The liver and intestine are closely connected through an integrated anatomical, metabolic, and immunological network known as the gut-liver axis. This axis is primarily established through the portal venous circulation, which channels blood from the small and large intestines carrying nutrients, xenobiotics, and microbial products directly to the liver (Pabst et al., 2023). This unique arrangement places the liver as the first organ exposed to both beneficial microbial metabolites such as short chain fatty acids (SCFAs) and potentially harmful substances including bacterial components and unconjugated drug metabolites (Tripathi et al., 2018; Tilg et al., 2022). Functioning as a metabolic and immune barrier, the liver performs essential roles in detoxification, biotransformation, and clearance, thereby preserving systemic homeostasis and maintaining equilibrium between immune tolerance and host defense in response to gut derived signals (Kubes and Jenne, 2018).

The enterohepatic circulation of bile acids (BAs) represents one of the most tightly regulated and essential forms of metabolic cooperation between the host and the gut microbiota (Grüner and Mattner, 2021; Li et al., 2024a). The liver synthesizes primary BAs and conjugates them with taurine or glycine to facilitate lipid digestion. When these conjugated BAs reach the intestine, commensal bacteria expressing bile salt hydrolase enzymes catalyze their deconjugation, producing secondary BAs such as deoxycholic acid and lithocholic acid. These secondary BAs are subsequently reabsorbed and transported back to the liver through the portal circulation (Urdaneta and Casadesús, 2017; Duszka, 2022). This continuous cycle plays a crucial role in maintaining hepatic metabolic balance through the activation of nuclear and membrane receptors, particularly the Farnesoid X receptor (FXR) and the Takeda G-protein coupled receptor 5 (TGR5), which collectively regulate BA synthesis, lipid metabolism, and inflammatory responses within the gut-liver axis (Chiang and Ferrell, 2020; Lin et al., 2025).

The ability of the gut microbiome to shape drug exposure kinetics, often referred to as the second genome governing pharmacokinetics, is central to understanding individual susceptibility to DILI. This regulatory capacity is exerted primarily through two interconnected biochemical pathways involving microbial enzymatic activity and transcriptional modulation of host xenobiotic receptors (Tsunoda et al., 2021; Burke and Li, 2025).

Microbial enzymes, particularly β-glucuronidase, interfere with host detoxification processes. During Phase II metabolism, the liver inactivates drugs or their toxic intermediates by conjugating them with glucuronic acid. When these glucuronide conjugates enter the colon, bacterial β-glucuronidase enzymes cleave the conjugation bond, regenerating the parent compound or an active metabolite that may exert hepatotoxic effects (Gao et al., 2022; Hillege et al., 2024). This reaction facilitates enterohepatic recycling, extends systemic drug exposure, and increases hepatic toxic burden, thereby altering both the effective dose and exposure duration within an individual.

In parallel, microbial metabolites serve as potent ligands for host nuclear receptors that regulate hepatic drug-metabolizing enzymes (Prakash et al., 2015). Indole derivatives produced from bacterial tryptophan metabolism, for instance, activate the aryl hydrocarbon receptor (AhR) and the pregnane X receptor (PXR), which in turn modulate the expression of cytochrome P450 enzymes such as CYP3A4 and efflux transporters such as MDR1 (Sun et al., 2020; Vyhlídalová, 2022; Haduch et al., 2023). Through these transcriptional networks, the microbiome indirectly governs hepatic detoxification capacity and influences overall drug handling. Consequently, personalized assessment of DILI risk requires functional metagenomic profiling of microbial enzymatic potential, including quantification of β-glucuronidase gene abundance, to capture the true microbial contribution to hepatotoxic susceptibility.

When DILI develops, the physiological harmony between the gut and the liver deteriorates, giving way to a pathological state (Huang et al., 2024). Exposure to certain drugs, particularly antibiotics such as amoxicillin-clavulanate, can trigger profound microbial dysbiosis characterized by the depletion of protective commensal species, including SCFA producers, and the overgrowth of opportunistic or pathogenic bacteria (Kesavelu and Jog, 2023). This ecological imbalance compromises intestinal barrier integrity and alters gut metabolic output.

Under healthy conditions, SCFAs are essential and serve as the principal energy source for colonocytes and play a crucial role in preserving the structure and function of tight junctions (Yue et al., 2022). Their depletion weakens mucosal defenses and increases intestinal permeability, a phenomenon often referred to as leaky gut (Camilleri, 2019). The disrupted barrier permits the translocation of microbial components, particularly LPS, a potent pathogen-associated molecular pattern derived from the outer membrane of Gram-negative bacteria such as Proteobacteria. These translocated microbial molecules enter the portal vein and reach the liver, where they activate Kupffer cells through the Toll-like receptor 4 (TLR4) signaling pathway (Sperandeo et al., 2017).

Activation of this pathway initiates the MyD88 and NF-κB cascade, leading to the production and release of proinflammatory cytokines including TNF-α, IL-1β, and IL-6 (Guijarro-Muñoz et al., 2014). The resulting inflammatory storm promotes oxidative stress, mitochondrial injury, and hepatocyte necrosis, which are key events in the progression of DILI (Andrade et al., 2019). In this context, the gut microbiota may contribute to hepatotoxicity, providing inflammatory co-factors such as LPS that lower the hepatic threshold for drug-induced cellular injury (Huang et al., 2024).

Both human clinical observations and preclinical animal studies of DILI reveal consistent alterations in gut microbial composition. These changes include a decline in microbial diversity, reflected by reduced alpha diversity, and by significant restructuring of community composition across subjects, reflected by altered beta diversity (Chu et al., 2023). A recurrent feature of this dysbiosis is a reduction in the Firmicutes-to-Bacteroidetes ratio, accompanied by an overrepresentation of Gram-negative bacteria, particularly members of the phylum Proteobacteria (Chu et al., 2023). This enrichment is of particular pathophysiological importance because Proteobacteria constitute the principal source of endotoxin, or LPS. The elevated presence of LPS producing taxa enhances activation of the TLR4 pathway in the liver, amplifying inflammatory responses and lowering the threshold for hepatocellular injury following drug exposure.

The gut microbiome exerts a dual influence on DILI, functioning both as a mediator of detoxification and as a contributor to hepatotoxicity. This bidirectional role underscores that therapeutic strategies for DILI cannot rely on a uniform approach (Mi et al., 2022; Chu et al., 2023). Precision medicine must instead determine whether intervention should focus on inhibiting harmful microbial activities or restoring beneficial ones that have been disrupted. Defining this balance is essential for effective management, as different pharmacological agents engage distinct microbiome-dependent pathways that either mitigate or exacerbate hepatic injury. The contrasting mechanisms observed with several high-risk drugs exemplify this complexity and highlight the need for tailored microbiome-targeted interventions in DILI prevention and therapy (Table 1).

SCFAs, principally acetate, propionate, and butyrate, are produced through the fermentation of dietary fiber by anaerobic gut bacteria. They play essential roles in regulating host energy metabolism, maintaining immune equilibrium, and preserving redox homeostasis (Li et al., 2024b). A consistent feature of DILI-associated dysbiosis is the marked reduction in SCFA production, reflecting the loss of beneficial microbial functions.

The hepatoprotective effects of SCFAs operate through multiple mechanisms. Butyrate, the primary energy substrate for colonocytes, reinforces intestinal epithelial tight junctions and strengthens the mucosal barrier, thereby limiting the passage of endotoxins into the portal circulation (Yan and Ajuwon, 2017). In addition to their structural role, SCFAs act as bioactive signaling molecules by activating G protein-coupled receptors GPR41 and GPR43 and by inhibiting histone deacetylases (Kim et al., 2013). These pathways influence immune cell differentiation, enhance the generation of regulatory T cells, and suppress proinflammatory signaling cascades. Through these combined effects, SCFAs help to stabilize the gut-liver interface and mitigate immune-mediated injury, underscoring their central importance in protecting against DILI (Verma et al., 2024).

The BA pool is continuously reshaped by the gut microbiota, exerting wide-ranging effects on host metabolic and immune signaling. Microbial bile salt hydrolase activity plays a decisive role in determining the concentration and composition of secondary BAs, which act as potent ligands for two major host receptors, FXR and TGR5 (Chiang and Ferrell, 2020; Song et al., 2025).

FXR is abundantly expressed in both the liver and intestine and serves as a central regulator of BA synthesis. When activated by BAs in the intestine, FXR induces the production of fibroblast growth factor 15 in mice or fibroblast growth factor 19 in humans. This hormone acts on the liver to suppress BA synthesis, thereby preventing the accumulation of toxic intermediates. Microbial transformations that generate secondary BAs capable of antagonizing FXR signaling, including tauro-β-muricholic acid, can disrupt this feedback mechanism, leading to dysregulated BA homeostasis and heightened susceptibility to drug-induced hepatotoxicity (Chiang et al., 2017).

In contrast, TGR5 is expressed on hepatic immune cells, including Kupffer cells, and on enteroendocrine L cells in the gut. Activation of this receptor by BAs promotes anti-inflammatory signaling within the liver and stimulates the release of glucagon like peptide 1, contributing to metabolic stability. Dysbiosis induced alterations in the BA profile weaken these protective signaling pathways and diminish hepatic resilience to toxic drug exposure, reinforcing the pivotal role of the microbiome in maintaining BA mediated liver protection (Chiang et al., 2017; Beaudoin et al., 2023).

Indole derivatives, including indole-3-propionate (IPA), indole-3-acetate, and indole-3-aldehyde, are key metabolites derived from the bacterial metabolism of dietary tryptophan (Hubbard et al., 2015). These compounds function as potent ligands for host nuclear receptors such as AhR and PXR. Activation of these receptors enhances the host’s intrinsic defense capacity by promoting antioxidant responses, strengthening epithelial barrier integrity, and regulating xenobiotic metabolism. Through these coordinated actions, indole derivatives establish a crucial microbiota-driven signaling pathway that protects the liver from oxidative stress and inflammation, thereby reducing susceptibility to drug-induced injury.

The harmful effects of the gut microbiome in DILI arise mainly from the production and translocation of proinflammatory microbial compounds. Among these, LPS is the best characterized toxin that activates hepatic TLR4 signaling and induces acute inflammatory responses (Soares and Pimentel-Nunes, 2010; Chen et al., 2021). High circulating levels of LPS reflect a toxic metabolic state associated with immune activation and hepatic stress. Other microbial metabolites, such as ethanol and p-cresol, can also aggravate oxidative damage and contribute to mitochondrial dysfunction, thereby amplifying hepatic injury (Huang et al., 2024).

These mechanisms may underlie metabotype-dependent hepatotoxicity, which proposes that DILI risk is determined by an individual’s prevailing metabolic environment, referred to as the metabotype (Iruzubieta et al., 2015). A high risk metabotype is defined by low concentrations of protective metabolites such as SCFAs, indole derivatives, and BAs that activate the FXR, together with elevated levels of proinflammatory microbial products such as LPS and increased β-glucuronidase activity. Understanding and quantifying this metabolite profile provides a molecular framework for explaining the wide variability in idiosyncratic DILI susceptibility and offers a path toward individualized prediction of hepatotoxic risk.

Current approaches to diagnosing DILI rely largely on exclusion and conventional biochemical testing, which often identify injury only after hepatic damage has occurred. Incorporating microbial and metabolic information offers a path toward proactive and individualized risk prediction (Segovia-Zafra et al., 2021).

Fecal and serum metabolite profiling provides a dynamic and noninvasive window into the metabolic communication between the gut and the liver. Measuring circulating levels of key protective metabolites such as IPA and indole-3-acetate, together with harmful compounds including secondary BAs and LPS, enables a quantifiable assessment of hepatotoxic risk (Zhang et al., 2022). In parallel, functional characterization of the gut microbiome through evaluation of gene abundance for critical enzymes such as microbial β-glucuronidase and bile salt hydrolase variants can help estimate an individual’s capacity for toxic drug reactivation or impaired detoxification (Feng et al., 2020).

Integrative multi-omics analysis is essential for capturing the full complexity of host-microbe interactions (Wang et al., 2019). Combining metagenomic, metabolomic, and transcriptomic data with clinical parameters provides the multidimensional information needed to construct robust predictive models (Sanches et al., 2024). By linking microbial functional genes with metabolite outputs and corresponding host inflammatory or detoxification gene expression patterns, researchers can identify microbial metabolic pathways that define individual susceptibility. Such integrative profiling enables advanced risk stratification, allowing clinicians to predict the likelihood of idiosyncratic DILI before drug administration and to move from reactive diagnosis toward preventive precision medicine.

The dual role of the gut microbiome in DILI, functioning both as a facilitator of detoxification and as a source of proinflammatory or toxic factors, defines two principal therapeutic directions. The first involves ecological restoration, aiming to reestablish a balanced microbial community and recover lost protective functions. The second focuses on targeted inhibition, seeking to suppress specific microbial activities that contribute to hepatotoxicity (Andrade et al., 2019). Together, these complementary strategies form the foundation for microbiome-based precision therapy designed to restore gut–liver homeostasis and reduce the risk of DILI.

DILI exhibits marked heterogeneity arising from the interplay of environmental, genetic, and microbial factors (García-Cortés et al., 2023). The clinical presentation and severity of DILI vary widely depending on the specific drug, patient characteristics, and underlying health status. Accounting for host variables such as metabolic disorders, including nonalcoholic fatty liver disease, or genetic predispositions, such as particular human leukocyte antigen alleles, is essential for accurately defining an individual’s susceptibility threshold. The additional influence of drug-drug interactions further complicates prediction, as concomitant therapies, including chemotherapeutic regimens in immune checkpoint inhibitor–associated DILI, can profoundly alter gut microbial composition and metabolic output, thereby modifying hepatotoxic risk.

Equally important is the recognition of racial and ethnic variability in DILI incidence and vulnerability (Lisboa et al., 2020). Baseline differences in microbiome structure, genetic background, and environmental exposures contribute to population-level disparities that remain underrepresented in existing research. Many retrospective studies fail to include sufficiently diverse cohorts, limiting the generalizability of their conclusions. Comprehensive, multi-center investigations incorporating diverse populations are therefore necessary to capture these variations and to identify groups at elevated risk, such as those with higher prevalence of metabolic disorders or distinct genetic profiles, ensuring that future DILI risk models reflect global population diversity and clinical reality.

A major limitation of current research on the gut microbiome in DILI is that most evidence remains observational and correlational (Huang et al., 2024). Although dysbiosis frequently accompanies hepatic injury, it has not been definitively established as a causal factor. Addressing this gap requires rigorous experimental validation to distinguish microbial association from mechanistic contribution.

The application of gnotobiotic animal models, in which germ-free hosts are colonized with defined microbial communities, provides a powerful tool to test causality. Similarly, targeted manipulation through selective antibiotic depletion or microbial reconstitution enables researchers to determine whether the presence or absence of specific microbial taxa or functional genes directly modifies DILI susceptibility. By demonstrating that the controlled introduction or elimination of a single microbial component can alter hepatic response to drugs, these models offer conclusive evidence for microbial causation and represent a critical step toward translating microbiome research into predictive and therapeutic applications.

Reproducing the complex multi-organ interactions that underlie DILI requires experimental systems capable of integrating both hepatic and intestinal physiology. Advanced in vitro and ex vivo platforms are emerging as indispensable tools for mechanistic validation and for evaluating individual patient susceptibility.

Human liver organoids derived from patient-specific stem cells enable high-throughput screening of intrinsic hepatotoxic potential while capturing interindividual genetic diversity (Han et al., 2024). These models provide valuable insight into direct drug cytotoxicity but remain limited in representing the systemic interactions that characterize DILI. Because DILI often results from gut-derived inflammatory mediators such as LPS, single-organ liver cultures cannot replicate the essential transport and activation processes that drive injury.

Microfluidic gut-liver chips, also known as organ-on-a-chip systems, are closing this gap (Palasantzas et al., 2023). These platforms physically and functionally connect an intestinal epithelial barrier, which can be colonized with commensal microbiota, to a hepatic module containing metabolically active hepatocytes and Kupffer cells. By allowing controlled transport of microbial metabolites and endotoxins between the gut and liver compartments, these chips faithfully reproduce the dynamic crosstalk responsible for DILI pathogenesis. Such integrative systems provide a predictive preclinical model for evaluating novel therapeutics and hold promise for personalized toxicological assessment that accounts for an individual’s unique microbial and metabolic profile.

The integration of microbiome, metabolome, and host data creates the foundation for developing clinically useful models that can stratify the risk of DILI. By combining pretreatment metagenomic and metabolomic profiles with genetic and clinical information, researchers can build predictive algorithms that estimate an individual’s likelihood of developing hepatotoxicity. This approach transforms DILI management from a reactive process into one centered on prevention and personalized assessment.

Such predictive systems can support clinical decision making by enabling proactive microbial interventions, including targeted enzyme inhibition or the administration of specific probiotics and prebiotics, and by guiding individualized drug dose adjustments based on predicted susceptibility (Yan et al., 2022). These developments mark a transition toward precision medicine in drug safety, where therapeutic strategies are tailored to each patient’s genetic, microbial, and metabolic characteristics to minimize the risk of hepatic injury.