The role of gut microbiota in prostate cancer (PCa) has received a lot of attention in recent years (Rizzo et al., 2022), and the concept of “gut-prostate axis” has been proposed (Fujita et al., 2023; Matsushita et al., 2023). Previous basic studies suggested that the underlying mechanisms of gut microbiota's effects on the incidence and progression of PCa may involve modulating inflammation, oxidative stress, and immune function and interfering with lipid metabolism (Porter et al., 2018; Wheeler and Liss, 2019). Nevertheless, evidence from population-based studies on the relationship between gut microbiota and PCa remains limited, and most of the existing studies are small-sample case-control studies.

Studies have suggested that patients with PCa exhibit an increased relative abundance of the bacterial genera Veillonella, Bacteroides (Alanee et al., 2019), Streptococcus (Liss et al., 2018), Rikenellaceae, Alistipes, and Lachnospira (Matsushita et al., 2021). While in control groups, the relative abundance of Faecalibacterium prausnitzii and Eubacterium rectalie was higher than that in cases (Golombos et al., 2018). However, the firm conclusions for the causal relationship and its direction between gut microbiota and PCa are not yet enough to draw according to the existing observational studies' evidence. Inherent defects limited traditional observational studies, and thus the causal role of gut microbiota in the risk of PCa remained to be clarified (Tong et al., 2020). In spite of the fact that randomized controlled trial is recognized as a gold standard for determining causality, its application in clinical settings is impractical due to the long incubation period from certain microbiota exposure to oncogenesis (Spieth et al., 2016). Under the circumstances, a novel method for investigating the causal association between gut microbiota and PCa is warranted.

Mendelian randomization (MR) analysis has become a widely used approach that exploits single nucleotide polymorphisms (SNPs) as unconfounded instrumental variants (IVs) to explore the potential causal relationships between environmental exposures and diseases (Davey Smith and Hemani, 2014; Sekula et al., 2016). MR can avoid reverse causality inferences and reflect the long-term effects of exposures on outcomes. A recent two-sample MR analysis assessed the causal effect of gut microbiota on cancer risk and showed that the relative abundance of Alphaproteobacteria, Rhodospirillales, Adlercreutzia, and Coprobacter was associated with PCa (Wei et al., 2023). However, the findings of Wei's study were not in accordance with those of other epidemiological studies, and they also did not explore the causal associations between gut microbiota and cancer risk at the species level.

Herein, in this two-sample MR study, we aimed to investigate the potential causal relationship between gut microbiota at different levels, especially at the species level, and PCa in order to provide some new ideas for exploring the methods for PCa prevention and treatment.

We conducted an MR analysis to explore the potential causal relationship between gut microbiota and PCa. The results showed that Melainabacteria (class level), Gastranaerophilales (order level), Prevotellaceae (family level), Acidaminococcaceae (family level), Ruminococcus torques group (genus level), Lachnospiraceae UCG-008 (genus level), Akkermansia muciniphila (species level), Bacteroides salyersiae (species level), and Eubacterium biforme (species level) were associated with PCa. To be specific, the relative abundance of Akkermansia muciniphila and Bacteroides salyersiae was both negatively associated with the odds of PCa, while that of Eubacterium biforme was positively associated with the odds of PCa. Besides, no reverse causality has been found between them.

The firm conclusions for the causal relationship between gut microbiota and PCa are not yet enough to draw according to the evidence from existing observational studies. In recent years, only a few studies have explored the association between gut microbiota and the risk of cancer using MR analyses, which is a widely used approach to explore the potential causal relationships between environmental exposures and diseases. For example, a univariable and multivariable MR study by Wei et al. (2023) assessed the causal effect of gut microbiota on five common cancers, including breast, endometrial, lung, ovarian, and PCa. Their results showed that a higher abundance of class Alphaproteobacteria was associated with a lower risk of PCa. Differently, in our study, we focused on the relationship between gut microbiota and PCa only and observed gut microbiota at phylum, class, order, family, genus, and species levels. We found that the relative abundance of Akkermansia muciniphila, Bacteroides salyersiae, and Eubacterium biforme was all associated with PCa. Our study detailed the group of gut microbiotas and found three specific species, which may provide some references for further basic and prospective research on the causal relationship between gut microbiota and PCa. In Wei's study, the associations between SNPs and PCa were obtained from the GWAS study from the Prostate Cancer Association Group to Investigate Cancer-Associated Alterations in the Genome (PRACTICAL) Consortium, which consists of cases diagnosed with PCa and controls of European descent. Another two-sample MR study by Long et al. (2023), also based on the population from the PRACTICAL Consortium, examined the causal relationship between gut microbiota and cancer. Although we used the GWAS study from the Finngen Consortium, the race of the study population was similar. Long's results showed that the genus Ruminococcustorquesgroup, class Verrucomicrobiae, family Verrucomicrobiaceae, order Verrucomicrobiales, genus Terrisporobacter, genus Roseburia, and class Alphaproteobacteria were causally associated with PCa. The different gut microbiota we explored may complement previous studies.

In fact, the underlying mechanisms of these gut microbiota and PCa are complex and unclear. The class Melainabacteria has been identified as an accurate biomarker of zinc (Zn) status in the human body (Chen et al., 2021). As Zn plays a growth-modulatory role in PCa, Melainabacteria may influence the occurrence and development of PCa by modulating Zn levels (To et al., 2020; Zhang et al., 2022). Melainabacteria exist in groundwater, wastewater treatment plants, and herbivorous mammal and human guts and have the function of synthesizing vitamins B and K, suggesting they are beneficial bacteria to their hosts (Di Rienzi et al., 2013). Similarly to the previous studies, our findings showed that an increased abundance of Melainabacteria was associated with low odds of PCa. Gastranaerophilales is one of the probiotics with impaired abundance in colitis (Wang et al., 2021; Wu et al., 2021). Inflammation is a risk factor for prostate carcinogenesis, with diet, chemical injury, and an altered microbiome being causally implicated (de Bono et al., 2020). However, no study has been conducted to describe the potential mechanisms by which Gastranaerophilales play a protective role in prostate carcinogenesis, which needs further basic research for clarification. Prevotellaceae is also a probiotic that plays an important role in colitis and some cancers (Zhang L. et al., 2019; Qu et al., 2021). Li et al. (2022) performed an analysis on alterations of gut microbiota diversity, composition, and metabonomics in benign prostatic hyperplasia rats and showed that there was a strong correlation between Prevotellaceae and differential metabolites. However, the specific mechanism of the potential protective role of Prevotellaceae in PCa has not been clear. According to previous studies, we speculated that Prevotellaceae may moderate inflammation (Qu et al., 2021), oxidative stress (Cui et al., 2018), and metabolic disorders (Li et al., 2022) and further influence PCa development.

The high relative abundance of the family Acidaminococcaceae, genus Ruminococcus torques group, and genus Lachnospiraceae UCG-008 was found to be associated with a high risk of PCa in this study. Acidaminococcaceae clustered into one group with the Veillonellaceae, which has been reported as a marker of dysbacteriosis and plays a possible role in carcinogenesis (Yan et al., 2015; Kasai et al., 2016). A bidirectional, two-sample MR study on the association between Graves' disease and the gut microbiome showed that the Ruminococcus torques group was identified as a risk factor (Cao et al., 2023). Ruminococcus torques group is a genera derived from the genus Mediterraneibacter from the family Lachnospiraceae and is identified as a butyrate-producing bacterium (Salyers et al., 1977). Butyrate-producing bacteria were found to promote fat deposition because they are able to convert dietary fiber to butyrate by fermentation (Yang et al., 2010; Blaut, 2015). The abundance of Ruminococcus torques group is thus supposed to be positively related to fat accumulation. Excess fat can accelerate the growth of prostate tumors by inducing inflammation (Hayashi et al., 2018). In addition, altered lipid metabolism, especially the excessive accumulation of cholesterol and fatty acids, promotes the malignant transformation of PCa via the formation of cholesteryl esters (Wang et al., 2022). Lachnospiraceae UCG-008 is considered a potential harmful genus in human colonic microbiota (Huang et al., 2019). Some medicines improve gut microbial dysbiosis by increasing beneficial bacteria and decreasing harmful bacteria, including Lachnospiraceae UCG-008 in order to relieve diseases (Huang et al., 2019; Yin et al., 2021). In conclusion, how gut microbiota influence the development of PCa by disturbing the metabolism is still unclear; further studies are needed to clarify the specific biological mechanisms.

We additionally explored the relationships between gut microbiota and the risk of PCa at the species level. Our findings indicated that the high relative abundance of Akkermansia muciniphila and Bacteroides salyersiae was both associated with a low risk of PCa, while that of Eubacterium biforme was positively associated with the risk of PCa. Akkermansia muciniphila is a Gram-negative anaerobic bacterium that contributes to homeostasis maintenance and barrier integrity in the gastrointestinal tract (Zhai et al., 2019; Zhang T. et al., 2019). Recently, it has been reported that intravenous injection of Akkermansia muciniphila-derived extracellular vesicles in immune-competent mice reduced the tumor burden of PCa without inducing obvious toxicity in normal tissues, indicating a potential association between the abundance of Akkermansia muciniphila and PCa (Luo et al., 2021). Similarly, our MR study supplemented and suggested a potential causal relationship between the high relative abundance of Akkermansia muciniphila and the low risk of PCa. Wang et al. (2020) found the membrane protein from Akkermansia muciniphila, namely Amuc_1100, was associated with tumourigenesis of colorectal cancer through modulation of CD8 T cells in mice. Gu et al. (2021) also suggested that Akkermansia muciniphila and its outer protein Amuc_1100 regulated the tryptophan metabolism in colitis. Another animal study also showed the effect of improved Amuc_1100 from Akkermansia muciniphila on metabolism in obese and diabetic mice (Plovier et al., 2017). As the mechanism of PCa development mentioned previously, Akkermansia muciniphila and Amuc_1100 may also influence PCa by improving inflammation and regulating glucose metabolism and lipid metabolism, but further mechanistic exploration is needed. Bacteroides salyersiae is a beneficial commensal with Akkermansia muciniphila and plays a similar role in the host (Derosa et al., 2020). Our study considered that the application of probiotics in the prevention and treatment of PCa was very promising. Gut microflora evolves with a complex polysaccharide-rich diet and dietary fermentation, resulting in the production of short-chain fatty acids such as butyrate, which represent a primary energy source for colonic epithelial cells and preserve them from inflammation (Ahmad et al., 2000; Atarashi et al., 2013). Eubacterium biforme is one of them. Aside from this, a lack of fiber in diets is associated with a low concentration of short-chain fatty acids (Pituch-Zdanowska et al., 2015). Daily attention to dietary fiber consumption may be the key to preventing inflammatory diseases by regulating intestinal flora. Daily consumption of bifidobacteria-rich foods such as yogurt may be a viable way to regulate probiotics; however, whether a high relative abundance of Eubacterium biforme is associated with a high risk of PCa is not determinate.

As mentioned before, MR is a relatively good study design to clarify the causal effect of potential risk factors on diseases of interest. By exploring the gut microbiota that modulate PCa risk, MR studies facilitate the recommendation of public health policies and clinical interventions that effectively reduce the incidence and social burden of PCa. With the rapid development of omics technologies in recent years, including genomics, proteomics, transcriptomics, metabolomics, and epigenomics, researchers can use a large number of novel exposures/intermediate phenotypes generated in observational studies to assess associations with clinical endpoints. The MR approaches summarize the results from published GWAS studies facilitately and do not need a separate study to carry out MR analyses (Sekula et al., 2016). Compared with observational studies, which commonly suffer from biased results due to confounding, MR can deal with any confounding by design as long as a valid genetic instrumental variable is available. However, two major limitations also influence the judgment of causation through MR approaches. First, no valid instrument for every research question is available because of a lack of knowledge about MR, and the publicly available sources of data to provide information on the associations of interest are not always available. Second, MR studies have potential limitations in the statistical power of the study design, which depends on several aspects, such as the proportion of variance in the exposure explained by the genetic instrumental variable and the magnitude of the causal association between exposure and outcome (Brion et al., 2013; Burgessm, 2014). Besides, in recent years, artificial intelligence has been explored to improve the diagnosis of PCa, including radiological and histological diagnoses (Ström et al., 2020; Mata et al., 2021). Ramírez-Mena et al. (2023) combined gene expression and AI for the detection and screening of PCa and found it can decrease the misclassification rates of anatomopathological analysis, thus reducing the need for repeated biopsies. Their algorithm may be applied to urine or blood samples rather than traditional living tissue, which could be a part of the liquid biopsy strategy in PCa in the future. To the best of our knowledge, AI has not been applied in the calculation of causal association between the gut microbiota and the risk of PCa, and we believe the potential causal relationship indicated by MR approaches can provide some references for further studies that focus on the underlying mechanisms.

There are some strengths and limitations in this study. The current study refined the taxonomic units of gut microbiota to the species level and explored their relationship with the risk of PCa, which may be more comprehensive compared to the previous studies. Nevertheless, there were two major limitations in this study. Our study data were extracted from the MiBioGen consortium, DMP, and FinnGen consortium, which may cause the selection bias. Furthermore, these databases only contain the European population, so the potential causal relationship between gut microbiota and the risk of PCa is limited, and whether the results are generalizable to other populations requires further studies to clarify. Another limitation was that the locations of the SNPs we used (whether they are present in the enhancer/promoter/coding region of the genome) are not available, which limited the explanation for the mechanisms of the potential causal association between gut microbiota and PCa.

Our study suggested a potential causal relationship between gut microbiota and the risk of PCa. Further studies are warranted to elucidate the causal association and specific underlying mechanisms of gut microbiota and the development of PCa.

Publicly available datasets were analyzed in this study. This data can be found here: the GWAS Catalog, https://www.ebi.ac.uk/gwas/.

QX and BH designed the study, collected, analyzed, and interpreted the data. QX wrote the manuscript. BH critically reviewed, edited, and approved the manuscript. Both authors read and approved the final manuscript.