Causal influence of gut microbiota on small cell lung cancer: a Mendelian randomization study

Previous studies have hinted at a significant link between lung cancer and the gut microbiome, yet their causal relationship remains to be elucidated.

occur in Asia. 4 In China, lung cancer has been the leading cause of death among all malignant tumors since 2011. 5Small cell lung cancer (SCLC), one of the primary types of lung cancer, is a high-grade neuroendocrine carcinoma predominantly found in smokers.Moreover, compared to other lung cancer types, SCLC has a notably poor prognosis. 6While smoking is the primary risk factor for SCLC, other risk factors such as exposure to asbestos, radiation, and environmental pollution should not be overlooked. 1,7,8he microbiome is increasingly recognized as a pivotal player in cancer onset, progression, and response to chemotherapy.In recent years, both preclinical and clinical studies have established a connection between the microbiome and lung cancer. 9The gut and lungs share a common embryonic origin, 10 and there is physical interaction between them, as ingested microbes can enter the gastrointestinal and respiratory tracts and gastroesophageal contents can be aspirated into the lungs. 11ue to this extensive dialogue between the gut and lungs, research targeting the gut-lung axis (GLA) has become paramount in disease studies in recent years. 12However, compared to other diseases like Crohn's disease and colitis, our understanding of how the microbiome affects the lungs remains limited.A study investigating female nonsmoking lung cancer patients revealed correlations between the gut microbiome and TNM staging and primary tumor size. 13Specifically, a significant positive correlation was found between the relative abundance of Faecalibacterium and primary tumor size, while a significant negative correlation was observed with Clostridium and Pseudomonas.Another study suggested that lung cancer patients have lower concentrations of Firmicutes and Proteobacteria and relatively higher levels of Pseudomonas and Clostridium compared to healthy individuals. 14Given the vast individual variations in gut microbiota, it is unsurprising to observe inconsistencies in the same microbial communities across different lung cancer patients.It is crucial to note that these studies were observational, conducted on clinical samples.The observed changes might be attributed to the effects of bacterial metabolic by-products or molecules on the immune system.For instance, in individuals with existing tumors, Pseudomonas and Faecalibacterium might suppress tumor proliferation by activating T ncells. 15owever, whether these microbial communities exert similar anti-tumor effects in tumor-free individuals remains uncertain.Moreover, in observational studies, the association between gut microbiota and SCLC can easily be confounded by factors like age, environment, dietary patterns, and lifestyle, which are challenging to control effectively.These conditions limit the causal inference between the gut microbiome and SCLC.
In this study, we employ Mendelian randomization (MR) as an innovative methodology to delve into the potential causal linkage between the gut microbiome and SCLC. 16MR capitalizes on genetic variations to formulate instrumental variables (IVs) for exposures, thereby seeking to elucidate the causal nexus between these exposures and disease manifestations. 17Since genetic variances are inherently determined, their associations with outcomes are insulated from prevalent confounders, rendering MR particularly suited for discerning authentic causal connections.The MR approach has been prolifically harnessed to investigate the causal interplay between the gut microbiome and a spectrum of ailments, encompassing gynecological maladies, 18 autoimmune disorders, 16,19 and metabolic syndromes. 20Drawing upon the genome-wide association study (GWAS) summary data from the MiBio-Gen and FinnGen consortia, we embarked on a unidirectional MR evaluation to gauge the causal interrelation between the gut microbiome and SCLC.

| Data sources
Genetic variation data for the gut microbiome were obtained from the most extensive gut microbiome composition genome-wide meta-analysis published by the MiBioGen consortium. 21This study encompasses 16SrRNA gene sequencing spectra and genotyping data from 18 340 individuals across 24 countries, including the United States, the United Kingdom, Finland, Sweden, Denmark, and the Netherlands.The summarized data from this research includes bacteria from nine phyla, 16 classes, 20 orders, 35 families, and 131 genera.Statistical data for SCLC were sourced from the FinnGen consortium. 22This GWAS comprises 218 792 European adult participants, with 179 cases and 218 613 controls.

| Selection of IVs
With the gut microbiome as the exposure and SCLC as the outcome, the IVs, that is, single nucleotide polymorphisms (SNPs), must satisfy three critical assumptions in Figure 1 23 : 1. Relevance assumption: The IV is strongly correlated with the exposure.2. Exclusion restriction assumption: The IV is not related to the outcome.3. Independence assumption: The IV is not associated with confounders that could lead to the outcome.
Consistent with many contemporary MR studies, 24 we employed a genome-wide significance threshold of p < 5 Â 10 À8 for SNP selection.Due to the scarcity of SNPs reaching this stringent threshold, a more lenient threshold of p < 5 Â 10 À6 was adopted for potential IVs for each exposure of interest.To ensure IV independence, we instituted linkage disequilibrium clumping using a clumping window of 10 MB and an R 2 threshold of <0.001, referencing European ancestry data from the 1000 Genomes Project.To mitigate bias from potentially weak instruments, we computed F-statistics for each SNP, retaining only those deemed as strong IVs (Fstatistics > 10).Additionally, ambiguous and palindromic SNPs, which could not be rectified during harmonization, were systematically excluded.

| Statistical analysis
We employed the inverse variance weighted (IVW) method as the primary analysis for MR.Additionally, four other methods, including the weighted median estimator (WME), MR-Egger analysis, simple mode analysis, and weighted mode analysis, were applied as supplementary references.The IVW method provides consistent estimates only when all SNPs serve as valid IVs, without considering the presence of an intercept term, implying no pleiotropy for all IVs. 25,26The WME is based on the assumption that over half of the SNPs are valid IVs. 27In contrast, the MR-Egger method assumes all SNPs are invalid IVs, defaulting to the presence of an intercept term. 28When IVW results are significant, further All results of Mendelian randomization (MR) analysis and sensitivity analysis between gut microbiota and small cell lung cancer (SCLC).
sensitivity analyses are incorporated.For exposures with only one IV, the Wald ratio method is employed to estimate the effect of exposure on the outcome.Sensitivity analyses encompassed heterogeneity and pleiotropy tests.The MR-Egger intercept test detects horizontal pleiotropy, where its intercept represents potential pleiotropy, and its p-value validates the significance of this pleiotropy. 29Cochran's Q test examines the heterogeneity of IVs, with a p-value greater than 0.05 indicating no heterogeneity. 30Moreover, we employed the "leaveone-out analysis" to determine if a significant association is driven by a single SNP, identifying potential outlier SNPs.Subsequent to this, SNPs exhibiting pleiotropy or heterogeneity were identified and removed based on the leave-one-out results.
Statistical F-statistics were computed using the for- 31 where n represents the sample size, k represents the number of IVs, and R 2 represents the proportion of variance in the exposure explained by genetic variation.When evaluating a single IV, k = 1.An F-statistic > 10 for a single SNP indicates no significant weak instrument bias; otherwise, that IV should be excluded.
After excluding the aforementioned non-compliant IVs, we repeated the MR analysis to obtain the final MR estimates.In the absence of heterogeneity and pleiotropy, IVW serves as a more reliable fitting model. 28For binary outcomes, effect estimates are presented as odds ratios (OR) with 95% confidence intervals.
MR analyses were conducted in the R computational environment (version 4.2.3) using the TwoSampleMR package (version 0.5.6).A p-value < 0.05 was considered statistically significant evidence of a causal effect.

| IV selection and preliminary MR analysis
Based on our IV selection criteria, we identified 2426 SNPs as instruments for examining 211 distinct gut microbiota taxa.Each selected SNP demonstrated an Fstatistic greater than 10, indicating a robust association with the microbiota taxa and minimizing the concern of weak instrument bias in our analysis.The relationships between these 211 bacterial taxa and SCLC, facilitated by the chosen SNPs, are visually represented in Figure 1.For an in-depth review of the SNPs utilized as IVs, including their association strengths and other relevant statistics, please refer to Table S1.Moreover, the comprehensive results of the MR analysis, detailing the influence of each microbiota taxon on SCLE risk, are systematically compiled in Table S2.

| Detailed MR analysis
The results of our MR analysis, depicted in Figure 2, demonstrate that the IVW method identified significant associations between five bacterial taxa and SCLC.To ensure the robustness of these findings, we further employed the WME and MR-Egger methods.While significant associations were observed with the IVW method, the analyses conducted using the WME and MR-Egger methods did not yield statistically significant results for the taxa under investigation.This variation in outcomes across different MR methods might indicate underlying heterogeneity or pleiotropy within the IVs selected for our study.The divergent results from these methods illuminate the intricate role gut microbiota may play in influencing SCLC, emphasizing the necessity of employing a multifaceted MR approach to corroborate causal relationships.Specifically, the genus Eubacterium ruminantium group (OR = 0.413, 95% CI: 0.223-0.767,p = 0.00513), genus Barnesiella (OR = 0.208, 95% CI: 0.0640-0.678,p = 0.00919), family Lachnospiraceae (OR = 0.319, 95% CI: 0.107-0.948,p = 0.03979), and genus Butyricimonas (OR = 0.376, 95% CI: 0.144-0.984,p = 0.04634) were found to have a protective effect against SCLC.

| Sensitivity analysis
The results of the sensitivity analysis are presented in Table 1.Cochran's Q test revealed that the test values for six bacterial genera were all greater than 0.05, indicating that there is no heterogeneity among the IVs for these taxa.Furthermore, the differences between the intercepts from the MR-Egger method and zero were not significant, confirming the absence of horizontal pleiotropy in the results.Scatter plots, as depicted in Figure S1, highlight the effect trends across different MR methods without any significant outliers.Concurrently, the leaveone-out analysis, presented in Figure S2, underscores the data's robustness, indicating that no individual SNP had a disproportionate impact on the overall findings.This further solidifies the reliability of our MR results.

| DISCUSSION
In this study, we employed MR to analyze 211 prevalent microbial taxa in the gut.Our findings suggest that four bacterial taxa, namely, the genus E. ruminantium group, genus Barnesiella, family Lachnospiraceae, and genus Butyricimonas, may have protective effects against SCLC.Conversely, the genera Intestinibacter, E. oxidoreducens group, Bilophila, and the order Bacillales might contribute to the onset of SCLC.
Previous studies have suggested that the gut microbiota of lung cancer patients differs significantly from F I G U R E 2 Forest plot of causal relationships estimated for gut microbiome and small cell lung cancer (SCLC) using the inverse variance weighted method.
T A B L E 1 Mendelian randomization sensitivity analysis of the influence of gut microbiota on small cell lung cancer risk.

No.
3][34][35][36] Moreover, the use of antibiotics or probiotics as adjunctive treatments for lung cancer, especially in conjunction with immunotherapy or mutation prevention, underscores the close association between alterations in the gut microbiome and lung cancer.However, due to the inherent limitations of observational studies and animal experiments, these studies are insufficient to determine the specific roles of different gut microbial taxa in lung cancer.Notably, current research is predominantly focused on non-small cell lung cancer (NSCLC), with scant studies on SCLC, which also has a high incidence rate.We conducted a comprehensive systematic analysis of 211 taxa of the gut microbiome, identifying specific bacteria that may have beneficial or detrimental effects on SCLC patients.This contributes to a deeper understanding of the pathogenesis of SCLC and offers potential avenues for novel clinical treatment strategies. 37In a previous study, it was also observed 38 that patients with SCLC exhibited a significant decrease in the abundance of the family Lachnospiraceae in their gut microbiota composition.Our research corroborates this finding, suggesting that family Lachnospiraceae might be one of the protective factors against SCLC.Earlier studies have also indicated that the Spirochaetaceae family can protect the host from cancer by producing butyrate. 39his could potentially explain the protective mechanism of family Lachnospiraceae against SCLC.Moreover, for the first time, we identified that the genus Butyricimonas might act as a protective factor for SCLC.Given that genus Butyricimonas is also a butyrate-producing bacterium, we speculate that it might share a similar mechanism with Spirochaetaceae, exerting antitumor effects through the anti-inflammatory properties of butyrate. 40owever, this hypothesis warrants further investigation for validation.
On the other hand, we found that the genus Intestinibacter, genus Bilophila, and order Bacillales might promote the onset of SCLC.The genus Intestinibacter has previously been associated with NSCLC. 41,42Its abundance is notably increased in the gut microbiota of NSCLC patients.Our study further suggests a causal link between this bacterial group and the occurrence of SCLC.We hypothesize that this might be achieved by recruiting myeloid-derived suppressor cells, tumor-associated macrophages, and regulatory T cells to inhibit the anti-tumor immune response, leading to the development of SCLC. 43s for the genus Bilophila, there are no direct reports linking it to tumor development.However, in a study on PD-1 treatment, it was observed that its abundance significantly increased in non-responders, indicating that it might not be beneficial in lung cancer patients.We speculate that since the genus Bilophila is a hydrogen sulfideproducing bacterium and hydrogen sulfide is a genotoxic compound, 44 it has been proven to damage DNA, leading to genomic or chromosomal instability.This might be one of the reasons it promotes the onset of SCLC or tumor mutations.A pilot study suggested that the abundance of order Bacillales increases in lung cancer patients, a finding that our research also confirms. 45We hypothesize that the increased number of order Bacillales might allow endotoxins to continuously enter the circulation through a compromised gut barrier, triggering a systemic inflammatory response, which in turn promotes tumor development. 46Targeting these harmful bacterial groups therapeutically might help reduce the incidence of SCLC or enhance the therapeutic effects against SCLC.
Our research has identified some beneficial and harmful bacterial groups involved in the pathogenesis of SCLC.However, this evaluation is based on causality, and further research is needed in the future to clarify its mechanism.It is also worth noting that our study was conducted at the genus level of the entire microbial community.Different species of bacteria within the same genus may have different pathological or physiological effects, and we need to be aware of the heterogeneity that exists within them.As we found in our study, the genus E. ruminantium group plays a protective role against SCLC, while the genus E. oxidoreducens group from the same genus might promote the onset of SCLC.Both of these bacterial groups are Firmicutes, which have been extensively studied and shown to undergo significant changes in lung cancer patients, [47][48][49] but their mechanisms are still not fully understood.On one hand, an increase in Firmicutes has been linked to increased gut permeability and chronic inflammation. 50,51The expression of inflammatory factors such as interleukin (IL)-6 and IL-1b increases, which may promote the onset of SCLC through pro-inflammatory effects.On the other hand, the abundance of Eubacterium in the gut is related to the level of short-chain fatty acids (SCFAs), and high levels of SCFAs have beneficial effects on tumors under clinical conditions. 47Therefore, it is not surprising that different species of bacteria from the same genus have different effects in SCLC.
In summary, the complex interactions among gut microbiota might explain the discrepancies between genetic prediction results and clinical observations.We speculate that this is largely due to the involvement of gut microbes in inflammation and immune regulation, thereby participating in the entire pathophysiological process of SCLC.However, further prospective randomized controlled trials are needed to validate our conclusions.Specifically, future research should explore the potential of gut probiotic interventions or antibiotic treatments for SCLC, especially when used in combination with immune checkpoint inhibitors.There are certain limitations in our study.Firstly, although our research analyzed common gut microbial communities, the gut microbiota is vast and highly heterogeneous.Moreover, our study only analyzed populations from Europe, which means that caution is needed when extrapolating our findings to individuals of other ethnicities.

| CONCLUSION
In summary, our research findings support the notion that the bacterial taxa genus E. ruminantium group, genus Barnesiella, family Lachnospiraceae, and genus Butyricimonas may have protective effects against SCLC.On the other hand, the bacterial taxa genus Intestinibacter, genus E. oxidoreducens group, genus Bilophila, and order Bacillales might promote the development of SCLC.It is particularly noteworthy that the genus Eubacterium might have dual effects on human health, with specific species and strains potentially having different impacts.Therefore, further research is needed to elucidate the potential mechanisms of these bacterial taxa in the context of SCLC.

AUTHOR CONTRIBUTIONS
Wenjing Yang formulated the research questions and design, conducted the analysis, and drafted the manuscript for submission; Yan Chen and Wangshu Li contributed to drafting the submitted manuscript; Xinxia Fan assisted in data analysis and performed statistical analysis.All authors have read and agreed to the published version of the manuscript.