Navigating the gut-bone axis: The pivotal role of Coprococcus3 in osteoporosis prevention through Mendelian randomization

Osteoporosis (OP) constitutes a notable public health concern that significantly impacts the skeletal health of the global aging population. Its prevalence is steadily escalating, yet the intricacies of its diagnosis and treatment remain challenging. Recent investigations have illuminated a profound interlink between gut microbiota (GM) and bone metabolism, thereby opening new avenues for probing the causal relationship between GM and OP. Employing Mendelian randomization (MR) as the investigative tool, this study delves into the causal rapport between 211 varieties of GM and OP. The data are culled from genome-wide association studies (GWAS) conducted by the MiBioGen consortium, in tandem with OP genetic data gleaned from the UK Biobank, BioBank Japan Project, and the FinnGen database. A comprehensive repertoire of statistical methodologies, encompassing inverse-variance weighting, weighted median, Simple mode, Weighted mode, and MR-Egger regression techniques, was adroitly harnessed for meticulous analysis. The discernment emerged that the genus Coprococcus3 is inversely associated with OP, potentially serving as a deterrent against its onset. Additionally, 21 other gut microbial species exhibited a positive correlation with OP, potentially accentuating its proclivity and progression. Subsequent to rigorous scrutiny via heterogeneity and sensitivity analyses, these findings corroborate the causal nexus between GM and OP. Facilitated by MR, this study successfully elucidates the causal underpinning binding GM and OP, thereby endowing invaluable insights for deeper exploration into the pivotal role of GM in the pathogenesis of OP.


Introduction
Osteoporosis (OP) has emerged as a paramount concern in public health, gravely compromising the skeletal well-being of the global aging populace. [1]It has significantly escalated socio-economic burdens among the elderly worldwide, while simultaneously compromising patient welfare.The escalating prevalence of OP, attributed to societal aging and shifts in dietary habits, underscores its profound detriment. [2]While preventative and therapeutic strategies have developed, [3] a vast majority of patients still lack timely and efficacious diagnosis and treatment.Recent observational studies on OP, through comprehensive genome-wide association studies (GWAS) and meta-analyses, have elucidated the role of the G-proteincoupled receptor, GPR43.This receptor, via the GPR43-βarr2 signaling pathway, suppresses NF-kB activity, [4][5][6] thus modulating bone metabolism.These revelations provide novel perspectives on the aetiology of OP from both genetic and immunological standpoints.
The human gastrointestinal tract harbours myriad symbiotic microbial communities, numbering in the trillions, [7] crucial for maintaining our health.With the relentless progress in high-throughput sequencing technologies, mounting evidence underscores the pivotal role of gut microbiota (GM) in bone metabolic processes. [8]The dynamic fluctuations of these microbial communities are intricately linked to the preservation of bone mass and quality. [9]Certain observational metaanalyses suggest a correlation between these microbial shifts and OP, with notable regional disparities in dominant microbial populations. [10]Research affirms that GM can influence bone metabolism through a multitude of mechanisms.These mechanisms are interwoven with various modifiable factors, such as gut metabolic byproducts, immune functionality, intestinal epithelial barrier functions, nutrient absorption metabolism, estrogens, and endocrinology. [11,12]endelian randomization (MR) is a robust technique for deducing causal relationships, utilizing genetic variations as instrumental variables (IV) to investigate the causal effects of exposures on outcomes.In this context, we employ GM as the exposure and OP as the outcome, utilizing MR analysis to explore the causal nexus between GM and OP.

Study design
The central focus of investigation lies in the 211 gut microbial species under scrutiny, while OP is defined as the outcome of study (see Fig. 1).A preliminary screening was carried out to identify gut microbial species significantly linked with OP, thus paving the way for an extensive MR analysis.This analysis is built upon 3 fundamental assumptions: IVs exhibit an association with the exposure, IVs remain independent from any confounding factors, and IVs impact the outcome solely through the conduit of the exposure pathway. [13]

Data sources
The genetic data pertaining to gut microbial communities is sourced from the latest Genome-Wide Association Study (GWAS) summary data by the MiBioGen consortium, encompassing a cohort of 18,340 individuals across 24 cohorts.We assessed 211 gut microbial communities spanning diverse taxonomic tiers, which include phylum, class, order, family, and genus. [14]This study scrutinized the composition of the GM, based on 3 distinct variable regions (V1-V2, V3-V4, and V4) in 16S rRNA sequencing.Moreover, it employed mapping of microbial quantitative trait loci (mbQTL) to pinpoint genetic variants that influence the relative abundance of microbial taxa. [14]The genetic information for OP comprises 7547 cases and 455,386 controls from the UK Biobank, 7788 cases and 204,665 controls from BioBank Japan, and 3203 cases and 209,575 controls from the FinnGen database.The GWAS summary data is drawn from databases with European ancestry, ensuring minimal likelihood of sample overlap.Details of exposures and outcomes are presented in Table 1.The foundational studies underpinning this data have received endorsement from the pertinent ethical review committees. [15,16]As a result, this study does not necessitate supplementary ethical clearance.

Selection of IV
The criteria guiding the selection of IVs encompass: (1) Preference for Single Nucleotide Polymorphisms (SNPs) associated with each genus at a genome-wide significance level (P < 1.0 × 10 -5 ) [17] ; (2) Utilization of European sample data from the 1000 Genomes Project as a reference panel to compute linkage disequilibrium (LD) among SNPs, retaining only the SNP with the lowest P value among those with R 2 < 0.001 (clustering window size = 10,000 kb); Exclusion of SNPs with a minor allele frequency ≤ 0.01; In cases of palindromic SNPs, deducing the forward strand allele using allele frequency information.This study employs a diverse range of methods, including Inverse-Variance Weighted (IVW), Weighted Median, Simple Mode, Weighted Mode, and MR-Egger regression techniques, to elucidate potential causal relationships between GM and OP.
The IVW method amalgamates meta-analysis techniques with estimates from each SNP to derive an overarching assessment of the impact of GM on OP.Under the absence of horizontal pleiotropy, IVW results remain unbiased. [18]MR-Egger regression operates under the assumption of instrument strength being independent of direct effects, allowing the intercept term to assess the presence of pleiotropy.A zero intercept signifies a lack of horizontal pleiotropy, as depicted in Figure 2. The findings from the MR-Egger regression align with IVW results. [19]he Weighted Median method can accurately estimate causal associations even when up to 50% of IVs are invalid. [20]The Weighted Mode estimation demonstrates superior capability in detecting causal effects, with reduced bias and a lower Type I error rate compared to MR-Egger regression, especially when the direct effects assumption is violated. [20]MR-PRESSO analysis identifies and seeks to mitigate horizontal pleiotropy by removing significant outliers.Cochran IVW Q statistic gauges the heterogeneity of IVs.To identify potential heterogeneous SNPs, a "Leave-one-out" analysis is performed by sequentially excluding each IV SNP. [21]y employing the q-value package for false discovery rate (FDR) correction, the threshold for the error rate q-value is set at 0.1.Significance is attributed to GM and OP when P < .05 and q < 0.1.Conversely, when P < .05 and q ≥ 0.1, the relationship between GM and OP is considered to be suggestively associated. [22]All statistical analyses were carried out using R software version 4.2.1 (R Foundation for Statistical Computing, Vienna, Austria).MR analyses were conducted using the TwosampleMR (version 0.5.6) [23]and MR-PRESSO (version 1.0) [24] R packages.

IV selection and preliminary MR analysis outcomes
Following the application of LD effect and palindromic quality control procedures, a total of 2249 SNPs, as identified by the MiBioGen consortium, emerged as IVs linked with the 211 microbial species DR (P < 1 × 10 −5 ).These microbial communities encompass 9 phyla (102 SNPs), 16 classes (178 SNPs), 20 orders (215 SNPs), 35 families (382 SNPs), and 131 genera (1372 SNPs).By applying a filter based on an IVW P value < .05 and an MR-PRESSO P value > .05,all initial positive results from the MR analysis are elaborated in Figure 2.
These findings imply an elevated risk of OP onset.Comprehensive results are presented in Tables 2, 3, 4, and Figure 3.While not attaining statistical significance after FDR correction (q > 0.1), we posit that these outcomes suggest a potential association, meriting further investigation in forthcoming studies.

Heterogeneity and sensitivity analysis
Within these 23 causal associations, the IVs demonstrated F-statistics ranging from 17.28 to 29.33, effectively mitigating biases arising from weak IVs.Cochran IVW Q-test revealed no significant heterogeneity among these IVs.No indication of heterogeneity emerged in the genetic variations of the GM (refer to the funnel plot in the table).None of the MR-Egger regression intercepts deviated from zero, signifying the absence of directional horizontal pleiotropy influence (all intercept P value > .05).Furthermore, the leave-one-out analysis unveiled no substantial deviations in the causal relationship between GM and OP, indicating that no single genetic variant propels any identified causal association, as depicted in Figure 4.

Discussion
OP, a global health concern, has garnered considerable attention.Despite extensive research efforts over the years, its precise etiology remains elusive.With the recent advent of microbiomics, researchers have shifted their focus toward understanding the intricate interplay between GM and various diseases, including OP. [10,25,26] The intricate relationship between GM and the health and diseases of their hosts has emerged as a focal point of scientific inquiry.Preliminary studies have unveiled the pivotal role of GM in skeletal health. [8,27]These microorganisms not only facilitate nutrient absorption [28] but also engage in profound interactions with the host immune system, [29,30] thereby influencing inflammation and bone metabolism. [31,32]For instance, foundational research has identified that GM can impact bone health through the modulation of inflammatory responses, [33] influence calcium uptake, [31,34] and interference with signaling pathways associated with bone metabolism. [35]n this study, we applied MR analysis, a robust statistical approach, to delve into potential causal relationships.Leveraging publicly available GWAS summary data, we embarked on a comprehensive exploration of the connection between GM and OP.Notably, our findings bolster the perspective of a definitive causal link between GM and OP.Research by Wei et al has identified associations between OP risk and the Bacteroidetes phylum, Bacteroides, Lactobacillus phylum, Escherichia, and Eggerthella genera. [36]The Bacteroidetes phylum comprises various Gram-negative bacteria in the gastrointestinal tract, including the Bacteroides genus, [37] aligning well with our findings.In contrast to these earlier findings, our study uniquely illuminates the significant, inverse association between the genus Coprococcus3 and OP risk.This focus on Coprococcus3 represents a novel contribution, emphasizing its critical role in potentially mitigating OP risk and distinguishing our research from the study conducted by Zeng et al. [38] These revelations not only provide novel insights into the pathogenesis of OP but also pave the way for innovative preventative and therapeutic strategies.For instance, modulating the composition of GM might mitigate OP risk.Furthermore, these insights offer potential new biomarkers for early OP diagnosis and risk assessment.
While this study reveals the causal relationship between GM and OP through MR analysis, it does come with certain limitations.Firstly, although MR can help mitigate the impact of confounding variables, our reliance on summary statistics rather than raw data, along with the lowest taxonomic level in the exposure dataset being the genus, restricts our ability to delve into the causal link at the species level between GM and OP.The SNPs employed in the analysis did not meet the conventional Genome-Wide Association Study (GWAS) significance threshold (P < 5 × 10 −8 ).Secondly, the sample used in this study originates from a specific database, potentially introducing geographical and population biases, thereby necessitating further validation for broader applicability.Lastly, the composition of GM is influenced by a multitude of factors, including dietary habits, lifestyle, and medication, which were not exhaustively accounted for in this research.Our emphasis on the genus Coprococcus3 as a novel insight into the GM role in OP prevention distinctly advances our understanding beyond current literature, setting a foundation for future research to build upon.

Conclusion
In conclusion, study offers novel insights into the role of GM in the pathogenesis of OP.Subsequent research is essential to further elucidate this relationship, delineate the mechanisms through which GM affects skeletal health, and leverage this knowledge for the prevention and treatment of OP.

Figure 2 .
Figure 2. Analysis of causal relationships between gut microbiota (GM) and osteoporosis (OP).Panels A, B, and C illustrate the genome-wide association analysis between GM and OP (P < 1 × 10 −5 ).Panels D, E, and F present the MR results of GM communities with a causative link to OP.

Figure 3 .
Figure 3. Scatterplot illustrating the causal relationship between gut microbiota and osteoporosis.Note: The subscripts exhibit scatterplots of taxon-SNP associations (x-axis) versus OP-SNP associations (y-axis), with horizontal and vertical lines representing the 95% confidence intervals for each association.The primary MR analysis was executed using the inverse-variance weighted method, followed by tests employing MR-Egger, weighted median, and other techniques.Lines sloping upwards from left to right denote a positive correlation with OP, suggesting a pathogenic causal influence.Downward sloping lines indicate a protective causal effect.MR = Mendelian randomization, SNP = single nucleotide polymorphism.

Figure 4 .
Figure 4. Leave-one-out analysis of the causal relationship between gut microbiota and osteoporosis.

Table 1
Details of the exposure and outcome.

Table 2
MR analysis results of gut microbiota and osteoporosis risk in the UKB database.

Table 3
MR analysis results of gut microbiota and osteoporosis risk in the Finn database.

Table 4
MR analysis results of gut microbiota and osteoporosis risk in the BBJ database.