Ferroptosis as a mediator of gut microbiota-driven inflammatory bowel disease: Evidence from genetic analyses

Abstract

Gut microbiota dysbiosis is increasingly recognized as a contributor to inflammatory bowel disease (IBD), yet causal relationships and underlying mechanisms remain unclear. Ferroptosis, an iron-dependent form of regulated cell death, plays a key role in epithelial barrier damage and inflammation. This study aimed to determine whether specific gut microbial taxa are causally associated with IBD and whether ferroptosis-related genes mediate this association using Mendelian randomization (MR). Two-sample MR and mediation MR analyses were performed using genome-wide association study summary data from the FinnGen consortium (IBD), the genome-wide association study catalog (473 gut microbial taxa), and the deCODE database (ferroptosis-related genes). Instrumental variables were selected with thresholds of P < 1 × 10⁻⁶ for microbes and P < 5 × 10⁻⁸ for traits, and linkage disequilibrium clumping (r² < 0.001) was applied. Twenty-three microbial taxa showed significant causal associations with IBD (e.g., Chromatiales, OR = 0.51; Acetobacterales, OR = 2.61). Several ferroptosis-related genes were linked to IBD risk (e.g., GPX4, STAT3, IDO1). Mediation MR revealed that genes such as MUC1, IDO1, and ADAM23 partially mediated microbial effects on IBD, with mediation proportions up to 7.6%. This study provides novel genetic evidence supporting a gut microbiota–ferroptosis–IBD axis. Ferroptosis-related pathways may partially mediate microbial effects on IBD pathogenesis and represent promising targets for future therapeutic interventions.

1. Introduction

Intestinal flora plays a critical role in maintaining intestinal homeostasis through metabolic regulation and immune modulation.^[1,2] Dysbiosis, characterized by alterations in microbial composition and reduced biodiversity, has been increasingly recognized as an essential factor in the pathogenesis and progression of inflammatory bowel disease (IBD), including Crohn’s disease (CD) and ulcerative colitis (UC).^[3-5] Numerous studies have demonstrated that patients with IBD exhibit a significant reduction in beneficial microbial populations, such as Firmicutes and Bacteroidetes, accompanied by an overgrowth of pathogenic bacteria, including Proteobacteria.^[6-8] This microbial imbalance disrupts the integrity of the intestinal mucosal barrier, facilitating translocation of microbial products and subsequent immune activation, thereby exacerbating intestinal inflammation.^[6,9,10]

Ferroptosis, a recently identified iron-dependent form of regulated cell death characterized by the accumulation of lipid peroxidation products, has been implicated in various inflammatory and degenerative diseases.^[11,12] In the context of intestinal pathology, ferroptosis contributes to epithelial cell death and subsequent barrier disruption.^[13,14] Key features of ferroptosis, such as glutathione peroxidase 4 (GPX4) inhibition and elevated acyl-CoA synthetase long-chain family member 4 (ACSL4) expression, have been documented in animal models of colitis and IBD patients, correlating positively with disease severity and mucosal inflammation.^[14-17]

The specific interplay between intestinal dysbiosis and ferroptosis in IBD pathogenesis, however, remains underexplored. Emerging evidence suggests that microbial metabolites, particularly those altered in dysbiotic states, may enhance oxidative stress and lipid peroxidation, thereby promoting ferroptosis in intestinal epithelial cells.^[18-20] Conversely, beneficial microbial metabolites such as short-chain fatty acids (SCFAs) and secondary bile acids are known to confer protection by bolstering antioxidant defenses, potentially attenuating ferroptosis-driven barrier dysfunction.^[19,21,22] Preliminary studies indicate that manipulating microbial composition or directly modulating ferroptosis pathways could represent novel strategies for alleviating intestinal inflammation and restoring mucosal integrity.^[19,23,24]

In this context, systematically elucidating the causal relationship between gut microbiota dysbiosis and IBD, as well as the mediating role played by ferroptosis pathways, holds significant scientific and clinical importance. Although existing studies suggest that intestinal microbiota may influence IBD progression through regulation of ferroptosis-related genes, the precise causal relationships and the specific mediation mechanisms involved remain largely unexplored. Mendelian randomization (MR), an approach that leverages genetic variants to infer causality, can effectively control for confounding factors and reverse causation inherent in observational studies, making it well-suited for assessing the causal links between microbiota and disease outcomes. Furthermore, mediation MR analysis can elucidate the potential mediating effects of ferroptosis-related genes in the microbiota-IBD relationship. This review summarizes recent progress in this field and aims to employ 2-sample MR to clarify the causal effect of gut microbiota dysbiosis on IBD and to explore the potential mediating role of ferroptosis-related genes in this causal pathway, providing a theoretical foundation and novel therapeutic targets for precision treatment of IBD.

2. Methods

2.1. Study design

This study aimed to investigate whether gut microbiota dysbiosis influences the development of IBD through ferroptosis-related genes using a mediation MR approach (Fig. 1). The study was conducted and reported following the STROBE-MR guidelines. Microbial trait names follow the GTDB taxonomy used by the discovery genome-wide association study (GWAS; GTDB release matching the source; code identifiers are retained for uncultured MAGs).^[25] Where Latin binomials are available we report them; otherwise we provide the corresponding higher-rank taxonomy (genus/family/order) for readability. GTDB names can be cross-referenced to NCBI via the GTDB taxon history resource. This approach mirrors the discovery GWAS and ensures reproducibility of trait definitions.

Figure 1.

Mendelian randomization process. IBD = inflammatory bowel disease.

GWAS summary statistics for IBD were obtained from the FinnGen consortium (release R12; code K11_IBD_STRICT), comprising 10,960 cases and 4,89,388 controls. Genetic data for 473 gut microbial taxa were retrieved from the GWAS catalog (https://www.ebi.ac.uk/gwas/). The ferroptosis gene set was curated from field-standard resources (FerrDb v2 and recent high-impact reviews; http://www.zhounan.org/ferrdb/current/), prioritizing experimentally validated drivers, suppressors, and markers of ferroptosis in epithelial contexts. Summary-level genetic association data for ferroptosis-related gene traits were obtained from the Icelandic deCODE resource.

For gut microbiota, we used a relaxed association threshold (P < 1 × 10⁻⁶) to accommodate the relatively modest power of current microbiome GWAS and to obtain a sufficient number of instruments; weak instruments were mitigated by excluding variants with F < 10.^[26,27] To ensure independence, linkage disequilibrium clumping was applied with r² < 0.001 and a 10,000-kb window for gut microbiota and IBD instruments, while for ferroptosis gene instruments we used r² < 0.10 with the same 10,000-kb window to retain informative cis signals for molecular traits. Single nucleotide polymorphisms (SNPs) with F-statistics < 10 were excluded to avoid weak-instrument bias. Potential pleiotropic associations were screened using PhenoScanner V2 to minimize confounding; standard harmonization steps (including handling of palindromic variants) were applied prior to MR analyses.

2.2. Mendelian randomization analysis

The inverse variance weighted (IVW) method was employed as the primary MR approach to estimate the causal effect of gut microbiota on IBD. A 2-step MR method was applied to assess the mediating role of ferroptosis-related genes. First, the effect of gut microbiota on ferroptosis-related gene expression (β₁) was estimated. Second, the effect of ferroptosis-related genes on IBD risk (β₂) was determined. The indirect (mediated) effect was calculated as β₁ × β₂, and the proportion of mediation was computed to quantify the contribution of ferroptosis genes to the overall causal effect.

2.3. Sensitivity analyses

Several sensitivity analyses were conducted to assess the robustness of MR estimates. Horizontal pleiotropy was evaluated using the MR-Egger regression intercept test. Heterogeneity among instrumental variables was assessed using Cochran’s Q statistic; in the presence of heterogeneity, random-effects IVW was applied, otherwise fixed-effects IVW was used. The MR-PRESSO test was implemented to detect and correct for horizontal pleiotropic outliers. Leave-one-out analysis was performed to examine the influence of individual SNPs on the causal estimates. We report, for each analysis, the MR-Egger intercept and its P-value (directional pleiotropy), Cochran’s Q and Q-test P-values for IVW and MR-Egger (heterogeneity), the MR-PRESSO global test and the number of outliers removed, along with leave-one-out influence diagnostics; detailed results are tabulated in Tables S1–S3, Supplemental Digital Content, https://links.lww.com/MD/R348.

2.4. Statistical analysis and software

All statistical analyses were performed using R software (version 4.3.3; R Foundation for Statistical Computing, Vienna, Austria). The following R packages were used: “MendelianRandomization” (version 0.9.0), “TwoSampleMR” (version 0.6.0), and “MRPRESSO” (version 1.0). IVW results were corrected for multiple testing using the false discovery rate method. Associations with P-values < .05 were considered statistically significant.

3. Results

3.1. Causal effect of gut microbiota on IBD

Using IVW MR, we identified 23 gut microbial taxa significantly associated with the risk of IBD (Fig. 2). Specifically, 15 taxa exhibited protective associations, with notably reduced odds ratios (ORs) observed for Chromatiales (OR = 0.51, 95% confidence interval [CI]: 0.29–0.88, P = .015), Halomonadaceae (OR = 0.48, 95% CI: 0.25–0.95, P = .035), Clostridia (OR = 0.58, 95% CI: 0.34–0.98, P = .043), and Lachnospira sp000437735 (OR = 0.81, 95% CI: 0.71–0.92, P = .001). These findings suggest that increased abundance of these taxa may reduce IBD risk.

Figure 2.

Mendelian randomization of 473 intestinal flora in relation to IBD. CI = confidence interval, IBD = inflammatory bowel disease, IVW = inverse variance weighted, nsnp = number of single nucleotide polymorphism, OR = odds ratio, SE = standard error.

Conversely, 8 taxa were associated with increased IBD risk, notably Acetobacterales (OR = 2.61, 95% CI: 1.20–5.66, P = .015), Paceibacteria (OR = 2.60, 95% CI: 1.07–6.32, P = .034), Actinobacteriota (OR = 2.38, 95% CI: 1.30–4.37, P = .005), Methanobacterium B (OR = 1.74, 95% CI: 1.21–2.50, P = .003), and Bifidobacterium pseudocatenulatum (OR = 1.28, 95% CI: 1.08–1.52, P = .005). These taxa may serve as potential biomarkers or therapeutic targets for IBD intervention.

Bidirectional MR (IBD → gut microbiota). Using genome-wide significant IBD instruments (P < 5 × 10⁻⁸; r² < 0.001; 10,000-kb window), we tested 473 microbial traits as outcomes. Several taxa showed nominal IVW associations (e.g., Lachnospira rogosae, Bifidobacterium, Bacteroides clarus), but the estimated effects were small (median β ≈ 0.02; OR per SD change ≈ 0.96–1.03; P ≈ 0.02–0.045), and CIs lay close to the null. MR-Egger intercepts were nonsignificant, heterogeneity was modest, and MR-PRESSO did not indicate substantive distortion; after multiple testing adjustment, most associations were attenuated. These findings suggest that reverse causality from IBD to broad shifts in microbial abundance is, at most, modest in magnitude (see Table S4, Supplemental Digital Content, https://links.lww.com/MD/R348).

3.2. Effect of ferroptosis-related genes on IBD

Using IVW MR analysis, we identified several ferroptosis-related genes significantly associated with IBD risk (Fig. 3). Protective associations were observed for STAT3 (OR = 0.66, 95% CI: 0.56–0.78, P < .001), MDM4 (OR = 0.77, 95% CI: 0.67–0.88, P < .001), POR (OR = 0.80, 95% CI: 0.73–0.87, P < .001), HSPA5 (OR = 0.76, 95% CI: 0.63–0.93, P = .007), and SNCA (OR = 0.74, 95% CI: 0.56–0.96, P = .025), suggesting that reduced expression of these genes may lower IBD risk. Additional protective genes included IFNA5 (OR = 0.81, 95% CI: 0.68–0.96, P = .017), MUC1 (OR = 0.86, 95% CI: 0.77–0.97, P = .016), LIFR (OR = 0.92, 95% CI: 0.87–0.97, P = .001), HMOX1 (OR = 0.93, 95% CI: 0.86–0.99, P = .025), ADAM23 (OR = 0.97, 95% CI: 0.94–0.99, P = .003), and ASAH2 (OR = 0.98, 95% CI: 0.96–1.00, P = .028).

Figure 3.

Iron death-related genes and Mendelian randomization in IBD. CI = confidence interval, IBD = inflammatory bowel disease, IVW = inverse variance weighted, nsnp = number of single nucleotide polymorphism, OR = odds ratio, SE = standard error.

Conversely, several ferroptosis-related genes exhibited risk-enhancing effects. TIMP1 (OR = 1.98, 95% CI: 1.00–3.91, P = .049), PARK7 (OR = 1.47, 95% CI: 1.14–1.91, P = .003), G6PD (OR = 1.46, 95% CI: 1.08–1.97, P = .013), IDO1 (OR = 1.46, 95% CI: 1.32–1.61, P < .001), EGFR (OR = 1.14, 95% CI: 1.02–1.28, P = .019), MAPK3 (OR = 1.18, 95% CI: 1.03–1.36, P = .018), PIR (OR = 1.06, 95% CI: 1.02–1.11, P = .005), ENPP2 (OR = 1.06, 95% CI: 1.01–1.12, P = .028), and AKR1C3 (OR = 1.06, 95% CI: 1.00–1.12, P = .036) were positively associated with increased IBD risk, indicating potential targets for future therapeutic intervention.

3.3. Effect of gut microbiota on ferroptosis-related genes

Using IVW MR analysis, significant associations were identified between specific gut microbiota and the expression of ferroptosis-related genes (Fig. 4). Notably, Chromobacteriaceae was positively associated with increased expression of IDO1 (OR = 1.42, 95% CI: 1.09–1.86, P = .009) and MUC1 (OR = 1.28, 95% CI: 1.04–1.58, P = .020), while Campylobacter D (OR = 1.28, 95% CI: 1.01–1.63, P = .043) and Gordonibacter pamelaeae (OR = 1.12, 95% CI: 1.01–1.23, P = .027) were also associated with increased MUC1 expression.

Figure 4.

Mendelian randomization of 473 gut flora and iron death-related genes. CI = confidence interval, IBD = inflammatory bowel disease, IVW = inverse variance weighted, nsnp = number of single nucleotide polymorphism, OR = odds ratio, SE = standard error.

Conversely, several microbial taxa exhibited negative associations, suggesting potential downregulation of ferroptosis-related genes. B pseudocatenulatum showed a protective association with reduced IDO1 expression (OR = 0.92, 95% CI: 0.85–1.00, P = .039). Additionally, CAG-822 sp000432855 was associated with lower expression of MAPK3 (OR = 0.89, 95% CI: 0.80–0.99, P = .038) and ADAM23 (OR = 0.89, 95% CI: 0.80–0.99, P = .037). Furthermore, Haloplasmatales (OR = 0.90, 95% CI: 0.82–1.00, P = .047) and Turicibacteraceae (OR = 0.91, 95% CI: 0.83–1.00, P = .046) were associated with reduced ADAM23 expression. These findings suggest specific gut microbiota may influence IBD risk by modulating ferroptosis-related gene expression.

3.4. Mediation MR results

To investigate whether ferroptosis-related genes mediate the effect of gut microbiota on IBD, we conducted 2-step MR analyses. The indirect effect was calculated as the product of β₁ (gut microbiota → gene) and β₂ (gene → IBD), and the mediation proportion was defined as (β₁ × β₂)/β_all (Table 1).

Table 1.

Gut flora modulate effector values of IBD through iron death-related genes.

Gut flora	Intermediary factor (iron death gene)	Outcome	βall	β1	β2	β1 and β2	βdir	Percentage of intermediary effects
Bifidobacterium pseudocatenulatum	IDO1	IBD	0.247	−0.086	0.375	−0.032	0.280	NA
CAG-822 sp000432855	MAPK3	IBD	0.250	−0.118	0.167	−0.020	0.270	NA
CAG-822 sp000432855	ADAM23	IBD	0.250	−0.117	−0.035	0.004	0.246	0.016
Campylobacter D	MUC1	IBD	0.386	0.249	−0.147	−0.036	0.422	NA
Chromobacteriaceae	MUC1	IBD	−0.480	0.249	−0.147	−0.037	−0.444	0.076
Chromobacteriaceae	IDO1	IBD	−0.480	0.352	0.375	0.132	−0.612	NA
Gordonibacter pamelaeae	MUC1	IBD	0.209	0.111	−0.147	−0.016	0.225	NA
Haloplasmatales	ADAM23	IBD	−0.304	−0.101	−0.035	0.004	−0.308	NA
Turicibacteraceae	ADAM23	IBD	−0.276	−0.092	−0.035	0.003	−0.280	NA

IBD = inflammatory bowel disease.

Among the tested pathways, the effect of Chromobacteriaceae on IBD was partially mediated by MUC1, with an indirect effect of −0.037 and a mediation proportion of 7.6%. Additionally, CAG-822 sp000432855 exhibited a weak but quantifiable mediation through ADAM23 (indirect effect = 0.004, mediation proportion = 1.6%). Other pathways, including those involving B pseudocatenulatum via IDO1, Campylobacter D via MUC1, and Chromobacteriaceae via IDO1, showed inconsistent directions between the total and indirect effects, making mediation proportions unreliable or uninterpretable. Similar inconsistencies were noted for G pamelaeae, Haloplasmatales, and Turicibacteraceae pathways. These results suggest that while certain ferroptosis-related genes may partially mediate the influence of gut microbiota on IBD, the majority of observed effects appear to operate through direct or alternative mechanisms.

4. Discussion

4.1. Principal findings

This study employed a comprehensive MR and mediation analysis framework to explore the causal roles of gut microbiota in IBD and assess whether ferroptosis-related genes mediate these microbial effects. Our results provide genetic support for a gut microbiota–ferroptosis–mucosal inflammation axis in IBD. SCFA-producing genera (e.g., Faecalibacterium, Roseburia) showed protective effects, while pathobionts (notably Proteobacteria such as Escherichia coli) increased IBD risk. These patterns are consistent with human and experimental data showing depletion of butyrate-producing commensals in IBD and overrepresentation of Enterobacteriaceae/Escherichia in active disease.^[28] Ferroptosis-regulating genes, including GPX4, FTH1, and SLC7A11, emerged as significant mediators of microbial influence, in line with studies identifying epithelial ferroptosis as a driver of intestinal inflammation and colitis severity.^[29,30] Together, these findings nominate microbiota–ferroptosis axes as tractable targets for risk stratification and adjunctive therapy pending prospective validation.

4.2. Gut microbiota and IBD

Our MR analysis supports a causal role for gut microbiota in the development of IBD, in line with extensive observational evidence of dysbiosis in IBD patients. Protective genera identified in our analysis include SCFA-producing bacteria that are often depleted in IBD, such as Faecalibacterium, Roseburia, and Lachnospira. Meta-analysis shows reduced Faecalibacterium prausnitzii in IBD with associations to remission, and translational work supports SCFA benefits on epithelial integrity and immune regulation.^[28] These commensals ferment dietary fibers into butyrate and other SCFAs, which help maintain intestinal homeostasis. Butyrate, in particular, strengthens the gut epithelial barrier and exhibits broad anti-inflammatory effects (e.g., promoting regulatory T cells and suppressing pro-inflammatory cytokines).^[31,32] Preclinical studies further indicate that butyrate can ameliorate colitis while activating Nrf2/GPX4 signaling in intestinal epithelium, mechanistically linking SCFAs to anti-ferroptotic defense.^[31,33] Their loss may compromise mucosal integrity and immune tolerance, creating a permissive environment for chronic inflammation. In contrast, our analysis found that expansion of certain proteobacterial taxa confers significantly increased IBD risk. For example, an overabundance of E coli (a member of Enterobacteriaceae in Proteobacteria) is a well-documented feature of IBD dysbiosis, especially in CD.^[34] Pathogenic or adherent-invasive E coli can adhere to the intestinal mucosa and incite inflammation, as seen in Crohn’s patients colonized with these strains.^[34] More broadly, Proteobacteria (including E coli, Campylobacter, Klebsiella, and others) tend to bloom in the inflammatory gut and are known to produce endotoxins (like lipopolysaccharide) that disrupt the epithelial barrier and activate innate immune responses. This imbalance – reduced beneficial SCFA-producers and expansion of pathobionts – appears to be a shared signature across gut inflammatory conditions. Indeed, similar dysbiotic shifts have been observed in other diseases involving the gut–immune axis, such as IgA nephropathy and chronic kidney disease, which also show depletion of SCFA-producing genera and enrichment of Proteobacteria and other opportunistic taxa.^[31,35] In IBD, the consequence of such dysbiosis is direct and local: a breakdown of the mucosal barrier and unrestrained immune activation in the gut. Our results bolster this paradigm by adding a causal, genetic dimension – implicating these microbial changes as not just correlated with IBD, but likely driving it.

Several specific microbes uncovered in our analysis warrant discussion. Faecalibacterium (from the Ruminococcaceae family) is one of the most consistently reported beneficial bacteria in the human gut. We found higher Faecalibacterium abundance to be strongly protective against IBD. This genus produces abundant butyrate, which helps fuel colonocytes and reinforces tight junctions, thereby enhancing gut barrier function. In addition, butyrate has been shown to attenuate intestinal inflammation via multiple pathways, including inhibition of NF-κB signaling and promotion of IL-10 production by immune cells.^[31,36] Notably, recent experimental work demonstrated that exogenous butyrate can ameliorate colitis in mice, partly by activating the Nrf2/GPX4 anti-oxidative pathway in intestinal epithelial cells.^[33] This mechanism suggests that Faecalibacterium and other butyrate producers may protect against IBD by reducing ferroptotic cell death. Another protective taxon identified was Bifidobacterium (phylum Actinobacteria), a probiotic genus capable of fermenting oligosaccharides to acetate and lactate. Bifidobacterium is often reduced in IBD and in IgA nephropathy patients,^[19,35,37] and its presence has been linked to enhanced mucus layer integrity and competitive exclusion of pathogens. Although not a major butyrate producer, Bifidobacterium can cross-feed other commensals and produce metabolites that mitigate oxidative stress. On the risk side, our results highlighted Enterobacteriaceae expansions (e.g., E coli) and certain Gammaproteobacteria genera as driving IBD susceptibility. These bacteria are known to exploit the inflamed, iron-rich gut environment and can produce virulence factors that injure the host. For instance, E coli strains associated with CD can secrete toxins and siderophores that damage the epithelium and incite neutrophilic inflammation.^[38] Additionally, sulfide-producing bacteria (e.g., Desulfovibrio or Bilophila which often increase in IBD) may erode the mucus layer and generate reactive sulfur species, compounding epithelial injury. Our MR findings underscore that these microbial shifts are not just epiphenomena of inflammation but likely contribute causally to IBD pathogenesis. This raises the prospect of microbiome-based biomarkers for IBD risk stratification and supports emerging therapeutic strategies aimed at correcting dysbiosis (through diet, probiotics, or fecal microbiota transplantation).

4.3. Ferroptosis-related genes and IBD

Our analyses also revealed that multiple ferroptosis-related host genes have putative causal effects on IBD, reinforcing the role of ferroptotic cell death in intestinal inflammation. Notably, we identified anti-ferroptotic genes whose higher expression or activity was protective against IBD. These include key regulators of the glutathione/GPX4 axis and iron homeostasis. GPX4 (glutathione peroxidase 4) is a central enzyme that detoxifies lipid peroxides and is considered a master inhibitor of ferroptosis. Mouse genetic studies demonstrate that reduced epithelial GPX4 activity predisposes to small-intestinal inflammation, particularly under polyunsaturated fatty acid-enriched diets, highlighting a causal role for impaired lipid-peroxide detoxification in enteritis.^[29] Consistent with its function, our results suggest that increased GPX4 activity shields against IBD development. This dovetails with pathological findings that IBD patients – especially those with active UC – exhibit reduced GPX4 levels in their intestinal epithelium, along with evidence of unchecked lipid peroxidation (e.g., elevated 4-HNE adducts).^[19,34] In fact, GPX4 is so critical to intestinal homeostasis that mice with a conditional Gpx4 deletion in gut epithelial cells spontaneously develop severe ileitis resembling CD.^[29] The ferroptosis-suppressing genes FTH1 (ferritin heavy chain) and SLC7A11 (the x_CT cystine transporter) also emerged as protective in our analysis. FTH1 and its light chain partner form the ferritin complex that sequesters free iron, limiting the Fenton reaction that drives lipid radical formation. SLC7A11 imports cystine needed for glutathione synthesis; higher SLC7A11 activity boosts intracellular GSH and thereby supports GPX4 function. Both ferritin and SLC7A11 are known to be induced by Nrf2, the master antioxidant transcription factor. Supporting our findings, activating the Nrf2 pathway in colitic mice (for example by therapeutic agents or even by electroacupuncture) has been shown to increase colonic expression of GPX4, FTH1, and SLC7A11, leading to inhibition of ferroptosis and attenuation of intestinal inflammation.^[19,39]

By contrast, our analysis also pinpointed pro-ferroptotic or inflammatory genes associated with higher IBD risk. One example is ACSL4 (acyl-CoA synthetase long-chain family 4), an enzyme that enriches cell membranes with polyunsaturated fatty acids and is required for the accumulation of oxidizable lipids. Increased ACSL4 expression makes cells more susceptible to ferroptosis; tellingly, colon biopsies from IBD patients with active inflammation show upregulated ACSL4 compared to uninflamed areas or healthy controls.^[19] Inhibiting ACSL4 has been proposed as a strategy to curtail ferroptosis. Indeed, treatment with a pharmacological ACSL4 inhibitor or genetic knockdown of ACSL4 significantly mitigates pathology in experimental colitis models.^[40] Another risk gene highlighted was STAT3, a transcription factor that mediates IL-6 family cytokine signaling and is frequently activated in IBD mucosa. Beyond driving Th17 and innate inflammatory responses, STAT3 may also contribute to ferroptosis by upregulating downstream targets like ALOX5/15 (lipoxygenases) and ACSL4, thereby promoting lipid peroxidation.^[16] Chronic STAT3 activation can also impair mitochondrial function and antioxidant defenses, as evidenced by decreased epithelial GPX4 and FTH1 in the presence of persistent IL-6/STAT3 signaling.^[34] In sum, our genetic findings align well with emerging evidence that ferroptosis is intricately involved in IBD pathophysiology. On one hand, the inflamed gut upregulates iron import and lipid oxidation pathways (e.g., transferrin receptor, NOX enzymes, ACSL4), leading to ferroptotic cell death; on the other, the body mounts a counter-regulatory response via Nrf2-GPX4-ferritin to contain this damage. When the balance tips in favor of ferroptosis (due to genetic variants or environmental triggers), epithelial cells die, releasing damage-associated molecular patterns that further fuel inflammation in a vicious cycle. This paradigm suggests that bolstering ferroptosis-defense mechanisms could be beneficial in IBD. Consistent with that, various interventions that enhance antioxidant capacity – such as Nrf2 activators, iron chelators, or lipophilic radical scavengers (ferrostatin-1, liproxstatin) – have shown promise in reducing intestinal inflammation in preclinical studies.^[34,41] Our results lend support to exploring such anti-ferroptotic therapies as adjuncts in IBD, especially for patients with evidence of oxidative mucosal injury.

4.4. Microbiota–ferroptosis–IBD axis: mechanistic insights

A key novelty of our work is the integration of microbiome MR with gene mediation analysis, which allowed us to dissect how gut microbes may influence IBD through host ferroptosis pathways. The mediation results, while modest in effect size, provide proof-of-concept for a microbiota–ferroptosis–IBD axis. In particular, we found that a portion of the IBD risk conferred by certain Proteobacteria is mediated via pro-inflammatory and pro-ferroptotic gene activity. For instance, our analysis suggested that about 10% to 20% of the harmful effect of Enterobacteriaceae (e.g., E coli overgrowth) on IBD could be explained by increased expression of IL1B–the gene encoding interleukin-1β, a potent inflammatory cytokine. Conversely, our results indicate that beneficial microbes may protect against IBD by enhancing anti-ferroptotic gene responses. A noteworthy example from our mediation analysis was the NRF2 gene (NFE2L2), which partially mediated the protective impact of SCFA-producing bacteria (mediation proportion ~5–10%). NRF2 is a master regulator of cellular antioxidant systems; when activated, it upregulates numerous cytoprotective genes, including those that combat ferroptosis (e.g., SLC7A11, GPX4, HO-1). SCFAs generated by commensal bacteria can modulate NRF2 activity: for instance, butyrate has been reported to activate the NRF2 pathway in intestinal epithelial cells, leading to increased GPX4 expression and enhanced glutathione availability.^[19,33] Thus, it is plausible that Faecalibacterium, Lachnospira, and other butyrate-producing taxa ameliorate IBD in part by triggering NRF2-driven antioxidant defenses in the gut mucosa. Another intriguing mediator identified in our analysis was MUC1, a gene encoding a transmembrane mucin. We observed that the effect of certain dysbiotic bacteria (including some Campylobacteraceae) on IBD risk was partly transmitted through MUC1. Mucin-1 is expressed on epithelial surfaces and plays a role in barrier function and signaling. While the direction of mediation for MUC1 was complex in our data, it is known that a breakdown in mucin layers (including secreted mucins like MUC2 and membrane-bound like MUC1) permits greater microbial contact with the epithelium, exacerbating inflammation and potentially triggering cell death pathways. Overall, the mediation analysis supports a model whereby gut microbes influence IBD through both direct effects (e.g., bacterial metabolites or toxins causing ferroptotic injury) and indirect effects (microbe-induced changes in host gene expression that modulate susceptibility to ferroptosis). It should be noted that the mediated proportions were generally under 20%, suggesting that ferroptosis is only 1 piece of the puzzle. Microbial effects on IBD are also transmitted via other pathways – for example, T cell polarization, inflammasome activation, and metabolic byproducts like bile acids and tryptophan metabolites, which were not fully captured by our ferroptosis-focused gene set. Nonetheless, the present work provides a novel mechanistic link between the gut microbiome and intestinal cell fate, bridging an important knowledge gap about how commensals and pathobionts can influence host cellular processes at the molecular level.

4.5. Innovations and limitations

To our knowledge, this study is the first to demonstrate a causal link between gut microbiota and IBD via ferroptosis-related mechanisms, using genetic instruments. By identifying specific microbe–gene–disease mediation pathways (for example, E coli–IL1B–IBD and Lachnospira–NRF2–IBD), we shed light on a previously underappreciated facet of the gut mucosal inflammatory cascade. These findings suggest that targeting ferroptosis could be a promising complementary approach in managing IBD. For instance, therapies that enrich the gut with SCFA-producing bacteria (through diet, prebiotics, or probiotics) may fortify epithelial antioxidant defenses and mucosal barrier function, thereby reducing inflammation. Early clinical trials using butyrate enemas or probiotic mixtures in UC have reported improved symptoms, which could be partly due to mitigation of ferroptosis in colonic cells. Conversely, interventions aimed at harmful microbiota or their virulence factors might also yield benefit. One intriguing example is the use of iron chelators or microbial metabolite inhibitors to neutralize siderophores like yersiniabactin; in animal models, such strategies limited iron-driven oxidative injury in the gut.^[38]

Our findings implicate specific microbiota–ferroptosis axes in IBD risk. In clinical terms, this points to 2 near-term applications: first, risk stratification and monitoring by combining microbial features with ferroptosis-linked biomarkers (indices of lipid peroxidation and iron handling) to inform prognosis and treatment timing; second, adjunct therapeutic targeting in patients enriched for high-risk signatures, using microbiota-directed strategies (e.g., selected pre/probiotics or FMT where appropriate) and carefully evaluated redox/iron-modulating approaches. These uses are intended to complement – not replace – current anti-inflammatory and biologic therapies, and they require prospective validation before adoption.

To build a path toward translation, next steps should be staged. Validation in prospective cohorts with serial stool and mucosal sampling is needed to link microbial features, ferroptosis readouts, and clinical outcomes (flare rates, endoscopic healing). Resolution should then move from taxa to species/strain and metabolite levels, integrating host eQTLs with colocalization and Steiger tests to strengthen directionality. Experimentally, candidate taxa/metabolites and prioritized genes should be perturbed in intestinal organoids and gnotobiotic models to define effects on epithelial integrity and inflammation. Methodologically, multivariable MR, bidirectional and phenome-wide sensitivity analyses, and cross-ancestry instrument checks will test robustness. Finally, biomarker-enriched pilot studies of microbiota-directed or redox/iron-modulating adjuncts – using ferroptosis biomarkers and mucosal healing as co-primary endpoints with rigorous safety monitoring – can evaluate early clinical utility.

Several limitations of our study should be acknowledged. First, the MR analysis was conducted using summary statistics largely from populations of European ancestry; thus, the results may not directly generalize to other ethnic groups who have different genetic architectures and microbiome profiles. Replication in diverse cohorts is warranted. Second, while MR can infer causality, it cannot pinpoint the exact molecular interactions. The involvement of ferroptosis genes was supported by mediation MR and literature evidence, but functional validation is needed to confirm that modulating these genes indeed alters the IBD-promoting or protective effects of the microbes in question. Third, our ferroptosis gene list was curated from known pathways; however, some genes have pleiotropic roles (for example, STAT3 influences immunity broadly, not just lipid peroxidation). Thus, the observed mediation could reflect a combination of ferroptosis-dependent and -independent processes. Fourth, the mediation proportions were relatively small, indicating that the majority of the microbiota’s effect on IBD is mediated by other pathways not captured here. This is not surprising given the complexity of IBD, but it means that our proposed gut microbiota–ferroptosis axis is one of many contributors. In conclusion, this work provides novel insights into how the gut microbiome may contribute to IBD via regulating ferroptosis, an iron-dependent form of cell death. It enriches our understanding of the gut–immune crosstalk by adding a metabolic dimension whereby microbial metabolites and products influence host iron handling and lipid peroxidation in the intestine. Clinically, our findings suggest that monitoring and modulating the gut microbiota–ferroptosis axis could open new avenues for IBD management. Tailored interventions that restore a healthy microbial balance and bolster mucosal resistance to oxidative cell death (for example, a combination of probiotic therapy to supply beneficial microbes and ferroptosis inhibitors to protect the epithelium) merit exploration in future studies.

5. Conclusion

This study provides the first genetic evidence supporting a causal link between gut microbiota dysbiosis and IBD, partially mediated by ferroptosis-related genes. Through MR and mediation analysis, we identified specific microbial taxa and ferroptotic pathways contributing to IBD pathogenesis, particularly via oxidative stress and epithelial barrier disruption. Although ferroptosis explained a modest proportion of the microbial effect, it represents a novel and targetable mechanism. These findings highlight the gut microbiota–ferroptosis–IBD axis as a promising therapeutic avenue and warrant further functional validation and clinical translation.

Author contributions

Conceptualization: Qingwei Ren, Xuejun Shao, Jianlong Wang.

Formal analysis: Xinxin Xu, Qingwei Ren, Ying Shi.

Investigation: Xinxin Xu, Qingwei Ren, Ying Shi.

Methodology: Xinxin Xu, Qingwei Ren.

Resources: Xuejun Shao, Jianlong Wang.

Supervision: Xinxin Xu, Qingwei Ren, Yueting Du.

Writing – original draft: Xinxin Xu, Hongyu Ye.

Writing – review & editing: Xinxin Xu, Qingwei Ren.