Potential link between Gordonibacter pamelaeae and phospholipase C gamma 1 in osteoarthritis

Abstract

The gut–joint axis has been increasingly implicated in osteoarthritis (OA), yet the causal contribution of specific gut microbes and their downstream molecular mechanisms remain unclear. We employed a multi-stage, two-sample Mendelian randomization (MR) framework. This approach utilized summary-level data from large-scale genome-wide association studies (GWAS) for gut microbiota, plasma proteins, and OA. The analysis involved three steps: (1) identifying gut microbial taxa with potential causal associations with OA risk; (2) screening of pyroptosis-related proteins using pQTL data; and (3) performing mediation analysis to evaluate potential intermediate mechanisms. We also performed transcriptomic analysis of human cartilage and in vitro experiments in chondrocytes to support the biological relevance of our findings. Our MR analysis identified Gordonibacter pamelaeae as a suggestive microbial taxon inversely associated with OA risk. Among 11 candidate pyroptosis-related proteins, only Phospholipase C Gamma 1 (PLCG1) showed a significant inverse association with OA. MR analysis further suggested that G. pamelaeae was positively associated with genetically predicted PLCG1 levels. Mediation analysis indicated that PLCG1 partially mediated the association between G. pamelaeae and OA. Consistent with these findings, PLCG1 mRNA levels were reduced in human OA cartilage. Furthermore, our in vitro experiments demonstrated that PLCG1 knockdown enhanced IL-1β-induced inflammatory and pyroptotic responses in chondrocytes. This study suggests a potential link involving gut microbiota and OA through PLCG1-related signaling. PLCG1 may act as a context-dependent regulator that limits excessive inflammatory responses under stress conditions. These findings refine the current understanding of the gut–joint axis and may help identify potential targets for OA intervention.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13568-026-02087-5.

Introduction

Osteoarthritis (OA) is a highly prevalent joint disorder characterized by progressive cartilage degeneration, chronic pain, and joint dysfunction, placing a substantial burden on patients and healthcare systems (Kloppenburg et al. 2025; Tang et al. 2025). Although mechanical loading contributes to disease onset, OA is increasingly recognized as a complex condition involving metabolic imbalance and persistent low-grade inflammation (Terkawi et al. 2022; Peng et al. 2025). Current therapies mainly provide symptomatic relief, and effective disease-modifying treatments are still lacking. This limitation highlights the need to identify upstream regulatory factors involved in OA development.

In recent years, the role of the gut microbiota in musculoskeletal disorders has gained growing interest (Biver et al. 2019). Emerging evidence supports a gut–joint axis, in which alterations in microbial composition may influence joint homeostasis through systemic pathways (Deng et al. 2024; Yang et al. 2025; He et al. 2025). Observational studies have shown distinct gut microbiota profiles in people with OA compared with healthy individuals (Boer et al. 2019; Favazzo et al. 2020; Wang et al. 2023). Nevertheless, such associations are susceptible to confounding by diet, lifestyle, and comorbidities, which complicates efforts to determine causal links between specific microbial taxa and OA.

At the cellular level, dysregulated chondrocyte death contributes directly to cartilage degeneration. Among the different forms of programmed cell death, pyroptosis has emerged as an important contributor to OA pathogenesis (Chen et al. 2024). The hallmarks of this process include inflammasome activation, caspase-1 cleavage, gasdermin D (GSDMD) pore formation, and the release of pro-inflammatory cytokines like interleukin-1β (IL-1β) and IL-18 (Shi et al. 2015; Broz and Dixit 2016). Accumulating evidence indicates that pyroptosis occurs in chondrocytes and synovial cells in OA, amplifying local inflammation and accelerating cartilage degradation (Liu et al. 2023; Lin et al. 2025). However, the upstream regulators of pyroptosis in chondrocytes remain poorly defined. In particular, it is unclear whether systemic factors such as gut microbiota contribute to the regulation of pyroptotic signaling.

Mendelian randomization (MR) provides an approach to estimate causal relationships. It leverages genetic variants as instrumental variables, thereby reducing confounding and reverse causation (Davey Smith and Hemani 2014; Sanderson et al. 2022). This approach is well suited to examining multi-layered biological pathways involving microbiota, host factors, and disease outcomes.

In this study, we employed a multi-stage MR strategy to explore the potential causal relationship among gut microbiota, pyroptosis-related proteins, and OA risk. We further integrated transcriptomic analysis and in vitro experiments to evaluate the biological relevance of key findings. Our goal was to identify potential mediators linking gut microbiota to OA and to provide new insights into the gut–joint axis.

Materials and methods

Study design

This investigation adopted a multi-stage, two-sample MR approach to systematically investigate potential causal relationships between gut microbiota and OA, and to further evaluate the mediating role of pyroptosis-related proteins. Our study protocol, which follows the STROBE-MR reporting guidelines, is schematically presented in Fig. 1. The analysis proceeded sequentially: (1) We first conducted a two-sample MR to identify specific gut microbiota taxa with potential causal associations with OA risk. (2) Next, we screened for causal links between a curated set of pyroptosis-related proteins and OA to identify potential mediators. (3) Finally, we performed a formal two-step mediation analysis to quantify the extent to which the identified pyroptosis-related protein, PLCG1, may partially mediate the association between the exposure (gut microbiota) and the outcome (OA). In addition, public cartilage transcriptomic data and in vitro functional experiments were used to provide biological validation of the lead candidate.

Fig. 1.

Study design and mediation framework for the gut microbiota–pyroptosis-related protein–osteoarthritis analysis. A A multi-stage Mendelian randomization (MR) framework was applied to investigate the potential causal relationships among gut microbiota, pyroptosis-related proteins, and osteoarthritis (OA). B Path a represents the effect of gut microbiota on pyroptosis-related proteins (βa); path b represents the effect of pyroptosis-related proteins on OA risk (βb); path c represents the total effect of gut microbiota on OA (βc), including the direct effect (c′) and the indirect (mediated) effect through pyroptosis-related proteins (ab)

Data acquisition

Gut microbiota and OA GWAS data

We obtained summary statistics for gut microbiota from genome-wide association studies (GWAS) Catalog (IDs: GCST90032172-GCST90032644). This dataset was determined through shotgun metagenomic sequencing of fecal samples, covering 11 phyla, 19 classes, 24 orders, 62 families, 146 genera, and 209 species of microorganisms. Supplementary Table S1 lists all analyzed gut microbial taxa. Genetic associations for OA were sourced from a GWAS meta-analysis involving 44,190 OA cases and 414,250 healthy controls of European ancestry.

pQTL data for mediators

To obtain genetic instruments for our potential mediators, we utilized protein quantitative trait loci (pQTL) dataset from the deCODE study in a population of 35,559 Icelanders (Ferkingstad et al. 2021). This pQTL dataset was derived from a European-ancestry population, which was broadly consistent with the ancestry background of the OA GWAS dataset. We initially downloaded summary statistics for 4,907 plasma proteins. Only proteins with available and valid instrumental variables after quality control were retained for subsequent MR analyses.

Curation of the pyroptosis-related gene set

Based on an extensive literature review, we compiled a candidate list of 33 genes involved in the pyroptosis signaling pathway (Supplementary Table S4A), including GSDMD, CASP1, NLRP3, IL1B, and IL18. We then intersected this gene list with the available pQTL data. This resulted in a final set of 11 pyroptosis-related proteins with valid genetic instruments (Supplementary Table S4B), thus enabling subsequent MR analysis.

Statistical analysis

Instrumental variables selection

We selected genetic instruments (SNPs) based on the three key MR assumptions. Detailed information on instrumental variables used in MR analyses is provided in Supplementary Tables S2 and S5. We used a significance threshold of P < 5 × 10⁻⁸, which was relaxed to P < 5 × 10⁻⁵ for exposures with few instruments. We performed LD clumping (r² < 0.001, 10,000 kb window) to ensure SNP independence. All instruments had an F-statistic > 10 to avoid weak instrument bias.

MR and sensitivity analyses

The random-effects Inverse Variance Weighted (IVW) method was used as the primary analysis to estimate the causal effects of (a) gut microbiota on OA, (b) pyroptosis proteins on OA, and (c) gut microbiota on pyroptosis proteins. Odds ratios (ORs) and 95% confidence intervals (CIs) were calculated to represent effect sizes. We used MR-Egger regression, the Weighted Median method, and the MR-PRESSO test as sensitivity analyses to assess pleiotropy. Cochran’s Q statistic was used to quantify heterogeneity. Benjamini–Hochberg false discovery rate (FDR) correction was performed based on IVW-derived P values to account for multiple comparisons.

Mediation analysis

A two-step MR mediation framework was used to assess whether PLCG1 mediated the association between gut microbiota and OA. The total effect of a specific gut microbe on OA was denoted as β_total. The effect of the microbe on the mediator (PLCG1) was estimated as β₁, and the effect of the mediator on OA was estimated as β₂. The indirect (mediated) effect was calculated as β₁ × β₂, and the proportion of the total effect mediated by PLCG1 was calculated as (β₁ × β₂) / β_total. The mediation results were interpreted as evidence supporting potential partial mediation within the MR framework. MR-based mediation analysis relies on several assumptions. First, the genetic instruments for both the exposure and the mediator should be valid and strongly associated with their corresponding traits. Second, the instruments should not be associated with confounders of the exposure–mediator, mediator–outcome, or exposure–outcome relationships. Third, the instruments should influence the outcome only through the specified exposure or mediator pathway, and the causal direction should be correctly specified. Therefore, the estimated mediation proportion should be interpreted cautiously, because violations of these assumptions, residual horizontal pleiotropy, weak instruments, sample overlap, or nonlinear and interactive effects may bias the mediation estimate.

Bioinformatic validation in human cartilage tissue

To determine the clinical relevance of our lead candidate gene, PLCG1, we analyzed its transcriptomic expression in human joint cartilage. We utilized the RNA-sequencing dataset GSE57218 from the GEO database. This dataset contains paired samples of OA-affected and preserved (macroscopically normal) cartilage from 33 patients. The significance level was set at P < 0.05.

In vitro validation in human chondrocytes

Cell culture

The immortalized human chondrocyte cell line IM-H488 (IMMOCELL, Xiamen, China) was cultured in Dulbecco’s Modified Eagle Medium/F-12 (DMEM/F-12) supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37 °C in a 5% CO₂ incubator. Cells were passaged at approximately 80–90% confluence. To maintain phenotypic stability, all experiments were performed using cells prior to passage 3.

Cell transfection

Small interfering RNA targeting human PLCG1 (siPLCG1) and a non-targeting negative control siRNA were purchased from RiboBio (Guangzhou, China). Chondrocytes were seeded in 6-well plates and transfected at approximately 60–70% confluence using riboFECT™ CP Transfection Reagent (RiboBio, Guangzhou, China). The final concentration of siRNA was 50 nM. Samples were collected 24–48 h later for subsequent analyses. The effectiveness of the gene silencing was assessed by Quantitative Real-Time PCR (qRT-PCR) and Western blot. Only experiments with effective PLCG1 reduction (> 50% decrease) were included in downstream analyses.

Inflammatory stimulation and experimental grouping

We established an in vitro OA model by treating cells with 10 ng/mL of IL-1β for 24 h. For combined treatment experiments, chondrocytes were first transfected with siRNA for 24 h before IL-1β was added. For experiments, cells were divided into four groups. The Control group received negative control siRNA. The IL-1β group received negative control siRNA followed by IL-1β stimulation. The siPLCG1 group was transfected with PLCG1-targeting siRNA alone. The IL-1β + siPLCG1 group was transfected with PLCG1-targeting siRNA and then stimulated with IL-1β.

RNA extraction and qRT-PCR

We extracted total RNA from chondrocytes with TRIzol reagent following the supplier’s protocol. A reverse transcription kit was then used to synthesize cDNA. We performed qRT-PCR using a SYBR Green master mix (Takara, Shiga, Japan) on a real-time PCR system. The relative expression of target genes (PLCG1, NLRP3, CASP1, IL1B, and MMP13) was calculated via the 2^-ΔΔCt method, with β-actin serving as the internal control. All experiments were performed in triplicate. Primer sequences used for qRT-PCR are listed in Supplementary Table S10.

Protein extraction and western blot analysis

RIPA lysis buffer, supplemented with protease inhibitors, was used to extract total cellular protein. We then determined protein concentrations using a BCA protein assay kit (Solarbio, Beijing, China). Equal amounts of protein from each sample were separated by SDS-PAGE and subsequently transferred to PVDF membranes. Following blocking, the membranes were incubated overnight at 4 °C with primary antibodies against PLCG1, NLRP3, cleaved caspase-1 (p20), GSDMD-N and β-actin. After incubation with HRP-conjugated secondary antibodies, we visualized protein bands using an enhanced chemiluminescence system. Band strengths were measured using ImageJ software. Target protein expression was normalized to β-actin. A detailed list of primary antibodies is provided in Supplementary Table S11.

Statistical analysis for in vitro experiments

All in vitro experiments were independently repeated a minimum of three times. Results are shown as the mean ± standard deviation (SD). We used one-way analysis of variance (ANOVA) with Tukey’s post-hoc test for comparisons between groups. The significance level was set at P < 0.05.

Overall statistical analysis

We performed all MR and bioinformatic analyses using R software. The TwoSampleMR package was used for MR analyses, and limma was used for transcriptomic differential expression analysis. All statistical tests were two-sided unless otherwise specified.

Results

Causal associations between gut microbiota and osteoarthritis risk

To establish the causal foundation of our study, we first performed a two-sample MR analysis to assess the effects of gut microbiota taxa on the risk of OA. The primary analysis using the IVW method identified a diverse set of 26 taxa that showed potential causal associations with OA (all P < 0.05). Among these, 9 taxa, such as Victivallis sp002998355 (OR = 1.342, 95% CI: 1.118–1.612, P = 0.002), were positively associated with OA risk. Conversely, 17 taxa were inversely associated with OA risk, including Gordonibacter pamelaeae (OR = 0.918, 95% CI: 0.853–0.988, P = 0.022) (Fig. 2). Notably, Gordonibacter pamelaeae showed a consistent inverse association with OA risk, suggesting a potential protective association for further investigation. After FDR correction across the screened gut microbial taxa, the association involving Gordonibacter pamelaeae did not reach the FDR significance threshold and was therefore interpreted as suggestive evidence. Nevertheless, this taxon remained nominally significant in the IVW analysis and showed an inverse association with OA risk. These results provide genetic evidence supporting potential causal associations between specific gut microbes and OA. Sensitivity analyses and complementary MR approaches produced consistent estimates, with no indication of heterogeneity or horizontal pleiotropy (Supplementary Figures S1–S3; Supplementary Table S6B), supporting the robustness of the observed associations. Complete FDR-adjusted results are provided in Supplementary Table S6A, and detailed estimates for the 26 taxa identified in the primary IVW screening are summarized in Supplementary Table S6B.

Fig. 2.

Causal effects of gut microbiota on osteoarthritis risk estimated by the inverse variance weighted (IVW) method. Forest plot showing the causal associations between gut microbial taxa and OA risk. Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using the IVW method. Taxa with OR > 1 indicate increased OA risk, whereas taxa with OR < 1 indicate decreased risk

PLCG1 identified as a candidate pyroptosis-related factor associated With OA

To explore whether pyroptosis-related proteins mediate the gut microbiota–OA relationship, we conducted an MR screening of 11 key pyroptosis-related proteins against OA. Among these, only Phospholipase C Gamma 1 (PLCG1) showed a significant inverse association with OA risk (OR = 0.894, 95% CI: 0.835–0.958, P = 0.002). No other protein in the curated panel demonstrated a significant association with OA risk (Fig. 3). Importantly, PLCG1 was the only protein that consistently showed a significant association across MR analyses, highlighting its potential as a candidate host molecule linking gut microbiota and OA. For readability, selected MR results are visualized in Fig. 3 and summarized in Supplementary Table S7. PLCG1 was therefore prioritized as the primary candidate mediator for subsequent analyses.

Fig. 3.

Causal effects of pyroptosis-related proteins on osteoarthritis risk estimated by the IVW method. Forest plot illustrating the associations between genetically predicted levels of pyroptosis-related proteins and OA risk. Among the tested proteins, PLCG1 showed a significant inverse association with OA risk, while other proteins were not significantly associated

PLCG1 partially mediates the association between Gordonibacter pamelaeae and OA

Having identified both OA-associated microbiota and a key pyroptosis-related protein, we proceeded to connect these elements. An MR analysis was conducted to evaluate the association between OA-related gut taxa and genetically predicted PLCG1 levels. This revealed that Gordonibacter pamelaeae was positively associated with PLCG1 levels (OR = 1.107, 95% CI: 1.010–1.214, P = 0.031) (Fig. 4), suggesting a potential link between this microbial taxon and PLCG1-related signaling. Similar effect directions were observed across different MR models (Supplementary Table S8).

Fig. 4.

Causal effects of gut microbiota on PLCG1 levels estimated by the IVW method. Forest plot showing MR estimates of associations between gut microbial taxa and genetically predicted PLCG1 levels. Gordonibacter pamelaeae demonstrated a positive association with PLCG1 levels, suggesting a potential link between this microbial taxon and PLCG1-related signaling

To further quantify this relationship, we performed a two-step mediation analysis focusing on the Gordonibacter pamelaeae–PLCG1–OA relationship. The analysis suggested a potential indirect effect through PLCG1. Specifically, PLCG1 was estimated to partially mediate 13.28% of the inverse association between Gordonibacter pamelaeae and OA risk (Table 1). These findings suggest that the association between Gordonibacter pamelaeae and reduced OA risk may be partly mediated through PLCG1-related signaling. Detailed mediation results are provided in Supplementary Table S9.

Table 1.

Mediation analysis of the Gordonibacter pamelaeae–PLCG1–osteoarthritis pathway

Gut Microbiota	Mediator	Outcome	Total effect β_total	Direct effect	Indirect effect	Mediated proportion (%)	Path a β1	Path b β2
Gordonibacter pamelaeae	PLCG1	Osteoarthritis	-0.086	-0.074	-0.011	13.28%	0.102	-0.112

Estimates of the total, direct, and indirect (mediated) effects of Gordonibacter pamelaeae on OA risk. The proportion mediated by PLCG1 was calculated as the ratio of the indirect effect to the total effect

PLCG1 expression is reduced in human OA cartilage

To validate the biological relevance of PLCG1 at the tissue level, we analyzed its expression in human cartilage using the GEO dataset GSE57218. This dataset contains paired OA-affected and normal-appearing cartilage samples from the same patients. The differential expression analysis demonstrated that PLCG1 mRNA levels were significantly reduced in OA-affected cartilage in comparison with the paired preserved cartilage (P < 0.05) (Fig. 5). This reduction in PLCG1 expression in diseased tissue suggests a context-dependent alteration of PLCG1 signaling during OA progression, rather than a simple linear protective effect. This observation provided important biological support for the MR findings while also indicating potential complexity in PLCG1 function.

Fig. 5.

Reduced expression of PLCG1 in osteoarthritis cartilage based on GEO transcriptomic analysis. Differential expression analysis of PLCG1 between OA-affected cartilage and paired preserved cartilage from the GEO dataset (GSE57218). (*P < 0.05)

PLCG1 knockdown enhances IL-1β-induced inflammatory and pyroptosis-related responses in chondrocytes

To provide functional support for the role of PLCG1 in chondrocyte inflammatory responses, we performed in vitro experiments using human chondrocytes (Fig. 6).

Fig. 6.

PLCG1 knockdown enhances IL-1β-induced inflammatory and pyroptosis-related responses in human chondrocytes. Human IM-H488 chondrocytes were transfected with either non-targeting control siRNA (Control and IL-1β groups) or PLCG1-targeting siRNA (siPLCG1 and siPLCG1 + IL-1β groups) for 24 h, followed by stimulation with or without 10 ng/mL IL-1β for another 24 h. A Representative Western blot images showing the protein expression of PLCG1, NLRP3, cleaved Caspase-1 (p20), and GSDMD-N. β-actin served as the loading control. GSDMD activation was evaluated by Western blot detection of GSDMD-N. B–E Semi-quantitative analysis of the protein levels shown in (A). F–J qRT-PCR analysis of the relative mRNA expression of PLCG1, NLRP3, CASP1, IL1B, and MMP13. Data are presented as mean ± SD from three independent experiments. Statistical significance was determined by one-way ANOVA with Tukey’s post hoc test (*P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001)

As shown in Fig. 6A-B, Western blot analysis demonstrated that transfection with PLCG1-targeting siRNA effectively reduced PLCG1 protein levels. This finding was further confirmed at the mRNA level by qRT-PCR (Fig. 6F), indicating successful knockdown efficiency. Stimulation with IL-1β alone significantly increased the protein expression of key pyroptosis-related markers, including NLRP3, cleaved Caspase-1 (p20), and GSDMD-N (Fig. 6A, C–E), indicating activation of pyroptosis-related signaling at the protein level.

Importantly, PLCG1 knockdown markedly enhanced this response. Compared with the IL-1β group, the combined treatment group (siPLCG1 + IL-1β) showed further increased expression of NLRP3, cleaved Caspase-1, and GSDMD-N (Fig. 6A, C–E). Consistent with these findings, qRT-PCR analysis showed that IL-1β stimulation upregulated the mRNA levels of NLRP3, CASP1, IL1B, and MMP13 (Fig. 6G–J). This effect was significantly amplified in PLCG1-silenced cells. Notably, PLCG1 knockdown alone had no significant effect on the basal expression levels of these genes, suggesting that PLCG1 primarily modulates inflammatory responses under stress conditions rather than affecting baseline cellular activity.

Collectively, these findings suggest that PLCG1 deficiency increases the sensitivity of chondrocytes to inflammatory stimulation, leading to enhanced activation of pyroptotic and catabolic pathways.

Discussion

This study provides genetic evidence supporting a potential link between gut microbiota and OA. Using a multi-stage MR framework, we identified Gordonibacter pamelaeae as a suggestive microbial taxon associated with reduced OA risk and highlighted PLCG1 as a candidate mediator within this pathway. By integrating genetic inference with transcriptomic and in vitro validation, our study supports a potential gut microbiota–PLCG1-related link involved in OA.

Several studies, including those by Boer and others, have reported changes in gut microbial composition in OA subjects, but most were observational and could not establish causality (Boer et al. 2019; Favazzo et al. 2020; Wang et al. 2023). Interventional studies using probiotics such as Lactobacillus or Bifidobacterium showed modest symptomatic benefits, yet the specific causal taxa remained unclear (Karim 2025; Zhao et al. 2026). In contrast, our MR analysis reduces confounding from diet, medication, and lifestyle, enabling more robust causal inference (Sanderson et al. 2022). A recent MR study by Huang et al. reported that genus-level Gordonibacter was associated with increased hip OA risk after considering medication use, which differs from our species-level finding for G. pamelaeae (Huang et al. 2024). This discrepancy may reflect differences in taxonomic resolution, OA subtype, microbiome profiling platform, and analytical framework.

The inverse association between G. pamelaeae and OA risk is biologically plausible. Gordonibacter species are anaerobic gut bacteria commonly found in the human colon and are involved in polyphenol metabolism, particularly the conversion of ellagic acid into bioactive urolithins (Yang et al. 2024; Bae et al. 2024; Dong et al. 2025). These microbial metabolites, especially urolithin A, can be absorbed into systemic circulation and have been reported to exert anti-inflammatory and antioxidant effects (Zhao et al. 2023). Previous studies suggest that urolithins may contribute to cartilage protection by improving mitochondrial function and reducing inflammatory signaling (He et al. 2021; D’Amico et al. 2022). Taken together, these findings support a biologically plausible link between G. pamelaeae and reduced OA risk, while further experimental studies are needed to clarify the taxon-specific and joint-site-specific roles of Gordonibacter in OA.

This study also highlights PLCG1 as a potential mediator linking gut microbiota to OA. Earlier studies have mostly focused on classical pyroptosis components, including NLRP3, CASP1, and GSDMD, which are typically upregulated in OA cartilage and promote inflammation (Huang et al. 2021; Tang et al. 2024; Chen et al. 2024). In contrast, PLCG1 has rarely been studied in OA and is not a canonical pyroptosis executor. It is a signaling enzyme that regulates phosphoinositide metabolism, intracellular calcium flux, and downstream kinase pathways (Kadamur and Ross 2013; Zhao et al. 2024). Interestingly, intracellular calcium signaling has been increasingly recognized as an important modulator of inflammasome activation (Murakami et al. 2012; Horng 2014; Zhou et al. 2024), suggesting that PLCG1 may influence pyroptosis through upstream signaling rather than acting as a direct executor. Recent studies have also identified other regulators of chondrocyte pyroptosis, such as FOXQ1 and the IRF1/GBP5 axis, further indicating that pyroptosis in OA may be regulated by multiple upstream signaling networks (Tang et al. 2024; Luo et al. 2024). Within this broader context, PLCG1 may represent an additional modulatory component involved in chondrocyte responses to inflammatory stress. Our MR results indicate that genetically higher PLCG1 levels are associated with reduced OA risk, suggesting that PLCG1 may participate in maintaining inflammatory signaling balance.

Our mediation analysis estimated that PLCG1 partially mediated 13.28% of the association between Gordonibacter pamelaeae and OA. Although this proportion is modest, it reflects the multifactorial nature of the gut–joint axis. Recent studies (e.g., Favazzo et al., Liu et al.) suggest that gut microbiota influence OA via multiple parallel mechanisms, involving the modulation of systemic inflammation, metabolic homeostasis, and immune cell function (Favazzo et al. 2020; Liu et al. 2025; Sun et al. 2025; You et al. 2025). In addition to urolithins derived from ellagic acid metabolism, recent MR-based evidence has identified other metabolites, such as 6-hydroxyindole sulfate and N-formylanthranilic acid, as potential mediators linking gut microbiota and knee OA risk (Wang et al. 2025). PLCG1-mediated signaling therefore represents one component of a broader regulatory network instead of a single dominant pathway. Future studies integrating microbiome, metabolomics, and functional validation are needed to clarify whether PLCG1 acts independently or in concert with microbiota-derived metabolic pathways.

To support the biological relevance of our genetic findings, we examined PLCG1 expression in human cartilage and performed functional experiments. This integrative approach allows for a more coherent interpretation of the relationship between genetic associations and cellular responses. Consistent with our MR results suggesting a protective role, PLCG1 expression was reduced in OA cartilage. This observation suggests that reduced PLCG1 expression may contribute to increased susceptibility to inflammatory stress in chondrocytes. Although PLCG1 has been implicated in immune inflammatory signaling and GSDMD-mediated pyroptotic cell death in immune cells (Kang et al. 2018; Tao et al. 2023), its biological effects may vary depending on cellular context and disease conditions. Accumulating evidence indicates that inflammasome regulation is highly cell-type- and context-dependent (Kelley et al. 2019; Feng et al. 2023; Vande Walle and Lamkanfi 2024). In non-immune cells such as chondrocytes, chronic inflammatory stress may shift PLCG1 toward a role in maintaining intracellular signaling homeostasis (Zhao et al. 2024). In this context, reduced PLCG1 expression may increase chondrocyte sensitivity to inflammatory stimulation.

Our in vitro findings further support this interpretation. PLCG1 knockdown alone did not significantly alter basal expression of pyroptosis-related markers. However, under IL-1β stimulation, PLCG1 deficiency markedly amplified the levels of NLRP3, CASP1, IL-1β, GSDMD-N, and MMP13. Importantly, these experimental observations provide functional support for the MR-derived association, linking genetic evidence with cellular responses under inflammatory conditions. Taken together, these findings suggest that PLCG1 may act as a modulatory factor that influences the sensitivity of chondrocytes to inflammatory stimuli, rather than directly driving pyroptosis. Collectively, our findings from genetic, transcriptomic, and functional analyses support a model of a gut–joint axis in which Gordonibacter pamelaeae is associated with PLCG1-related signaling, which may modulate inflammatory and pyroptosis-related responses (Fig. 7).

Fig. 7.

Proposed gut microbiota–PLCG1-related signaling model in osteoarthritis. Schematic diagram illustrating a potential model in which Gordonibacter pamelaeae is associated with PLCG1-related signaling, which may modulate inflammatory and pyroptosis-related responses in chondrocytes under stress conditions. In the canonical inflammasome branch, inflammatory stress promotes NLRP3 inflammasome activation and caspase-1 activation, leading to GSDMD cleavage and GSDMD-N pore formation. Activated caspase-1 also mediates pro-IL-1β / pro-IL-18 maturation, followed by mature IL-1β / IL-18 release. These inflammatory and pyroptosis-related responses may contribute to cartilage degeneration and OA-related pathological changes. Dashed arrows indicate genetically inferred or not directly validated associations, whereas solid arrows indicate canonical pathways or experimentally supported processes

From a translational perspective, our findings may have several potential applied and biotechnological implications. First, microbiota-based interventions, potentially including probiotic- or prebiotic-related approaches, could represent a promising strategy for OA prevention or management. Strategies targeting urolithin-producing bacteria, such as Gordonibacter pamelaeae, or dietary approaches aimed at enhancing polyphenol metabolism (e.g., ellagic acid from berries and nuts) may promote the production of beneficial metabolites such as urolithins, which have been reported to exert anti-inflammatory effects and may contribute to joint health (D’Amico et al. 2022). Second, specific microbial taxa such as Gordonibacter pamelaeae, along with PLCG1-related signaling pathways, may serve as potential biomarkers for OA susceptibility or progression. Finally, PLCG1-associated signaling pathways may represent a potential therapeutic target, and future strategies aimed at restoring or maintaining PLCG1-related signaling in chondrocytes may help limit excessive inflammatory and pyroptosis-related responses. However, further experimental and clinical studies are required to validate these potential applications.

This study has several limitations. First, our MR analyses were primarily based on GWAS data from European populations, which may limit generalizability of the findings to other populations. Although multiple sensitivity analyses were performed, residual horizontal pleiotropy and potential influences from medication exposure, diet, and metabolic status cannot be completely excluded (Sanderson et al. 2022). Second, given the large number of microbial taxa and proteins screened, multiple-testing burden should be considered when interpreting the MR findings. Therefore, the identified associations should be interpreted together with effect consistency, biological plausibility, and downstream validation. In particular, the association involving Gordonibacter pamelaeae did not survive FDR correction and should be interpreted as suggestive rather than definitive microbiome-wide evidence. Third, while MR provides evidence for potential causal associations, it cannot fully elucidate the underlying biological mechanisms. In addition, the two-step MR mediation framework assumes valid instruments, absence of horizontal pleiotropy, correct causal direction, and no unmeasured confounding across the exposure–mediator–outcome relationships. Thus, the estimated mediation proportion should be interpreted as potential partial mediation within a simplified MR framework rather than definitive evidence of a complete biological pathway. The specific microbial metabolites linking Gordonibacter pamelaeae to PLCG1-related signaling remain to be identified. Finally, although our transcriptomic and in vitro experiments support the biological relevance of PLCG1 in chondrocyte inflammatory and pyroptosis-related responses, further validation using primary human chondrocytes, animal models, and microbiota/metabolite intervention systems is warranted.

In conclusion, this study suggests a potential gut microbiota–PLCG1-related link involved in OA. Genetically predicted Gordonibacter pamelaeae abundance was inversely associated with OA risk, and PLCG1 may partially mediate this association by modulating inflammatory and pyroptosis-related responses in chondrocytes. These findings provide new insights into the gut–joint axis and may inform future studies on microbiota-related biomarkers and therapeutic strategies for OA.

Authors’ contributions

S. L. designed the study, performed the MR and bioinformatic analyses, and wrote the original draft. S. Z. supervised and analyzed the in vitro experiments. C. Z. conceived the study, supervised the project, and reviewed & edited the manuscript. S. L. and S. Z. contributed equally to this work. All authors read and approved the final manuscript.

Funding

This research was funded by the Natural Science Foundation of Guangdong Province (Grant No. 2024A1515011231).

Data availability

The GWAS summary statistics for gut microbiota and OA are publicly available from the GWAS Catalog. The pQTL data are available from the deCODE genetics website. The GEO dataset (GSE57218) is accessible from the NCBI GEO database. All other data generated or analyzed during this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Not applicable. This study was a secondary analysis of existing, publicly available, and anonymized summary-level data. All contributing studies had previously obtained the necessary ethics approvals and participant consent.

Competing interests

The authors declare no competing interests.