The interplay of the gut microbiome, tryptophan metabolites, and uremic toxin generation in chronic kidney disease progression

The kynurenine pathway

Approximately 90–95% of dietary tryptophan that escapes protein synthesis enters the kynurenine pathway,^[6] normally a low-grade hepatic flux driven by tryptophan 2,3-dioxygenase (TDO). In CKD, the dominant rate-limiting enzyme shifts to indoleamine 2,3-dioxygenase 1 (IDO1), an extra-hepatic enzyme strongly induced by interferon-γ and the chronic, low-grade inflammation that accompanies declining GFR.^[18] The kynurenine-to-tryptophan ratio (Kyn: Trp) rises progressively across CKD stages and accelerates sharply once estimated glomerular filtration rate (eGFR) falls below 30 mL/min/1.73 m².^[19] This is more than a stoichiometric shift. Kynurenine itself is an AhR ligand and binds GPR35; its downstream products carry distinct effects.

Kynurenic acid (KYNA), generated by kynurenine aminotransferase, is a non-selective glutamate-receptor antagonist with affinity for both N-methyl-D-aspartate (NMDA) and α7-nicotinic acetylcholine receptors. Plasma concentrations rise several-fold in advanced CKD; while neurally KYNA may be protective, in vascular tissue it associates with endothelial dysfunction and atherogenesis.^[20] Quinolinic acid, by contrast, is an NMDA receptor agonist with potent neurotoxic potential. It accumulates in dialysis patients, correlates with cognitive impairment in ESKD cohorts, and may explain part of the high prevalence of executive dysfunction in this population.^[21] 3-Hydroxykynurenine generates reactive oxygen species through autoxidation, contributes to mitochondrial damage, and has been linked to cardiovascular events in hemodialysis.^[22] The clinical signature of CKD kynurenine perturbation is therefore a syndrome (neurocognitive change, oxidative stress, immune dysregulation) that maps onto the lived experience of advanced CKD better than any single solute.

The indole/serotonin pathway

The bacterial route is the more clinically prominent. Tryptophanase, expressed by Escherichia coli, Bacteroides fragilis, several Clostridium species and others, cleaves tryptophan into indole, pyruvate, and ammonia in the colonic lumen.^[23] Indole crosses the gut epithelium and undergoes hepatic CYP2E1-mediated 2-hydroxylation to 3-hydroxyindole, then SULT1A1-catalysed sulfation to IS. Once formed, IS is more than 90% protein-bound, principally to albumin’s site II (the indole site), and conventional hemodialysis removes it poorly because only the unbound fraction crosses the membrane.^[24,25] Plasma IS rises by an order of magnitude or more between CKD G3 and ESKD, accumulates further between dialysis sessions, and persists despite high-flux and online hemodiafiltration techniques.^[26]

IAA arises through bacterial decarboxylation and oxidation of indole-3-pyruvate; it is also AhR-active, protein-bound, and carries strong epidemiological signals for cardiovascular events in CKD.^[27] IPA, generated by Clostridium sporogenes and other fiber-fermenting commensals, is the most interesting member of the family because it appears largely protective: A high-affinity AhR ligand that paradoxically dampens nuclear factor kappa B (NF-κB) activity, supports gut barrier integrity, and inversely associates with type 2 diabetes incidence in population studies.^[6,28] IPA concentrations are reduced in CKD, mirroring loss of the fiber-fermenting commensals, a striking example of the same receptor system producing opposite outcomes depending on which ligand prevails. Peripheral serotonin, derived from enterochromaffin cells under microbial regulation, accounts for approximately 90% of body 5-HT^[29] and modulates platelet function and bone metabolism. Its CKD relevance is real but indirect, contributing to platelet dysfunction and bone homeostasis without dominating the toxicity narrative.

Cross-pathway interactions and the net uremic tryptophan phenotype

The kynurenine and indole branches do not run in parallel. They compete for the same substrate pool and overlap on the same downstream receptors. The net CKD phenotype is consistent across cohorts: Elevated kynurenine and Kyn: Trp ratio, elevated IS and IAA, depleted IPA, and reduced 5-HT precursor availability.^[30] Diet shapes this balance materially. Fiber-rich, plant-based intake favors IPA generation and tryptophan-sparing fermentation, while Western-style diets (high in animal protein, low in fermentable carbohydrate) push indole and p-cresol production toward the toxic end of the spectrum.^[31] The clinical implication is direct: The same patient with the same eGFR can carry very different uremic toxin burdens depending on how they eat.

AhR signaling as a central mediator

The AhR sits at the apex of much of the tryptophan-derived toxicity, and it is also the source of the field’s deepest conceptual problem. AhR is a cytosolic ligand-activated transcription factor of the bHLH-PAS family, classically described in xenobiotic metabolism for its response to dioxin. Endogenous activation comes from kynurenine, kynurenic acid, IS, IAA, and indole derivatives such as 6-formylindolo[3,2-b] carbazole.^[32] Upon ligand binding, AhR translocates to the nucleus, dimerizes with ARNT, and binds xenobiotic response elements, inducing CYP1A1 and CYP1B1 (the canonical readouts) alongside non-canonical NF-κB cross-talk and transforming growth factor beta 1 (TGF-β1) induction.

Not all AhR ligands produce the same outcome – this is the central conceptual problem of the field. IS, IAA, and kynurenine drive a pathological transcriptional program that engages NF-κB and NOD-, LRR-, and pyrin domain-containing 3 (NLRP3) priming, induces TGF-β1, and produces tubulointerstitial fibrosis, endothelial dysfunction, and immune dysregulation.^[33-36] IPA, by contrast, is a high-affinity AhR ligand whose binding produces a fundamentally different program: Dampened NF-κB activity, preserved gut barrier integrity, and an inverse association with type 2 diabetes incidence in population studies.^[6,28,37] The same receptor therefore mediates both arms of the gut–kidney axis depending on which ligand prevails. This is more than a pharmacological footnote. It is a unifying organizing principle, and it explains why pan-AhR antagonism is unlikely to succeed therapeutically: Blocking the receptor cannot distinguish toxic ligands from protective ones.

In proximal tubular cells, IS activates AhR to drive epithelialto-mesenchymal transition with upregulation of TGF-β1, α-smooth muscle actin, and type I collagen, providing a molecular bridge between gut-derived solutes and tubulointerstitial fibrosis.^[33,34] Gondouin et al., documented in human endothelial cells that indolic uremic solutes induce tissue factor production through an AhR-dependent pathway, with relevance to thrombotic risk,^[35] and recent work has linked AhR activation to NLRP3 inflammasome priming in macrophages.^[36] The receptor is therefore better viewed as a molecular hub than as a single pathway: It integrates dietary, microbial, and host inputs and produces context-dependent outputs ranging from immune tolerance to fibrosis. Any therapeutic strategy that succeeds will need to preserve the protective IPA arm while blocking the toxic IS/IAA/kynurenine arm – a central pharmacological challenge for the next decade.

Gut barrier dysfunction and systemic translocation

CKD impairs intestinal epithelial integrity through several reinforcing mechanisms. Urea hydrolysis to ammonia in the colonic lumen raises pH and damages the mucus layer, while butyrate depletion compromises the bioenergetics of colonocytes that depend on it as primary fuel.^[38] Tight-junction proteins ZO-1, claudin-1, and occludin are reduced in CKD intestinal biopsies and animal models.^[39] The result is increased intestinal permeability, with translocation of bacterial lipopolysaccharide and bacterial DNA into the portal and systemic circulations. McIntyre’s group documented elevated circulating endotoxin across CKD stages, peaking in dialysis patients, with strong correlations to inflammation and mortality.^[40] This endotoxemia engages TLR4, contributes to monocyte activation, and reinforces the kidney’s exposure to a chronic inflammatory milieu, producing a vicious cycle in which dysbiosis drives barrier failure, barrier failure drives systemic inflammation, and systemic inflammation drives further renal injury.

Cardiovascular toxicity

The cardiovascular signal is the most clinically consequential. IS, IAA, and quinolinic acid have been independently associated with cardiovascular events and all-cause mortality in CKD and dialysis cohorts.^[41,42] Mechanistically, IS impairs endothelial nitric oxide synthesis through eNOS uncoupling, drives smooth-muscle cell osteogenic transition through RUNX2 induction (a contributor to vascular calcification), and amplifies platelet hyperactivity.^[43,44] Lekawanvijit et al., demonstrated direct IS-induced cardiac fibrosis in animal models that paralleled uremic cardiomyopathy in humans.^[45]

The strongest human evidence comes from prospective cohorts. In the European prospective CKD cohort (Amiens University Hospital) (n = 139 CKD patients, 25 deaths over 605 ± 217 days of follow-up), Barreto et al., reported an unadjusted hazard ratio (HR) of 17.1 (95% confidence interval [CI] 2.27– 129.1) for all-cause mortality in the highest IS tertile compared with the lowest, and a per-1-standard deviation increment in log-IS HR of 2.10 (95% CI 1.37–3.23) after adjustment for serum albumin, hemoglobin, and phosphate.^[41] The wide Cis reflect the small sample size, but the effect size and direction are robust across all four multivariable models tested. In an independent Taiwanese cohort of 268 CKD patients followed for 21 ± 3 months (35 renal-progression events, 14 deaths), Wu et al., reported HR 1.092/μmol/L (95% CI 1.060–1.126; p < 0.001) for serum p-cresyl sulfate predicting renal progression, and HR 1.099 (95% CI 1.053–1.148; p < 0.001) for all-cause mortality, both effects surviving adjustment for traditional risk factors and baseline eGFR.^[46] Similar associations have been reported for IS and incident heart failure in hemodialysis cohorts.^[47] and for IS and cardiovascular events in chronic hemodialysis.^[41,48] The data remain observational, and residual confounding by GFR and inflammation cannot be excluded; instrumental variable approaches that might disentangle these are still few. By GRADE criteria, the prognostic association between IS or p-cresyl sulfate (PCS) and cardiovascular events is moderate-certainty; the causal step from biomarker to clinical outcome remains low-certainty pending interventional confirmation.

Renal fibrosis and CKD progression

Animal data linking IS to kidney fibrosis are robust. The 5/6 nephrectomy model with IS supplementation accelerates tubulointerstitial fibrosis through AhR-TGF-β1 signaling and concurrent NRF2 suppression with consequent oxidative stress.^[33,49] Bolati’s group showed that IS induces epithelialto-mesenchymal transition in cultured proximal tubular cells through both Smad-dependent and Smad-independent pathways.^[34] The translational gap is human renal histology data: while tubulointerstitial fibrosis is the final common pathway of CKD progression and tryptophan-derived toxins clearly accumulate, definitive linkage in humans remains largely indirect, through correlations between plasma IS concentration and rate of eGFR decline.^[46] We are not aware of a published human biopsy series that has co-localized intrarenal IS concentration with histological fibrosis scores; closing that gap would substantially strengthen the causal inference. Overall certainty of evidence linking tryptophan-derived toxins to human renal fibrosis is low: The animal literature is robust, but human biopsy-level confirmation is absent.

Immune dysregulation and inflammaging

Tryptophan-derived metabolites drive substantial immune perturbation. IS and IAA expand pro-inflammatory CD14+ CD16+ monocytes, alter neutrophil chemotaxis, and accelerate T-cell senescence.^[50] Kynurenine, via AhR, expands regulatory T-cell populations while paradoxically also promoting exhausted T-cell phenotypes.^[51] The combined net effect (chronic immune activation alongside impaired pathogen-specific responses) fits the clinical observation of high infection susceptibility and impaired vaccine responses in dialysis patients.^[52] This inflammaging phenotype connects the gut-tryptophan axis to the elevated cardiovascular and infection mortality of advanced CKD and helps explain why two patients with similar eGFR can carry markedly different prognostic burden.

Dietary interventions

Dietary fiber is the most actionable intervention, supported by the most coherent mechanistic and clinical evidence. Fermentable fibers feed SCFA-producing commensals, lower colonic pH, support mucin production, displace tryptophan-fermenting indole producers, and reduce plasma IS and p-cresyl sulfate.^[56,57] The CRIC and other cohorts have shown inverse associations between dietary fiber intake and CKD progression and mortality.^[58] Plant-dominant diets, including Mediterranean and plant-dominant low-protein diet (PLADO) patterns, produce similar effects: Small randomized controlled trials (RCTs) and prospective cohort studies report reductions in plasma IS and p-cresyl sulfate of roughly 20– 40% compared with usual care.^[59] Implementation requires balancing fiber intake against potassium load (manageable through cooking technique, leaching, and selection of low-potassium plant sources) and ensuring adequate protein in advanced CKD where sarcopenia risk is real. The KDIGO 2024 guideline endorses plant-dominant patterns within individualized nutrition prescriptions.^[55] The traditional low-protein diet remains a separate intervention: Its tryptophan-relevant feature is reduced substrate for indole and kynurenine generation, but the clinical benefit at the level of hard endpoints is inconsistent and unlikely to be driven by tryptophan metabolites alone. The principal clinical practice implications and outstanding research priorities arising from the available evidence are summarized in Boxes 1 and 2, respectively.

A subtler point follows from this evidence base. Most trials of dietary fiber, plant-dominant patterns, and synbiotics in CKD were conducted in European, Australian, or East Asian cohorts whose background diets, microbiome compositions, and CKD etiologies do not generalize universally. Patients eating Mediterranean, South Asian vegetarian, traditional sub-Saharan African, or Middle Eastern patterns carry different baseline microbiome and metabolite profiles, and a fiber-supplementation strategy that achieves a 25% PCS reduction in a Brisbane cohort may behave differently in a Cairo, Karachi, or Lagos setting. Cultural adaptation of dietary advice, validated against local microbiome and metabolite data, is therefore a methodological requirement for translatable practice and not a soft equity gesture. The same logic applies to the AST-120 evidence base, which is dominated by East Asian cohorts and may not transfer to populations with different dietary protein burdens, transit times, and adherence patterns.

Prebiotics and probiotics

The clinical evidence for pre- and probiotics in CKD is moderate but heterogeneous. The SYNERGY trial (n = 37 randomized, 31 completed; predialysis CKD with eGFR 10–30 mL/min/1.73 m²; 6-week crossover with 4-week washout) tested a synbiotic of inulin HP/FOS/GOS plus a 9-strain probiotic versus matched placebo.^[60] The primary outcome of serum IS was not met (mean change −2 μmol/L, 95% CI −5 to 1), but synbiotic therapy significantly reduced p-cresyl sulfate by a mean of 14 μmol/L (95% CI −27 to −2; ~13% reduction). The treatment effect was substantially larger in antibiotic-naive participants (n = 21): PCS reduction −25 μmol/L (95% CI −38 to −12) and IS reduction −5 μmol/L (95% CI −8 to −1). An unexpected 38 mg/24 h increase in albuminuria (95% CI 1–295; p = 0.03) was reported in the synbiotic arm, prompting caution. This albuminuria signal has neither been clearly replicated nor formally refuted in subsequent small synbiotic trials, in part because most have not reported urinary protein as a pre-specified outcome. Subsequent small RCTs and a feasibility extension (SYNERGY II) have produced mixed signals, with poor strain standardization, variable doses and durations, and heterogeneous outcomes complicating interpretation.^[61] The honest summary is that selected synbiotics modestly reduce specific uremic toxin concentrations in early-to-mid CKD; whether this translates into clinical endpoint benefit remains unproven.

Fecal microbiota transplantation

Fecal microbiota transplantation (FMT) has compelling preclinical evidence in CKD models, with reduction in renal injury and uremic toxin burden after transfer of healthy microbiota into nephrectomized mice.^[62] Human data are limited to small case series and pilot studies, with positive but heterogeneous metabolomic signals.^[63] Major safety concerns apply in CKD: Immunosuppressed transplant recipients face heightened risk of donor-derived pathogen transmission, reinforced by reported transmission events with extended-spectrum β-lactamase-producing organisms, while ESKD patients with multiple antibiotic exposures and central venous access carry independent infection risk.^[64] FMT remains firmly investigational in CKD.

Oral sorbents

AST-120 (spherical carbon adsorbent) binds indole and other organic compounds in the gut lumen, reducing absorption and downstream IS formation. The evidence trajectory is instructive. Multiple Japanese and Korean studies and registry data support eGFR slope improvement and reduced cardiovascular events.^[65,66] The pivotal multinational EPPIC-1 and EPPIC-2 trials, reported by Schulman et al., in 2015, randomized 2,035 adults with moderate-to-severe CKD (serum creatinine 2.0–5.0 mg/dL in men, 1.5–5.0 mg/dL in women) to AST-120 9 g/day or placebo, with a primary composite endpoint of dialysis initiation, kidney transplantation, or doubling of serum creatinine; each trial continued until 291 primary endpoints accrued.^[67] The primary endpoint was not met in either trial: 35.6% (AST-120) versus 35.3% (placebo) in EPPIC-1 (HR 1.03, 95% CI 0.84–1.27, p = 0.78) and 34.4% versus 36.8% in EPPIC-2 (HR 0.91, 95% CI 0.74–1.12, p = 0.37). Pooled analysis confirmed no benefit. Median time to primary endpoint in the placebo groups (189.0 and 170.3 weeks for EPPIC-1 and EPPIC-2, respectively) was substantially longer than the 124 weeks assumed for power calculations, so progression was more gradual than expected. Post hoc analyses suggested potential benefit in the US per-protocol subgroup with ≥67% adherence (HR 0.74, 95% CI 0.56–0.97) and in heavily proteinuric patients receiving ACE inhibitors or ARBs (HR 0.74, 95% CI 0.56–0.96), but these are exploratory findings.^[68] The result has been regulatory discordance: AST-120 is approved and reimbursed in Japan, Korea, the Philippines, and Taiwan, but not in the United States or the European Union. The lessons are several: Trial design must account for adherence to high-volume oral sorbents, patient selection by tryptophan toxin burden may be more rational than eGFR alone, and a metabolite-targeted intervention is unlikely to alter outcomes if downstream biology is not also modulated.

AhR antagonism and emerging molecular targets

AhR antagonists (CH-223191, GNF-351, and tissue-selective candidates) have demonstrated efficacy in preclinical CKD models, attenuating renal fibrosis and endothelial dysfunction.^[69] IDO inhibitors, repurposed from oncology where epacadostat and indoximod were developed for tumor immune escape, are mechanistically attractive for the kynurenine pathway but lack human CKD data.^[70] Tryptophanase inhibitors (boronic acid and oxime analogues) are at the early discovery stage.^[71] Realistic timelines for any of these reaching clinical translation in CKD are a decade or more, with ligand-specificity the central pharmacological challenge.

Synthesis: Where the evidence stands

Current evidence supports plant-dominant, fiber-rich dietary patterns as the strongest available intervention, with biomarker improvements and concordant epidemiological signals for hard endpoints (moderate-to-strong evidence). Synbiotics produce modest, surrogate-level improvements in early-to-mid CKD (moderate evidence, conflicting on hard endpoints). AST-120 is region-dependent and best characterized as conflicting evidence: Real biomarker effects, real biological plausibility, but unconvincing primary endpoint trial performance. FMT, AhR antagonism, and IDO inhibition remain investigational with no clinical-readiness signal. Table 2 summarizes the comparison.

Limitations of this review

Three limitations of this synthesis warrant explicit acknowledgement. First, the literature search was restricted to English-language publications, which may underrepresent work from regions where CKD epidemiology and gut microbiome composition differ substantially from European and East Asian cohorts that dominate the available evidence base. Second, much of the cardiovascular and progression evidence cited rests on observational cohorts. Residual confounding by GFR, inflammation, dietary patterns, and concomitant medications is plausible, and only randomized intervention with hard endpoints can fully resolve causality. Third, geographic representation in the cited cohorts is uneven: AST-120 evidence is heavily weighted toward Japan, Korea, and Taiwan; IS-mortality data come predominantly from European and Taiwanese cohorts; FMT pilot evidence is preclinical and small-series. Generalizability to other practice settings, including the Middle East and sub-Saharan Africa, cannot be assumed without local validation.

A fourth limitation is the underrepresentation of several CKD populations in the tryptophan-axis evidence base. Pediatric CKD, pregnancy, and the post-transplant period beyond the 1^st year are particularly thin in prospective data, and the developmental trajectory of the microbiome confounds cross-sectional comparison in children. Cohorts from the Middle East and North Africa, sub-Saharan Africa, South Asia, and Latin America are similarly sparse, despite carrying a disproportionate share of the global CKD burden and likely differing in microbiome composition, dietary protein source, and CKD etiology (diabetic nephropathy in MENA, hypertensive and HIV-related disease in sub-Saharan Africa, Mesoamerican nephropathy in Central America). The biomarker–outcome relationships described here cannot be assumed to generalize without explicit validation in these populations.

Priority research directions

Personalized approaches are plausible but not yet operational. A patient with high baseline IS and a fiber-fermenter– depleted microbiome may respond differently to dietary intervention than a patient with similar eGFR but a different metabolic phenotype. Translating this into clinical practice will require validated, affordable biomarker panels, decision thresholds, and trial evidence that current strata-blind interventions miss benefit in some patients. The field would benefit from a standardized biomarker consortium pooling samples across cohorts and harmonizing assays, analogous to the CKD Biomarkers Consortium for tubular injury markers.

Priority research directions include well-designed RCTs of dietary fiber supplementation with renal and cardiovascular endpoints; head-to-head synbiotic trials with consistent strains and dosing; regulatory-quality FMT trials in selected CKD populations with rigorous safety monitoring; AST-120 trials that select for high-toxin-burden patients and minimize adherence noise; and first-in-human studies of tissue-selective AhR antagonists or IDO inhibitors. Equity considerations matter throughout: dietary interventions need cultural adaptation across diverse settings, and microbiome therapeutics must not become another biomarker that widens nephrology care disparities.

Data availability

Not applicable. This is a narrative review article and no original primary data were generated. All cited evidence is available from the original sources referenced in the bibliography.