher glucose to supply an accelerated metabolic rate for cell growth, Coprococcus could serve as a potential target of an anticancer drug due to its function as a glycolysis inhibitor.

A cross-feeding relationship is also reported between Actinobacteria and butyrogenic genera. The lactate and acetate produced by Bifidobacterium, a genes of Actinobacteria, could be further utilized by butyrogenic microbes to generate butyrate, which in turn improves the abundance of Bifidobacterium,⁴⁶ indicating a potential synergistic effect between Actinobacteria and Coprococcus in cancer treatment. It is promising to anticipate that the oral administration of prebiotics and probiotics, such as Actinobacteria and Coprococcus, might serve as new preventive and therapeutic strategies to treat HNC.

Desulfovibrio is a type of sulfate-reducing bacterium (SRB) that is common in the human gut. However, the species and abundance of Desulfovibrio in the gut can change when the host suffers from various diseases. Current research on diseases such as diabetes, chronic kidney disease (CKD), Alzheimer’s disease (AD), and Parkinson’s disease has revealed varying degrees of increased abundance of Desulfovibrio in patients’ gut microbiomes.^47-51 Currently, studies have shown significant enrichment of Desulfovibrio in the intestines and tumor sites of patients with CRC. A high-fat diet increases the abundance of Desulfovibrio and enhances the potential for liver metastasis in CRC patients.⁵² Other studies have shown that D. desulfuricans reduces the effectiveness of FOLFOX chemotherapy and promotes the progression of colorectal cancer.⁵³ Desulfovibrio may play a crucial role in the development, progression, and treatment of colorectal cancer.^54,55

In this study, we evaluated the causal relationship between gut microbiota and HNC incidence using two-sample MR analysis, confirming the association between gut microbiota and the risk of developing HNC. MR analysis can infer causal associations from exposure to outcomes that are unaffected by confounding factors. In this study, large-scale population samples were used to improve the reliability of the results by using GWAS data. In addition, we performed sensitivity analyses using multiple methods to ensure the credibility of the results.

However, our study has several limitations. First, the microbiota GWAS data were obtained from people of multiple ethnicities but primarily individuals of European descent, and the GWAS data for HNC were all from European populations. This may limit the generalizability of our findings to other ethnicities. Second, we used a relatively lenient p-value threshold (p < 1×10^-5) for instrumental variable selection to retain a sufficient number of SNPs, which may introduce weak instrument bias. Nevertheless, we calculated F-statistics for all instruments, and all exceeded the conventional threshold of 10, indicating that weak bias is unlikely to be substantial. Third, although we applied FDR correction for multiple testing, the lack of significant associations after correction (i.e., no taxa passed FDR < 0.05) suggests that our findings should be considered preliminary and hypothesis-generating. Fourth, we found that no evidence of horizontal pleiotropic of exposure factors was detected when using the MR-Egger regression detection. The results of Cochran’s IVW Q test showed no significant heterogeneity of IVs. Similarly, the leave-one-out sensitivity analyses indicated that no single SNP significantly affected the causal association. Fifth, MR estimates reflect lifelong genetic predisposition rather than transient environmental exposures, which is a strength for causal inference but also limits direct translation to clinical interventions. Sixth, we did not conduct subtype-specific analyses of HNC (e.g., by anatomical site or HPV status), as such stratified GWAS data are not yet available. Finally, the mechanistic interpretations we provided regarding Actinobacteria and Coprococcus are speculative and based on existing literature from other cancer types or model systems. Direct experimental evidence linking these gut microbes to HNC pathogenesis is lacking, and our findings should be validated by functional studies, animal models, and prospective cohorts before any clinical application.

5. Conclusion

In summary, this two-sample Mendelian randomization study provides genetic evidence that specific gut microbiota are causally associated with the risk of head and neck cancer. Among the 211 taxa examined, Actinobacteria and Coprococcus emerged as the most robustly associated, with the former potentially exerting protective effects through immune modulation (e.g., regulatory T-cell induction, downregulation of inflammatory cytokines) and the latter via production of butyrate, a short-chain fatty acid known to inhibit glycolysis and reinforce epithelial barrier function. These findings lend support to the emerging “gut–head and neck axis” hypothesis, whereby gut-derived metabolites and immune cells remotely influence the upper aerodigestive tract microenvironment. By combining Mendelian randomization—which mitigates confounding and reverse causation—with biologically plausible pathways, our study bridges genetic causality and cancer biology at the human level. From a translational perspective, the identified microbial taxa may serve as non-invasive biomarkers for early risk stratification, particularly given the current lack of effective screening tools for HNC, and could inform future microbiome-based preventive strategies (e.g., probiotics, dietary interventions). However, these conclusions are hypothesis-generating, and further functional studies, animal models, and prospective cohorts are required to validate the causal mechanisms and clinical applicability before any intervention can be recommended.

Supplemental Material

Supplemental Material - Genetically Predicted Gut Microbiota in Correlation With the Risk of Head and Neck Cancer

Supplemental Material for Genetically Predicted Gut Microbiota in Correlation With the Risk of Head and Neck Cancer by Zhaoyi Lu, Liqing Zhang, Qin Zhang, Qianqian Duan, Xuewen Liu, Xiaoli Deng, Dongsheng Chen, and Xi Chen in Cancer Informatics.

Supplemental Material

Supplemental Material - Genetically Predicted Gut Microbiota in Correlation With the Risk of Head and Neck Cancer

Footnotes

ORCID iD

Xi Chen

Ethical Considerations

The data utilized in this study were openly accessible and authorized by their respective institutions. An ethical clearance for the present research is unnecessary.

Consent to Participate

No animal participants were involved in the investigation.

Author contributions

Zhaoyi Lu, Liqing Zhang, Xi Chen and Qin Zhang conceived and designed the research. Material preparation and data collection were performed by Qianqian Duan, Xuewen Liu, Dongsheng Chen and Xiaoli Deng. Data analysis, interpretation and visualization were performed by Qin Zhang. The draft manuscript was written by all author. Xi Chen and Dongsheng Chen revised and edited the manuscript. Xi Chen approved the version to be published. All authors confirm that they had full access to all the data in the study and accept the responsibility to submit for publication.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was financially supported by the Natural Science Foundation of Jiangsu Province (Grants No. BK20230739), Jiangsu Province Hospital (the First Affiliated Hospital with Nanjing Medical University) Clinical Capacity Enhancement Project (Grants No. JSPH-MA-2023-1), The Natural Science Foundation of the Jiangsu Higher Education Institution of China (Grants No. 1020210958) and Jiangsu Province Industry-University-Research Collaboration (Grants No. BY20230230).

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data Availability Statement

The GWAS data utilized in this study were openly accessible and their sources were adequately explained in the manuscript (https://mibiogen.gcc.rug.nl/menu/main/home). All code required to reproduce the analyses presented in this study has been deposited on GitHub and is freely accessible at .

Supplemental Material

Supplemental material for this article is available online.

Appendix

References

Sung

Ferlay

Siegel

, et al. Global Cancer Statistics 2020: GLOBOCAN Estimates of Incidence and Mortality Worldwide for 36 Cancers in 185 Countries. CA Cancer J Clin. 2021;71(3):209-249.

Zumsteg

Luu

Rosenberg

, et al. Global epidemiologic patterns of oropharyngeal cancer incidence trends. J Natl Cancer Inst. 2023;115(12):1544-1554.

Katirachi

Grønlund

Jakobsen

Grønhøj

von Buchwald

. The Prevalence of HPV in Oral Cavity Squamous Cell Carcinoma. Viruses. 2023;15(2):451.

Gormley

Creaney

Schache

Ingarfield

Conway

. Reviewing the epidemiology of head and neck cancer: definitions, trends and risk factors. Br Dent J. 2022;233(9):780-786.

Guo

Xiao

Chen

Zhao

Lin

Chen

. Epidemiological Trends of Head and Neck Cancer: A Population-Based Study. Biomed Res Int. 2021;2021:1738932.

Kaidar-Person

Gil

Billan

. Precision medicine in head and neck cancer. Drug Resist Updat. 2018;40:13-16.

Jethwa

Khariwala

. Tobacco-related carcinogenesis in head and neck cancer. Cancer Metastasis Rev. 2017;36(3):411-423.

Teng

NMY

Price

McKee

Hall

Robinson

. Exploring the impact of gut microbiota and diet on breast cancer risk and progression. Int J Cancer. 2021;149(3):494-504.

Cullin

Azevedo Antunes

Straussman

Stein-Thoeringer

Elinav

. Microbiome and cancer. Cancer Cell. 2021;39(10):1317-1341.

10.

Matsushita

Fujita

Hayashi

, et al. Gut Microbiota-Derived Short-Chain Fatty Acids Promote Prostate Cancer Growth via IGF1 Signaling. Cancer Res. 2021;81(15):4014-4026.

11.

Behary

Amorim

Jiang

, et al. Gut microbiota impact on the peripheral immune response in non-alcoholic fatty liver disease related hepatocellular carcinoma. Nat Commun. 2021;12(1):187.

12.

Dorobisz

Zatoński

. The Microbiome's Influence on Head and Neck Cancers. Curr Oncol Rep. 2023;25(3):163-171.

13.

Jiang

Wang

Xia

, et al. A randomized, double-blind, placebo-controlled trial of probiotics to reduce the severity of oral mucositis induced by chemoradiotherapy for patients with nasopharyngeal carcinoma. Cancer. 2019;125(7):1081-1090.

14.

Hes

Desilets

Tonneau

, et al. Gut microbiome predicts gastrointestinal toxicity outcomes from chemoradiation therapy in patients with head and neck squamous cell carcinoma. Oral Oncol. 2024;148:106623.

15.

Spieth

Kubasch

Penzlin

Illigens

BMW

Barlinn

Siepmann

. Randomized controlled trials - a matter of design. Neuropsychiatr Dis Treat. 2016;12:1341-1349.

16.

Smith

Ebrahim

. Mendelian randomization': can genetic epidemiology contribute to understanding environmental determinants of disease? Int J Epidemiol. 2003;32(1):1-22.

17.

Davey Smith

Hemani

. Mendelian randomization: genetic anchors for causal inference in epidemiological studies. Hum Mol Genet. 2014;23(R1):R89-R98.

18.

Sekula

Del Greco M

Pattaro

Köttgen

. Mendelian Randomization as an Approach to Assess Causality Using Observational Data. J Am Soc Nephrol. 2016;27(11):3253-3265.

19.

Cui

Zheng

, et al. Genetically proxied gut microbiota and cancer risk: a scoping review of Mendelian randomization studies. Arch Public Health. 2025;83(1):209.

20.

Liu

Xie

, et al. Association of gut microbiome and oral cavity cancer: A two sample mendelian randomization and case-control study. J Stomatol Oral Maxillofac Surg. 2024;125(4):101736.

21.

Lian

Sun

Han

Baranova

Cao

Zhang

. Gut microbiome's causal role in head and neck cancer: findings from mendelian randomization. Front Oncol. 2024;14:1453202.

22.

Sun

Bai

Hao

Jiang

Chen

. Gut microbiota and oral cavity cancer: a two-sample bidirectional Mendelian randomization study. Front Oncol. 2024;14:1389678.

23.

Rao

Zhang

Zha

, et al. Plasma proteins mediate the effects of the gut microbiota on the development of head and neck cancer: a two-sample and mediated Mendelian randomized study. Discov Oncol. 2025;16(1):202.

24.

Raudenská

Bugajová

Kalfeřt

, et al. The interplay between microbiome and host factors in pathogenesis and therapy of head and neck cancer. Biochim Biophys Acta Rev Cancer. 2024;1879(6):189216.

25.

Fan

Chen

. Oral microbiota-driven immune modulation along the oral-gut axis: from local signals to systemic inflammation. NPJ Biofilms Microbiomes. 2026;12(1):46.

26.

Nouri

Choi

, et al. Exploring Connections between Oral Microbiota, Short-Chain Fatty Acids, and Specific Cancer Types: A Study of Oral Cancer, Head and Neck Cancer, Pancreatic Cancer, and Gastric Cancer. Cancers (Basel). 2023;15(11):2898.

27.

Kurilshikov

Medina-Gomez

Bacigalupe

, et al. Large-scale association analyses identify host factors influencing human gut microbiome composition. Nat Genet. 2021;53(2):156-165.

28.

Burgess

Dudbridge

Thompson

. Combining information on multiple instrumental variables in Mendelian randomization: comparison of allele score and summarized data methods. Stat Med. 2016;35(11):1880-1906.

29.

Kulinskaya

Dollinger

. An accurate test for homogeneity of odds ratios based on Cochran's Q-statistic. BMC Med Res Methodol. 2015;15:49.

30.

Hemani

Tilling

Davey Smith

. Orienting the causal relationship between imprecisely measured traits using GWAS summary data. PLoS Genet. 2017;13(11):e1007081.

31.

Yavorska

Burgess

. MendelianRandomization: an R package for performing Mendelian randomization analyses using summarized data. Int J Epidemiol. 2017;46(6):1734-1739.

32.

Pongen

Thirumurugan

Ramasubburayan

Prakash

. Harnessing actinobacteria potential for cancer prevention and treatment. Microb Pathog. 2023;183:106324.

33.

Chen

Liu

Zou

, et al. A high-fat diet promotes cancer progression by inducing gut microbiota-mediated leucine production and PMN-MDSC differentiation. Proc Natl Acad Sci U S A. 2024;121(20):e2306776121.

34.

Zeng

, et al. Gut microbiota dysbiosis promotes the development of epithelial ovarian cancer via regulating Hedgehog signaling pathway. Gut Microbes. 2023;15(1):2221093.

35.

Al-Qadami

Bowen

Van Sebille

, et al. Baseline gut microbiota composition is associated with oral mucositis and tumour recurrence in patients with head and neck cancer: a pilot study. Support Care Cancer. 2023;31(1):98.

36.

Bibb

. Understanding and manipulating antibiotic production in actinomycetes. Biochem Soc Trans. 2013;41(6):1355-1364.

37.

Macfarlane

Cummings

Allison

. Protein degradation by human intestinal bacteria. J Gen Microbiol. 1986;132(6):1647-1656.

38.

Kim

Park

. Conjugated Linoleic Acid: Potential Health Benefits as a Functional Food Ingredient. Annu Rev Food Sci Technol. 2016;7:221-244.

39.

McCarville

Dong

Caminero

, et al. A Commensal Bifidobacterium longum Strain Prevents Gluten-Related Immunopathology in Mice through Expression of a Serine Protease Inhibitor. Appl Environ Microbiol. 2017;83(19).

40.

Cano

Santacruz

Trejo

Sanz

. Bifidobacterium CECT 7765 improves metabolic and immunological alterations associated with obesity in high-fat diet-fed mice. Obesity (Silver Spring). 2013;21(11):2310-2321.

41.

O'Mahony

Scully

O'Mahony

, et al. Commensal-induced regulatory T cells mediate protection against pathogen-stimulated NF-kappaB activation. PLoS Pathog. 2008;4(8):e1000112.

42.

Notting

Pirovano

Sybesma

Kort

. The butyrate-producing and spore-forming bacterial genus Coprococcus as a potential biomarker for neurological disorders. Gut Microbiome (Camb). 2023;4:e16.

43.

Hamer

Jonkers

Venema

Vanhoutvin

Troost

Brummer

. Review article: the role of butyrate on colonic function. Aliment Pharmacol Ther. 2008;27(2):104-119.

44.

Liu

Yang

Zhang

Gatlin

Ringø

Zhou

. Effects of dietary microencapsulated sodium butyrate on growth, intestinal mucosal morphology, immune response and adhesive bacteria in juvenile common carp (Cyprinus carpio) pre-fed with or without oxidised oil. Br J Nutr. 2014;112(1):15-29.

45.

Geng

Yin

Zhang

Gong

Yang

. Butyrate Suppresses Glucose Metabolism of Colorectal Cancer Cells via GPR109a-AKT Signaling Pathway and Enhances Chemotherapy. Front Mol Biosci. 2021;8:634874.

46.

Louis

Flint

. Diversity, metabolism and microbial ecology of butyrate-producing bacteria from the human large intestine. FEMS Microbiol Lett. 2009;294(1):1-8.

47.

Chen

Wang

Zhan

Zhang

Wang

. Fecal metabonomics combined with 16S rRNA gene sequencing to analyze the changes of gut microbiota in rats with kidney-yang deficiency syndrome and the intervention effect of You-gui pill. J Ethnopharmacol. 2019;244:112139.

48.

Nie

Chen

Geng

Nie

. Arabinoxylan ameliorates type 2 diabetes by regulating the gut microbiota and metabolites. Food Chem. 2022;371:131106.

49.

Wang

Zhang

. Alterations to the Gut Microbiota and Their Correlation With Inflammatory Factors in Chronic Kidney Disease. Front Cell Infect Microbiol. 2019;9:206.

50.

Wang

Zheng

Yang

. Oral Administration of Resveratrol-Selenium-Peptide Nanocomposites Alleviates Alzheimer's Disease-like Pathogenesis by Inhibiting Aβ Aggregation and Regulating Gut Microbiota. ACS Appl Mater Interfaces. 2021;13(39):46406-46420.

51.

Murros

Huynh

Takala

Saris

PEJ

. Desulfovibrio Bacteria Are Associated With Parkinson's Disease. Front Cell Infect Microbiol. 2021;11:652617.

52.

Cai

Yang

, et al. High-Fat Diet Enhances the Liver Metastasis Potential of Colorectal Cancer through Microbiota Dysbiosis. Cancers (Basel). 2022;14(11):2573.

53.

Liu

, et al. Desulfovibrio desulfuricans and its derived metabolites confer resistance to FOLFOX through METTL3. EBioMedicine. 2024;102:105041.

54.

van Liere

de Boer

NKH

Dekker

van Leerdam

de Meij

TGJ

Ramsoekh

. Systematic review: non-endoscopic surveillance for colorectal neoplasia in individuals with Lynch syndrome. Aliment Pharmacol Ther. 2022;55(7):778-788.

55.

Yan

Drew

Markowitz

, et al. Structure of the Mucosal and Stool Microbiome in Lynch Syndrome. Cell Host Microbe. 2020;27(4):585-600.e4.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.31 MB

0.09 MB

0.00 MB