Abstract
Background:
Kidney transplantation is the only curative treatment for end-stage kidney disease, providing markedly improved outcomes over dialysis. In search of optimized long-term outcomes, recent studies have demonstrated alterations in the metabolic state of kidney transplant recipients (KTRs) may contribute to chronic graft outcomes. Here, we sought to benchmark metabolic perturbations associated with the post-transplant state using a set of extensively characterized kidney transplant donors as healthy control.
Methods:
In this cross-sectional, single-center study, we used untargeted GC-MS to profile the 24 h urine from 121 stable KTRs (0.2–38.9 yrs post-transplant) and 94 extensively screened healthy potential kidney donors (HC) from the TransplantLines Biobank and Cohort Study. We assessed group differences using linear models adjusted for baseline covariates including eGFR, and the multivariable OPLS-DA model for global separation. ROC analysis was performed on the top discriminants. GlobalTest Pathway enrichment was used to map differentially abundant metabolites in metabolic pathways from the KEGG database.
Results:
In KTRs, 28 metabolites were significantly altered compared to HC; 13 of which were elevated and 15 decreased in KTRs. Among them, methylmalonic acid (AUROC 0.886), citric acid (AUROC 0.870), and glycolic acid (AUROC 0.804) were all reduced in KTRs and showed the strongest in-sample separation between the groups. OPLS-DA achieved good group separation (R2Y = 0.733, Q2 = 0.621) with no signs of overfitting. Twenty pathways were significantly perturbed, led by the TCA cycle, glyoxylate and dicarboxylate metabolism and branched-chain amino acid degradation. Sensitivity analyses examining alternative normalizations, excluding patients with diabetes or proteinuria, adjusting for statin use, and stratifying by eGFR and transplant vintage, were highly concordant with the main analysis (Spearman ρ > 0.87 for all pairwise comparisons).
Conclusion:
Urinary metabolic profiles in KTRs are consistent with alterations in TCA cycle, glyoxylate and dicarboxylate metabolism, and branched chain amino acid degradation after transplantation. Although causality cannot be established from this cross-sectional design, these findings may reflect altered energy metabolism in the post-transplant state.
Introduction
While kidney transplantation is the only definitive treatment for patients with end-stage kidney disease, it does not restore normal physiology (, ). Instead, it sets a new and persistent baseline that is different from a healthy, untransplanted individual. As a consequence, kidney transplant recipients (KTRs) experience an elevated burden of metabolic and cardiovascular complications compared to the general population (, ). Since such systemic changes can threaten long-term allograft health, and late graft loss remains a major challenge in transplantation (–), it is critical to understand the drivers of chronic graft pathology. This requires looking beyond routine clinical measures of kidney function, and assessing more granular, underlying cellular and systemic biology.
Untargeted profiling of small-molecule metabolites offers a method for a wide and relatively unbiased exploration of the post-transplant metabolic landscape (). As the downstream product of genetic, transcriptional, environmental, and pharmacological exposures, the metabolome also provides an integrated and functional snapshot of an individual's physiological state (, ). Urine is an especially relevant biofluid in this context; it can be collected non-invasively, reflects both glomerular filtration and tubular metabolic activity, and contains metabolic signatures shaped by systemic biochemical processes (, , ). This positions it well for detecting chronic, low-grade metabolic shifts associated with the post-transplant state.
Existing applications of urinary metabolomics in transplantation often prioritize differentiating between graft injury states (–). While clinically valuable, interpretation of these injury-specific perturbations is fraught and would be strengthened by a more defined baseline of the post-transplant state. Given the generally modest effect sizes and limited sample sizes in high-throughput experiments, even small phenotypic imprecisions can dilute the biological contrast a study is designed to detect (, ). Therefore, the quality of the healthy reference group is highly relevant. Prospective living kidney donors from the TransplantLines Biobank are a particularly well-suited reference group. These individuals are extensively screened in a lengthy evaluation process before donation, ensuring a well-characterized and clinically healthy reference group. While living kidney donors differ from KTRs in renal function and comorbidity burden, and cannot substitute for matched non-transplant chronic kidney disease (CKD) controls, factors such as reduced kidney function, chronic medication, and cardiometabolic comorbidity can appropriately be accounted for via covariate adjustment and study design.
Here, we leveraged the TransplantLines Biobank to comprehensively profile the urinary metabolome in pre-donation urine from 94 such healthy potential kidney donors alongside that of 121 KTRs using untargeted gas chromatography-mass spectrometry (GC-MS) analyses. We hypothesized that stable KTRs harbor a distinct urinary metabolic signature associated with the post-transplant state and aimed to characterize these metabolic perturbations and establish a reference baseline for future studies investigating specific graft outcomes.
Materials and methods
Study design and patient cohort
This cross-sectional study forms part of the several decades long, ongoing, single-center, prospective TransplantLines Biobank and Cohort Study (clinicaltrials.gov NCT03272841). The study includes (potential) transplant recipients and transplant donors of hearts, lungs, kidneys, livers, and small intestines. From June 2015, all solid organ transplantation patients and transplant donors of the University Medical Center Groningen (UMCG, The Netherlands) were invited to participate. Written informed consent was acquired during enrollment. The overall objective of TransplantLines is to identify and investigate factors related to development of graft failure and to inform patient-centered outcomes in solid organ transplantation. The study is in line with the WMA Declaration of Helsinki and Declaration of Istanbul and received approval from the Medical Ethical Committee of the UMCG (METc 2014/077). Data collection occurred before, during, and at defined intervals after transplantation (that is, at 3, 6, 12, 24, 60 months, and every subsequent five years) for the transplant recipients. For living donors, data were acquired before donation.
The cohort in this study consisted of 121 adult KTRs and 94 living kidney donors, in whom urinary metabolites were measured. All participants underwent clinical and laboratory evaluations pre-donation within the TransplantLines study. Living kidney donors were designated as healthy controls (HC). Included participants visited the UMCG outpatient clinic between September 2015 and November 2017. Trained nursing staff collected data during morning screening visits at the UMCG, including blood, feces, nails, and hair samples. Pooled 24-hour urine samples were collected according to a strict standardized protocol, as previously described (). Participants were instructed to discard the first morning void on the day before the clinic visit, collect all subsequent urine over the next 24 h and include the following morning's first void. Potential inadequacies in 24 h urine collections were defined as a difference of more than 2.5 standard deviations between estimated and measured urine volumes. The estimated urine volume was derived from the formula creatinine clearance = [(urinary creatinine) x 24-hour urine volume]/[serum creatinine], where creatinine clearance was estimated using the Cockcroft-Gault formula (). Participants with potentially inadequate collections were excluded as part of a sensitivity analysis. KTRs were eligible if they were ≥18 years of age. Participants were excluded in the event of known or overt systemic illness, history of drug or alcohol addiction, inability to communicate in Dutch or inability to comprehend study questionnaires. KTRs were considered clinically stable at the time of sampling if confirmed by an outpatient visit without active rejection, acute infection or recurrent disease, as described previously (). The 94 healthy controls were potential kidney donors who had completed the standard UMCG donor screening protocol.
Gas chromatography-mass spectrometry (GC-MS)
For urinary metabolite evaluation, a 1 mL urine aliquot was combined with an internal control solution of 4-phenylbutyric acid (0.5 mg/mL in 0.1M HCl, 100 μL) and ethoxy-amine HCl solution (200 mg/mL in H2O, 100 μL). This mixture was then derivatized in glass containers at 60°C for 30 min. Following the addition of a spatula tip of NaCl and two drops of 37% HCl, the mixture was extracted using a 1:1v/v solution of ethyl-acetate/diethyl-ether. The organic fraction containing the derivatized acids was isolated and evaporated under a vacuum. Next, a 200 μL aliquot of derivatizing reagent consisting of BSTFA, pyridine, and trichloro-methyl-silane (5/1/0.06v/v/v) was introduced for a second derivatization process at 60°C for 30 min.
Prepared samples (1 µL) were injected into the GC-MS. Organic acids were analyzed using a Thermo Scientific Trace 1,310 gas chromatograph coupled to an ISQ LT single quadrupole mass spectrometer (Interscience, Breda, The Netherlands), operated in EI positive ionization mode at 70 eV and registering a total ion scan (m/z range 50–650). Separation was achieved using a (14% cyanopropylphenyl) methylpolysiloxan column (30 m × 0.250 mm × 0.25 µm film thickness; Restek Chromatography, Bellefonte, PA, USA), with helium as the carrier gas (0.8 mL/min).
Data compilation and quantitative evaluations were performed using Chromeleon Version 7.2 SR4 software. Untargeted GC-MS metabolic profiles were pre-processed with PyChromics, a python-based pre-processing tool developed by the UMCG and Rijksuniversiteit Groningen specifically designed for organic acid assessments, to prepare the data for downstream statistical analysis. Concurrently, an enzymatic analysis was conducted on the same samples to quantify creatinine levels using a Roche Clinical Chemistry Analyzer (Cobas, Roche Diagnostics, Mannheim, Germany). Quantitative results were computed by dividing the target compound area by the internal control area. Finally, data were normalized to the corresponding urinary creatinine concentrations to account for variations in urine water content. Metabolite values that fell below the limit of detection (signal area <0.00043) were substituted with random noise values within the sub-detection range. Metabolite identification was based on GC-MS spectral library matching, consistent with putative annotation. No technical replicates were performed. Quality control samples at two concentration levels were analyzed across the measurement period to monitor analytical reproducibility (Supplementary Table S1). Batch effects were assessed by principal component analysis; visual inspection of the scores showed no meaningful batch-related clustering (Supplementary Figure S1).
Statistical and bioinformatics analysis
Baseline characteristics of the study cohort were summarized and compared between KTRs and HC. Depending on data distribution, continuous variables were presented as mean ± standard deviation (SD) or median [Q1–Q3] and compared using a Student's t-test or Mann–Whitney U-test, respectively. Categorical variables were expressed as proportions and compared using the χ2-test or Fisher's exact test as appropriate.
Univariable analyses were performed to identify differentially abundant urinary metabolites between groups. Specifically, we calculated fold changes and false discovery rate (FDR)-adjusted p-values using the Benjamini-Hochberg procedure and visualized them using volcano plots. We evaluated the in-sample discriminatory ability of individual metabolites by calculating the area under the receiver operating characteristic (AUROC) curve; these values reflecting within-sample group separation only, which should not be interpreted as validated diagnostic performance. Then, we constructed linear models to account for potential confounders and to assess metabolite significance before and after adjusting for relevant baseline clinical covariates. Covariates were selected a priori based on clinical relevance. Age, sex and BMI were included as standard demographic potential confounders. eGFR was included because it may affect urinary metabolite excretion and is expected to differ between KTRs and healthy subjects. Other baseline characteristics that differed significantly between groups were additionally included as covariates. Diabetes was included given its known influence on metabolic pathways. Transplant vintage, immunosuppressive regimen and donor type were not included as covariates, as these variables only apply to KTRs and cannot be modeled in a between-group comparison. Additional sensitivity analyses adjusting for statin use, analyzing metabolite excretions and concentrations not normalized for creatinine, excluding patients with diabetes or proteinuria, and stratified analyses by eGFR and transplant vintage within KTRs were performed to assess the robustness of our findings. FDR correction using the Benjamini-Hochberg procedure was applied to both adjusted and unadjusted models. Metabolite values were log2-transformed and scaled prior to analysis; distributions were visually assessed and found to approximate normality.
Next, we applied a supervised multivariable pattern recognition algorithm to maximize separation between KTRs and HC groups. Specifically, the orthogonal projections to latent structures discriminant analysis (OPLS-DA) model was used. OPLS-DA was implemented in R using the ropls package with one predictive component, the number of orthogonal components was determined via 7-fold cross-validation. Model-level significance was assessed by permutation testing with 1,000 permutations. To evaluate feature importance and ranking stability, VIP scores were independently extracted from models trained within a separate 5-fold cross-validation scheme and averaged across folds. A hierarchically clustered heatmap was used to visualize the top discriminative features across samples. Finally, to provide biological context to the observed metabolic shifts, pathway enrichment analysis was applied. We mapped metabolites to known biochemical networks in the KEGG database using the GlobalTest algorithm implemented in MetaboAnalystR. Pathway impact and significance were assessed to identify heavily perturbed networks and to contextualize important discriminant features with established pathways. All statistical analyses and data visualizations were performed using R software (version 4.4.0).
Results
Baseline characteristics of the study population
Our study included a total of 215 individuals, comprising 121 KTRs and 94 HC. The baseline clinical and demographic characteristics are summarized in Table 1. The two groups were comparable across age [60.0 [54.3, 67.0] vs. 61.0 [53.0, 67.0] years, p = 0.81], sex distribution (59.6% vs. 51.2% male, p = 0.28), and BMI (27.1 ± 3.6 vs. 27.3 ± 5.5 kg/m2, p = 0.72). Consistent with the clinical profile of this population, KTRs had a significantly lower renal function compared to HC, evidenced by a lower eGFR (49.6 ± 16.9 vs. 66.5 ± 17.2 mL/min/1.73m2, p < 0.001). Furthermore, KTRs demonstrated higher levels of systemic inflammation [C-reactive protein; 1.3 [0.7, 3.0] vs. 2.2 [1.1, 4.6] mg/L, p = 0.019] and elevated systolic blood pressure (141.7 ± 19.9 vs. 133.4 ± 16.0 mmHg, p = 0.001). The prevalence of comorbid diabetes (30.6% vs. 2.1%, p < 0.001) and the use of statins (52.1% vs. 12.1%, p < 0.001) were also substantially higher in the KTRs cohort. Within the KTRs group, most patients were maintained on standard immunosuppressive regimens, including prednisolone (97.5%), proliferation inhibitors (81.0%), and calcineurin inhibitors (76.0%).
Table 1
| Characteristica | Overall (n = 215) | HC (n = 94) | KTR (n = 121) | P-valueb |
|---|---|---|---|---|
| Male, n (%) | 118 (54.9) | 56 (59.6) | 62 (51.2) | 0.28 |
| Age, yrs | 61.0 [54.0, 67.0] | 60.0 [54.3, 67.0] | 61.0 [53.0, 67.0] | 0.81 |
| BMI, kg/m2 | 27.2 (4.8) | 27.1 (3.6) | 27.3 (5.5) | 0.72 |
| eGFR, mL/min/1.73 m² | 57.0 (19.0) | 66.5 (17.2) | 49.6 (16.9) | <0.001 |
| CRP, mg/L | 1.8 [0.8, 3.9] | 1.3 [0.7, 3.0] | 2.2 [1.1, 4.6] | 0.019 |
| Systolic BP mmHg | 138.1 (18.7) | 133.4 (16.0) | 141.7 (19.9) | 0.001 |
| Diastolic BP, mmHg | 78.9 (10.8) | 77.7 (11.1) | 79.9 (10.5) | 0.14 |
| Post-transplant time, yrs | 5.89 [1.34, 12.14] | – | 5.89 [1.34, 12.14] | – |
| Diabetes, n (%) | 39 (18.1) | 2 (2.1) | 37 (30.6) | <0.001 |
| Proteinuria, >0.5 g/24h | 20 (11.6) | – | 20 (20.4) | – |
| Calcineurin inhibitors, n (%) | 92 (43.4) | – | 92 (76.0) | – |
| Proliferation inhibitors, n (%) | 98 (46.2) | – | 98 (81.0) | – |
| Prednisolone, n (%) | 118 (55.7) | – | 118 (97.5) | – |
Baseline patient characteristics.
Data are presented as mean (SD) for normally distributed continuous variables, median [IQR] for non-normally distributed continuous variables, and n (%) for categorical variables.
p-values compare HC vs. KTR groups using Student's t-test for normally distributed variables, Mann–Whitney U-test for non-normally distributed variables, and Chi-square test (or Fisher's exact test for small cell counts) for categorical variables. A two-sided p-value of <0.05 was considered statistically significant.
Bold values indicate statistical significance p < 0.05.
Urinary metabolome is significantly altered after transplantation
We conducted an initial univariable differential analysis between KTRs and HC to identify the primary metabolic perturbations that distinguish a transplanted individual from a healthy population. This analysis identified 28 metabolites that were significantly altered under an FDR-adjusted p < 0.05 and an absolute fold change (FC) ≥1.5 (Figure 1A). Of those, 13 metabolites were increased and 15 were decreased in KTRs. AUROC curves and boxplots subsequently demonstrated a more detailed view for the most discriminant metabolites (Figures 1B,C). Top hits included methylmalonic acid (mean in-sample AUROC = 0.886), citric acid (mean in-sample AUROC = 0.870), and glycolic acid (mean in-sample AUROC = 0.804), which were all downregulated in KTRs. These values notably require external validation before any inference can be made on clinical utility. In an additional differential analysis using linear regression models, with adjustment for sources of potential confounding, including age, sex, BMI, eGFR, systolic blood pressure, CRP, and diabetes status, we assessed the independence of the metabolites. Although minor changes were observed, the overall shift, and specifically the most prominent hits, remained highly significant and unaffected (Figure 1D, Table 2). Three metabolites (1,2,3-trihydroxybenzene, 2-hydroxy-3-methylbutyric acid, and suberic acid) reached significance only after covariate adjustment, likely consistent with negative confounding.
Figure 1
Table 2
| Metabolite | Status | Unadjusted model | Adjusted model | ||||
|---|---|---|---|---|---|---|---|
| log₂FC | 95% CI | FDR | log₂FC | 95% CI | FDR | ||
| Glutaric acid | Both | 0.814 | (0.553, 1.074) | 0.000 | 0.907 | (0.611, 1.202) | 0.000 |
| Oxoglutaric acid | Both | −0.923 | (−1.182, −0.663) | 0.000 | −0.888 | (−1.189, −0.588) | 0.000 |
| Methylmalonic acid | Both | −1.026 | (−1.284, −0.768) | 0.000 | −0.865 | (−1.159, −0.572) | 0.000 |
| Beta-Tyrosine | Both | 0.730 | (0.468, 0.992) | 0.000 | 0.834 | (0.529, 1.138) | 0.000 |
| Glycolic acid | Both | −0.997 | (−1.255, −0.739) | 0.000 | −0.788 | (−1.073, −0.504) | 0.000 |
| 2-Hydroxybutyric acid | Both | −0.624 | (−0.887, −0.361) | 0.000 | −0.715 | (−1.011, −0.419) | 0.000 |
| 3-Hydroxyisovaleric acid | Both | −0.948 | (−1.207, −0.689) | 0.000 | −0.672 | (−0.958, −0.386) | 0.000 |
| Hippuric acid | Both | −0.761 | (−1.022, −0.499) | 0.000 | −0.686 | (−0.987, −0.384) | 0.000 |
| 2-Methyl−3-hydroxybutyric acid | Both | −0.857 | (−1.117, −0.597) | 0.000 | −0.674 | (−0.974, −0.375) | 0.000 |
| p-Hydroxyhippuric acid | Both | 0.696 | (0.434, 0.958) | 0.000 | 0.676 | (0.369, 0.983) | 0.000 |
| Hydroxypropionic acid | Both | −0.654 | (−0.917, −0.392) | 0.000 | −0.665 | (−0.968, −0.363) | 0.000 |
| Catechol | Both | −0.780 | (−1.041, −0.518) | 0.000 | −0.632 | (−0.934, −0.331) | 0.000 |
| 3-Hydroxysebacic acid | Both | 0.592 | (0.329, 0.855) | 0.000 | 0.624 | (0.319, 0.930) | 0.000 |
| (S)-3-Hydroxyisobutyric acid | Both | −0.557 | (−0.820, −0.293) | 0.000 | −0.557 | (−0.851, −0.262) | 0.002 |
| 3-Hydroxyphenylacetic acid | Both | −0.656 | (−0.919, −0.394) | 0.000 | −0.540 | (−0.843, −0.238) | 0.003 |
| Pyruvic acid | Both | −0.491 | (−0.755, −0.227) | 0.001 | −0.539 | (−0.852, −0.227) | 0.003 |
| 2-Ethylhydracrylic acid | Both | −0.589 | (−0.852, −0.326) | 0.000 | −0.526 | (−0.829, −0.222) | 0.004 |
| Ethylmalonic acid | Both | −0.697 | (−0.959, −0.435) | 0.000 | −0.521 | (−0.821, −0.220) | 0.004 |
| Citric acid | Both | −0.978 | (−1.236, −0.719) | 0.000 | −0.498 | (−0.784, −0.212) | 0.005 |
| 4-Hydroxybenzoic acid | Both | 0.561 | (0.297, 0.824) | 0.000 | 0.485 | (0.181, 0.789) | 0.008 |
| Malic acid | Both | −0.511 | (−0.775, −0.248) | 0.001 | −0.480 | (−0.790, −0.171) | 0.009 |
| 2-Furoylglycine | Both | −0.458 | (−0.722, −0.193) | 0.002 | −0.474 | (−0.785, −0.163) | 0.010 |
| Succinic acid | Both | 0.427 | (0.163, 0.692) | 0.004 | 0.461 | (0.154, 0.768) | 0.012 |
| Urea | Both | −0.317 | (−0.582, −0.052) | 0.038 | −0.413 | (−0.718, −0.107) | 0.027 |
| 4-Hydroxycyclohexylcarboxylic acid | Both | −0.433 | (−0.697, −0.168) | 0.004 | −0.401 | (−0.716, −0.087) | 0.035 |
| 4-Deoxyerythronic acid | Before adj | −0.379 | (−0.644, −0.115) | 0.013 | −0.366 | (−0.676, −0.055) | 0.058 |
| 3-(3-Hydroxyphenyl)propanoic acid | Before adj | 0.349 | (0.084, 0.614) | 0.023 | 0.318 | (0.005, 0.630) | 0.111 |
| 4-Deoxythreonic acid | Before adj | −0.425 | (−0.689, −0.160) | 0.004 | −0.308 | (−0.607, −0.009) | 0.116 |
| 3-Hydroxybenzoic acid | Before adj | 0.376 | (0.112, 0.641) | 0.013 | 0.304 | (−0.009, 0.618) | 0.129 |
| 4-hydroxyvaleric acid | Before adj | −0.366 | (−0.631, −0.102) | 0.016 | −0.293 | (−0.605, 0.020) | 0.146 |
| Salicyluric acid | Before adj | 0.335 | (0.070, 0.600) | 0.030 | 0.291 | (−0.022, 0.605) | 0.146 |
| 5-Hydroxyhexanoic acid | Before adj | −0.524 | (−0.787, −0.260) | 0.000 | −0.269 | (−0.576, 0.037) | 0.187 |
| Ketoleucine | Before adj | −0.330 | (−0.595, −0.065) | 0.032 | −0.262 | (−0.575, 0.050) | 0.201 |
| 2-Methylbutyrylglycine | Before adj | 0.303 | (0.038, 0.568) | 0.045 | 0.263 | (−0.053, 0.579) | 0.201 |
| Phenoxyacetic acid | Before adj | −0.309 | (−0.574, −0.044) | 0.043 | −0.243 | (−0.543, 0.058) | 0.240 |
| Lactic acid | Before adj | 0.486 | (0.222, 0.750) | 0.001 | 0.242 | (−0.064, 0.547) | 0.242 |
| Isocitric acid | Before adj | −0.303 | (−0.568, −0.038) | 0.045 | −0.216 | (−0.531, 0.099) | 0.308 |
| (S)-3,4-Dihydroxybutyric acid | Before adj | −0.446 | (−0.710, −0.182) | 0.003 | −0.203 | (−0.502, 0.096) | 0.316 |
| Aconitic acid | Before adj | −0.544 | (−0.808, −0.281) | 0.000 | −0.198 | (−0.497, 0.100) | 0.325 |
| Acetoacetic acid | Before adj | −0.321 | (−0.586, −0.057) | 0.036 | −0.197 | (−0.512, 0.117) | 0.337 |
| p-Cresol | Before adj | −0.319 | (−0.584, −0.054) | 0.037 | −0.178 | (−0.486, 0.130) | 0.385 |
| Vanillylmandelic acid | Before adj | −0.303 | (−0.568, −0.038) | 0.045 | −0.119 | (−0.427, 0.188) | 0.592 |
| Hydroxyphenyllactic acid | Before adj | 0.383 | (0.119, 0.648) | 0.012 | 0.104 | (−0.194, 0.403) | 0.633 |
| 3-Methylglutaconic acid | Before adj | 0.326 | (0.061, 0.591) | 0.034 | 0.107 | (−0.208, 0.422) | 0.633 |
| 1,2,3-Trihydroxybenzene | After adj | 0.101 | (−0.164, 0.367) | 0.521 | 0.451 | (0.138, 0.765) | 0.015 |
| 2-Hydroxy-3-methylbutyric acid | After adj | −0.282 | (−0.547, −0.017) | 0.063 | −0.435 | (−0.749, −0.121) | 0.020 |
| Suberic acid | After adj | −0.273 | (−0.538, −0.008) | 0.072 | −0.385 | (−0.702, −0.069) | 0.045 |
Linear models.
Bold values indicate statistical significance FDR < 0.05.
Supervised OPLS-DA globally separates kidney transplant recipients and healthy controls
To evaluate systemic metabolic shifts through an integrated analysis of the metabolome, while accounting for inherent multicollinearity of metabolic networks, we trained and evaluated a supervised OPLS-DA model with one predictive component and automatic selection of orthogonal components. The final model retained two orthogonal components. The scores plot derived from this model showed a clear metabolic separation of KTRs and HC (Figure 2A). The corresponding loadings plot shows the contribution of individual metabolites to this separation (Figure 2B). The final model had an R2Y of 0.733 and a Q2 of 0.621 (permutation p = 0.001, Figure 2C). We then examined cross-validated VIP scores for stability across folds to identify the metabolites contributing most strongly to group separation (Figure 2D). Consistent with the independent univariable and multivariable linear models, methylmalonic acid, citric acid, and glycolic acid were the most prominent, though not the sole, markers of this metabolic shift. Finally, a hierarchically clustered heatmap of the top 25 VIP features was generated to visualize their relative abundance across individuals and groups (Figure 2E).
Figure 2
Transplantation significantly disturbs the urinary metabolome
To contextualize the identified metabolic shifts within a broader systems biology framework, we performed pathway enrichment analysis. We used the GlobalTest algorithm to evaluate coordinated differential abundance across metabolite sets mapped to curated KEGG pathways (Figure 3A). The x-axis represents pathway impact, a topological metric reflecting the centrality of detected hits within the pathway network architecture, while the y-axis represents statistical significance after FDR correction.
Figure 3
In total, 20 pathways reached significance at FDR < 0.05 (Table 3, Figure 3A). Among these, the strongest statistical enrichment was observed for the citrate cycle (TCA cycle; 6 hits out of 20 mapped metabolites; FDR = 6.69 × 10−22), alanine, aspartate and glutamate metabolism (4/28; FDR=3.39 × 10−20), and glyoxylate and dicarboxylate metabolism (7/32; FDR =4.38 × 10−20). Other notably enriched pathways included fatty acid degradation (1/39; FDR = 5.98 × 10−14), valine, leucine and isoleucine degradation (4/40; FDR = 1.25 × 10−12), propanoate metabolism (3/22; FDR = 1.72 × 10−9), butanoate metabolism (5/15; FDR = 6.64 × 10−9), and pyruvate metabolism (3/23; FDR = 5.01 × 10−8). It must be noted that pathways supported by a few annotated metabolite hits or negligible topological impact should be interpreted with caution; both metrics are reported for all pathways in Table 3. Next, we visualized the network of pathway interconnectivity, with edges drawn when ≥5 of pathway members were shared (Figure 3B).
Table 3
| Pathway | Total | Hits | P | FDR | Impact |
|---|---|---|---|---|---|
| Citrate cycle (TCA cycle) | 20.000 | 6.000 | 0.000 | 0.000 | 0.317 |
| Alanine, aspartate and glutamate metabolism | 28.000 | 4.000 | 0.000 | 0.000 | 0.048 |
| Glyoxylate and dicarboxylate metabolism | 32.000 | 7.000 | 0.000 | 0.000 | 0.483 |
| Fatty acid degradation | 39.000 | 1.000 | 0.000 | 0.000 | 0.000 |
| Valine, leucine and isoleucine degradation | 40.000 | 4.000 | 0.000 | 0.000 | 0.061 |
| Lipoic acid metabolism | 28.000 | 2.000 | 0.000 | 0.000 | 0.000 |
| Arginine biosynthesis | 14.000 | 2.000 | 0.000 | 0.000 | 0.000 |
| Propanoate metabolism | 22.000 | 3.000 | 0.000 | 0.000 | 0.000 |
| Butanoate metabolism | 15.000 | 5.000 | 0.000 | 0.000 | 0.111 |
| Pyruvate metabolism | 23.000 | 3.000 | 0.000 | 0.000 | 0.220 |
| beta-Alanine metabolism | 21.000 | 1.000 | 0.000 | 0.000 | 0.000 |
| Glycolysis or Gluconeogenesis | 26.000 | 2.000 | 0.000 | 0.000 | 0.098 |
| Tyrosine metabolism | 42.000 | 6.000 | 0.000 | 0.000 | 0.149 |
| Ubiquinone and other terpenoid-quinone biosynthesis | 19.000 | 3.000 | 0.000 | 0.000 | 0.045 |
| Cysteine and methionine metabolism | 33.000 | 1.000 | 0.000 | 0.000 | 0.000 |
| Arginine and proline metabolism | 36.000 | 2.000 | 0.001 | 0.001 | 0.000 |
| Glycine, serine and threonine metabolism | 33.000 | 3.000 | 0.001 | 0.002 | 0.025 |
| Valine, leucine and isoleucine biosynthesis | 8.000 | 1.000 | 0.016 | 0.021 | 0.000 |
| Purine metabolism | 70.000 | 1.000 | 0.021 | 0.026 | 0.000 |
| Phenylalanine, tyrosine and tryptophan biosynthesis | 4.000 | 1.000 | 0.032 | 0.038 | 0.000 |
| Pyrimidine metabolism | 39.000 | 1.000 | 0.429 | 0.490 | 0.047 |
| Tryptophan metabolism | 41.000 | 2.000 | 0.487 | 0.525 | 0.014 |
| Glycerolipid metabolism | 16.000 | 1.000 | 0.525 | 0.525 | 0.093 |
| Pentose phosphate pathway | 23.000 | 1.000 | 0.525 | 0.525 | 0.000 |
Pathway analysis results.
When considering topological impact alongside statistical significance, glyoxylate and dicarboxylate metabolism were the most prominent across all enriched pathways with an impact score of 0.483 (Supplementary Figure S2). The primary contributing metabolite was glycolic acid, the third ranked individual discriminant in both univariable and multivariable analyses. The TCA cycle enrichment was driven predominantly by citric acid and oxoglutaric acid (Supplementary Figure S3), both identified as significant markers in all previous analyses. Methylmalonic acid, the top-ranked discriminant overall, contributed significantly to the enrichment of the BCAA degradation pathway (Supplementary Figure S4). The remaining enriched pathways returned relatively negligible to modest topological impact scores.
Sensitivity analyses
To assess the stability of our findings, we performed multiple sensitivity analyses. We repeated the adjusted differential analysis after excluding KTRs with diabetes (n = 37), after excluding KTRs with proteinuria (n = 20), after excluding potentially inadequate urine collection samples (n = 3), and after additional adjustment for statin use. We additionally analyzed absolute metabolite excretion rates and urinary concentrations instead of creatinine-normalized values, and stratified KTRs by eGFR and transplant vintage (below vs. above median). Across all sensitivity analyses, effect size estimates were highly concordant with the main analysis (Spearman ρ > 0.87 for all pairwise comparisons of log2-fold changes), and the top discriminant metabolites remained consistent (Figure 4).
Figure 4
Discussion
Kidney transplantation alters the metabolic profile of recipients, and a growing body of evidence links the post-transplant state to elevated cardiometabolic risk and chronic graft dysfunction (, ). Much remains to be learned about the nature and the drivers of the metabolic changes associated with kidney transplantation and their downstream effects (, ). In this study, we used an extensively screened cohort of living kidney donors as a healthy reference group to benchmark the urinary metabolome in KTRs. Our findings reveal that the metabolic signature of KTRs and healthy individuals quantifiably differs and the distinctions persist after adjustment for major clinical covariates.
Specifically, we identified 28 metabolites that differed significantly between groups (FDR < 0.05, FC ≥ 1.5), of which 13 were decreased and 15 increased in KTRs. Across the analyses, three metabolites consistently emerged as the strongest discriminants: methylmalonic acid, citric acid, and glycolic acid. All three metabolites showed a reduced concentration in the urine of KTRs. The OPLS-DA model demonstrated a clear metabolic separation between KTRs and healthy controls and had strong performance metrics (R2Y = 0.733, Q2 = 0.621; permutation p = 0.001), indicating a stable and reproducible signal and reinforcing the robustness of the metabolic signature. Pathway enrichment analysis showed that differential metabolites clustered within coherent biochemical routes. The most prominent pathways included the TCA cycle, glyoxylate and dicarboxylate metabolism, and BCAA degradation.
Among the enriched pathways, the TCA cycle showed the strongest statistical enrichment. Pyruvic acid and the TCA cycle intermediates citric acid, 2-oxoglutaric acid, and malic acid were reduced in KTRs. Succinic acid, however, was elevated. Persistence of these changes after adjusting for potential clinical confounders argue against a purely filtration-dependent explanation, though the underlying basis cannot be established based on this observational data alone. Our findings share similarities with those in non-diabetic CKD patients, for which Hallan et al. reported comparable reductions in the urinary excretion of these metabolites, and a decreased expression of ten TCA cycle enzymes in kidney tissue (). Separate studies in KTRs similarly observed lower circulating citric acid and pyruvic acid (). Though speculative, that these patterns persist after transplantation is consistent with the possibility that mitochondrial energy metabolism is not fully restored by engraftment and could reflect long-lasting injury from the transplantation process or ongoing metabolic stress within the graft (, ). It should be acknowledged that the overlap with findings in non-transplant CKD suggests that these TCA cycle perturbations may partly reflect reduced renal mass or chronic kidney injury rather than transplant-specific changes. The notion of overlap between CKD and transplantation warrants the further assessment of these observations in future work to understand the distinctions between the metabolomic profile of renal pathology and the transplant process.
Next, the elevation of succinate stands in contrast to the broader pattern of TCA intermediate reductions and notably diverges from what Hallan et al. found in non-diabetic CKD. In their study, they reported an overall reduction in TCA cycle metabolites, including decreased urinary succinate (). This suggests that an elevation of succinate in our KTRs cohort may reflect a mechanism more specific to the transplantation context. One possible explanation involves ischemia-reperfusion injury (IRI). In experimental models of IRI, succinate dehydrogenase has been shown to operate in reverse during hypoxia, reducing fumarate to succinate. Upon reperfusion, rapid reoxidation of accumulated succinate by complex II has been reported to drive electron leakage and reactive oxygen species generation at complex I. This burst of ROS has been shown to inflict oxidative damage on proteins, lipids, and DNA (–). More broadly, succinate accumulation is thought to reflect an imbalance between oxygen supply and metabolic demand, where its buildup is driven by perturbations in the electron transport chain or other metabolic stress (). Elevated succinate in KTRs may be consistent with ongoing mitochondrial stress beyond the perioperative period, though this interpretation requires direct experimental validation. Notably, succinate has also been reported to function as a signaling molecule through the renal receptor SUCNR1, where it has been implicated in inflammatory and immune responses (, ). Whether elevated succinate in our cohort relates to mitochondrial dysfunction, SUCNR1-mediated inflammatory signaling, or both, cannot be determined from this study.
Perturbations in BCAA catabolism and fatty acid β-oxidation align with the broader pattern of metabolic changes observed in this cohort, which together are consistent with altered metabolic flux. MMA, the most discriminant metabolite in our cohort, was significantly reduced in KTRs, its production stemming from the mitochondrial conversion of propionyl-CoA, linking the degradation of BCAAs and odd-chain fatty acids to the TCA cycle via succinyl-CoA. Reduced concentrations of MMA, 3-hydroxyisovaleric acid and 2-methyl-3-hydroxybutyric acid may also indicate diminished leucine and isoleucine catabolism, potentially reflecting a difference in dietary intake in KTRs (), or changes in mitochondrial oxidative processes. MMA is also produced in part by gut microbial propionate metabolism (–), which raises the additional possibility that post-transplant changes in gut microbial activity may contribute to its reduction. It is important to note, however, that the present cross-sectional design does not allow mechanistic inference.
Several dicarboxylic acids including adipic, glutaric, and 3-hydroxysebacic acids were conversely elevated in KTRs. High urinary levels of these dicarboxylic acids have previously been associated with oxidative stress in membranous nephropathy (), and glutaric acid has been identified as a urinary marker for allograft injury (). Several other microbiome-derived metabolites were differentially abundant between groups, including hippuric acid, catechol, 4-hydroxybenzoic acid and other aromatic and phenolic metabolites. While shifts in gut microbial composition after transplantation have been documented (, ), these differences could equally reflect dietary changes, immunosuppression, or chronic medication use. Any microbiome-related interpretation remains hypothetical in the absence of microbiome data.
Among all enriched pathways, glyoxylate and dicarboxylate metabolism carried the highest topological impact, driven primarily by glycolic acid. Hallan et al. reported similar patterns in the same pathway in CKD patients (), while Dong et al. identified it as the strongest discriminator of graft injury (). Further supporting its clinical relevance, Post Hospers et al. () found that higher pre-transplant plasma glycolic acid levels were associated with a lower risk of delayed graft function, itself a determinant of long-term graft survival. The consistency across cohorts and clinical contexts suggests a potentially meaningful signal that may carry prognostic information in the transplanted kidney.
Overall, the metabolites identified in this study are significantly and consistently altered in stable KTRs compared to healthy donors and may inform non-invasive monitoring strategies. The enrichment of pathways related to the TCA cycle and glyoxylate metabolism identifies metabolic domains that could be targeted for future therapeutic or nutritional interventions. Moreover, the baseline metabolic profile established here provides a reference against which injury-specific or outcome-specific metabolic changes can be compared and interpreted in future targeted metabolomics studies. However, as this is an exploratory cross-sectional study, causality cannot be established and requires future mechanistic experiments.
Several steps would need to be undertaken to advance these findings toward clinical and biological relevance. First, longitudinal studies with pre- and post-transplant sampling within the same individuals would be relevant to understand the temporal relationship between transplantation and the observed metabolic changes. Second, comparison with matched non-transplant CKD controls would help disentangle metabolic changes attributable to reduced kidney function from those specific to the transplant process. Third, investigating the top discriminant metabolites for associations with specific graft injuries or clinical outcomes would clarify their clinical relevance. Fourth, extension to multi-center cohorts would provide insights on the reproducibility and generalizability of these findings. Finally, extending the analytical platform to other biological compartments, including serum or plasma, could provide a more global and comprehensive view of the post-transplant metabolic phenotype.
Limitations
Several limitations merit consideration. The cross-sectional design precludes causal inference and residual confounding despite covariate adjustment cannot be ruled out. Additionally, the urinary metabolome of stable KTRs likely represents an aggregate picture, which is, in part, shaped in part by chronic immunosuppressive therapy. Although this means that individual metabolite changes cannot be attributed solely to the transplant process itself, it does not diminish the value of this benchmark. Mapping the metabolic profile of stable transplant recipients provides a reference that can aid in the interpretation of future stratified studies, especially when metabolic differences between disease states may have modest effect sizes despite having clinically meaningful ramifications. Additionally, chronic immunosuppression is, by definition, part of the clinically realized post-transplant state that this study aims to characterize. Still, overall, the biological interpretations discussed above should be viewed as hypothesis-generating. Moreover, residual non-transplant-specific factors, such as dietary patterns, variation in muscle mass and post-transplant care may also influence urinary metabolite levels independently of transplant-specific biology and should be taken into consideration when interpreting the metabolic signature of KTRs.
Conclusion
In this cross-sectional study, we identify covariate-independent metabolic alterations in the urine of stable kidney transplant recipients compared to healthy living kidney donors, among which the TCA cycle, glyoxylate and dicarboxylate metabolism, and BCAA catabolism are the most prominently perturbed pathways. These findings are consistent with alterations in oxidative metabolism as a feature of the post-transplant metabolic phenotype. Future studies are warranted to elucidate causal vs. correlational frameworks and establish contributions from reduced renal function, chronic medication use, and comorbidity.
Statements
Data availability statement
The data concern health care data of kidney transplant recipients. Even in a pseudonymized dataset, individual patients would be recognizable based on the combination of parameters age, sex, eGFR, medication use, transplantation date, donor age and donor sex. The pseudonymized data can be shared upon reasonable request from individual researchers, provided a data transfer agreement is in place between our institution and the institution of the requesting researcher.
Ethics statement
The studies involving humans were approved by the Medical Ethical Committee of the UMCG (METc 2014/077) and were conducted in accordance with the Declaration of Helsinki, the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.
Author contributions
NS: Investigation, Writing – review & editing, Software, Project administration, Writing – original draft, Methodology, Visualization, Formal analysis, Validation, Resources. KJ: Writing – original draft, Writing – review & editing. MR: Data curation, Validation, Methodology, Investigation, Writing – original draft, Resources, Writing – review & editing, Software. SE: Writing – review & editing, Writing – original draft, Supervision, Project administration. JA: Writing – review & editing, Writing – original draft, Supervision. MB: Writing – review & editing, Conceptualization, Writing – original draft, Supervision. KB: Writing – review & editing, Supervision, Formal analysis, Validation, Writing – original draft, Methodology, Software. SB: Writing – original draft, Investigation, Supervision, Writing – review & editing, Methodology, Data curation, Conceptualization, Resources, Funding acquisition, Validation, Project administration. MH-F: Data curation, Methodology, Resources, Conceptualization, Investigation, Formal analysis, Writing – original draft, Writing – review & editing, Supervision, Project administration, Validation.
Funding
The author(s) declared that financial support was received for this work and/or its publication. The author(s) declare that the TransplantLines Biobank and Cohort study was supported by grants from Astellas BV (TransplantLines Biobank and Cohort study), Chiesi Pharmaceuticals BV (PA-SP/PRJ 2020–9136), and NWO/TTW via a partnership program with DSM, Animal Nutrition and Health, The Netherlands (14939). The project was co-financed by the Dutch Ministry of Economic Affairs and Climate Policy by means of so-called PPP-allowances, made available by the Top Sector Life Sciences & Health to stimulate public-private partnerships (PPP-2019-032 and PPP-2022-015). M.H.d.B. is funded by the European Union (ELIMINATE-CKD, ERC-COG 101125516). Funders had no role in the study design, data collection, analysis, reporting, or decisions to submit for publication.
Acknowledgments
We thank the colleagues at the University Medical Center Groningen for their assistance and collaboration, and we thank the patients who enabled this work and entrusted us with their samples.
Conflict of interest
The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
The author(s) SE, JA declared that they were an editorial board member of Frontiers, at the time of submission. This had no impact on the peer review process and the final decision.
Generative AI statement
The author(s) declared that generative AI was used in the creation of this manuscript. The authors used generative AI to assist with language editing and improvement of sentence structure while preparing the manuscript. The authors take responsibility for all content created.
Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.
Publisher’s note
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Supplementary material
The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/frtra.2026.1833811/full#supplementary-material
TransplantLines Investigators
Coby Annema, Stephan J.L. Bakker, Stefan P. Berger, Hans Blokzijl, Frank A.J.A. Bodewes, Marieke T. de Boer, Kevin Damman, Martin H. de Borst, Isabelle J.C. Dielwart, Arjan Diepstra, Gerard Dijkstra, Caecilia S.E. Doorenbos, Rianne M. Douwes, Michele F. Eisenga, Michiel E. Erasmus, C. Tji Gan, Antonio W. Gomes Neto, Eelko Hak, Bouke G. Hepkema, Jip Jonker, Frank Klont, Tim J. Knobbe, Daan Kremer, Henri G.D. Leuvenink, Willem S. Lexmond, Vincent E. de Meijer, Hubert G.M. Niesters, Gertrude J. Nieuwenhuijs-Moeke, L. Joost van Pelt, Robert A. Pol, Anna M. Posthumus, Adelita V. Ranchor, Jan Stephan F. Sanders, Marion J. Siebelink, Riemer H.J.A. Slart, J. Casper Swarte, Daan J. Touw, Marius C. van den Heuvel, Coretta van Leer-Buter, Marco van Londen, Charlotte A. te Velde Keyzer, Erik A.M. Verschuuren, Michel J. Vos, Rinse K. Weersma.
References
1.
LiabeufSCheddaniLMassyZA. Uremic toxins and clinical outcomes: the impact of kidney transplantation. Toxins (Basel). (2018) 10:229. 10.3390/toxins10060229
2.
WojtaszekEOldakowska-JedynakUKwiatkowskaMGlogowskiTMalyszkoJ. Uremic toxins, oxidative stress, atherosclerosis in chronic kidney disease, and kidney transplantation. Oxid Med Cell Longev. (2021) 2021:6651367. 10.1155/2021/6651367
3.
AnderssonCHansenDSørensenSSMcGrathMMcCauslandFRTorp-PedersenCet al. Long-term cardiovascular events, graft failure, and mortality in kidney transplant recipients. Eur J Intern Med. (2024) 121:109–13. 10.1016/j.ejim.2023.10.026
4.
WyldMLRDe La MataNLMassonPO’LoneEKellyPJWebsterAC. Cardiac mortality in kidney transplant patients: a population-based cohort study 1988–2013 in Australia and New Zealand. Transplantation. (2021) 105:413–22. 10.1097/TP.0000000000003224
5.
HariharanSIsraniAKDanovitchG. Long-Term survival after kidney transplantation. N Engl J Med. (2021) 385:729–43. 10.1056/NEJMra2014530
6.
LangewischEMannonRB. Chronic allograft injury. Clin J Am Soc Nephrol. (2021) 16:1723–9. 10.2215/CJN.15590920
7.
PoggioEDAugustineJJArrigainSBrennanDCScholdJD. Long-term kidney transplant graft survival—making progress when most needed. Am J Transplant. (2021) 21:2824–32. 10.1111/ajt.16463
8.
HocherBAdamskiJ. Metabolomics for clinical use and research in chronic kidney disease. Nat Rev Nephrol. (2017) 13:269–84. 10.1038/nrneph.2017.30
9.
PattiGJYanesOSiuzdakG. Innovation: metabolomics: the apogee of the omics trilogy. Nat Rev Mol Cell Biol. (2012) 13:263–9. 10.1038/nrm3314
10.
BouatraSAziatFMandalRGuoACWilsonMRKnoxCet al. The human urine metabolome. PLoS One. (2013) 8:e73076. 10.1371/journal.pone.0073076
11.
EmwasAHLuchinatCTuranoPTenoriLRoyRSalekRMet al. Standardizing the experimental conditions for using urine in NMR-based metabolomic studies with a particular focus on diagnostic studies: a review. Metabolomics. (2015) 11:872–94. 10.1007/s11306-014-0746-7
12.
KalimSRheeEP. An overview of renal metabolomics. Kidney Int. (2017) 91:61–9. 10.1016/j.kint.2016.08.021
13.
SuhreKSchwartzJESharmaVKChenQLeeJRMuthukumarTet al. Urine metabolite profiles predictive of human kidney allograft Status. J Am Soc Nephrol. (2016) 27:626–36. 10.1681/ASN.2015010107
14.
KostidisSBankJRSoonawalaDNevedomskayaEvan KootenCMayborodaOAet al. Urinary metabolites predict prolonged duration of delayed graft function in DCD kidney transplant recipients. Am J Transplant. (2019) 19:110–22. 10.1111/ajt.14941
15.
KalantariSChashmniamSNafarMSamavatSRezaieDDaliliN. A noninvasive urine metabolome panel as potential biomarkers for diagnosis of T cell-mediated renal transplant rejection. OMICS. (2020) 24:140–7. 10.1089/omi.2019.0158
16.
BroadhurstDIKellDB. Statistical strategies for avoiding false discoveries in metabolomics and related experiments. Metabolomics. (2007) 2:171–96. 10.1007/s11306-006-0037-z
17.
LeekJTScharpfRBBravoHCSimchaDLangmeadBJohnsonWEet al. Tackling the widespread and critical impact of batch effects in high-throughput data. Nat Rev Genet. (2010) 11:733–9. 10.1038/nrg2825
18.
EisengaMFGomes-NetoAWVan LondenMZiengsALDouwesRMStamSPet al. Rationale and design of TransplantLines: a prospective cohort study and biobank of solid organ transplant recipients. BMJ Open. (2018) 8:e024502. 10.1136/bmjopen-2018-024502
19.
CockcroftDWGaultH. Prediction of creatinine clearance from serum creatinine. Nephron. (1976) 16:31–41. 10.1159/000180580
20.
IwamotoHOkiharaMAkashiIKiharaYKonnoOKawachiSet al. Metabolomic profiling of plasma, urine, and Saliva of kidney transplantation recipients. Int J Mol Sci. (2022) 23:13938. 10.3390/ijms232213938
21.
HallanSAfkarianMZelnickLRKestenbaumBSharmaSSaitoRet al. Metabolomics and gene expression analysis reveal down-regulation of the citric acid (TCA) cycle in non-diabetic CKD patients. EBioMedicine. (2017) 26:68–77. 10.1016/j.ebiom.2017.10.027
22.
StanimirovaIBanasikMZąbekADawiskibaTKościelska-KasprzakKWojtowiczWet al. Serum metabolomics approach to monitor the changes in metabolite profiles following renal transplantation. Sci Rep. (2020) 10:17223. 10.1038/s41598-020-74245-z
23.
PavlovićNKrižanacMKumrićMVukojevićKBožićJ. Mitochondrial dysfunction: the silent catalyst of kidney disease progression. Cells. (2025) 14:794. 10.3390/cells14110794
24.
FunkJASchnellmannRG. Persistent disruption of mitochondrial homeostasis after acute kidney injury. Am J Physiol Renal Physiol. (2012) 302:F853–64. 10.1152/ajprenal.00035.2011
25.
WeinbergJMVenkatachalamMARoeserNFNissimI. Mitochondrial dysfunction during hypoxia/reoxygenation and its correction by anaerobic metabolism of citric acid cycle intermediates. Proc Natl Acad Sci U S A. (2000) 97:2826–31. 10.1073/pnas.97.6.2826
26.
Martínez-ReyesIChandelNS. Mitochondrial TCA cycle metabolites control physiology and disease. Nat Commun. (2020) 11:102. 10.1038/s41467-019-13668-3
27.
BhargavaPSchnellmannRG. Mitochondrial energetics in the kidney. Nat Rev Nephrol. (2017) 13:629–46. 10.1038/nrneph.2017.107
28.
ChouchaniETPellVRGaudeEAksentijevićDSundierSYRobbELet al. Ischaemic accumulation of succinate controls reperfusion injury through mitochondrial ROS. Nature. (2014) 515:431–5. 10.1038/nature13909
29.
MurphyMPChouchaniET. Why succinate? Physiological regulation by a mitochondrial coenzyme Q sentinel. Nat Chem Biol. (2022) 18:461–9. 10.1038/s41589-022-01004-8
30.
ArizaACDeenPMTRobbenJH. The succinate receptor as a novel therapeutic target for oxidative and metabolic stress-related conditions. Front Endocrinol. (2012) 3:22. 10.3389/fendo.2012.00022
31.
GuoYChoSWSaxenaDLiX. Multifaceted actions of succinate as a signaling transmitter vary with its cellular locations. Endocrinol Metab (Seoul). (2020) 35:36–43. 10.3803/EnM.2020.35.1.36
32.
BernardiABiasiaFPatiTPivaMD’AngeloABuccianteG. Long-term protein intake control in kidney transplant recipients: effect in kidney graft function and in nutritional status. Am J Kidney Dis. (2003) 41:S146–52. 10.1053/ajkd.2003.50105
33.
WangXCaiZWangQWuCSunYWangZet al. Bacteroides methylmalonyl-CoA mutase produces propionate that promotes intestinal goblet cell differentiation and homeostasis. Cell Host Microbe. (2024) 32:63–78.e7. 10.1016/j.chom.2023.11.005
34.
ShaferMLowVLiZBlenisJ. The emerging role of dysregulated propionate metabolism and methylmalonic acid in metabolic disease, aging, and cancer. Cell Metab. (2025) 37:316–29. 10.1016/j.cmet.2025.01.005
35.
TejeroJLazureFGomesAP. Methylmalonic acid in aging and disease. Trends Endocrinol Metab. (2024) 35:188–200. 10.1016/j.tem.2023.11.001
36.
GaoXChenWLiRWangMChenCZengRet al. Systematic variations associated with renal disease uncovered by parallel metabolomics of urine and serum. BMC Syst Biol. (2012) 6(Suppl 1):S14. 10.1186/1752-0509-6-S1-S14
37.
SigdelTKSchroederAWYangJYCSarwalRDLibertoJMSarwalMM. Targeted urine metabolomics for monitoring renal allograft injury and immunosuppression in pediatric patients. J Clin Med. (2020) 9:2341. 10.3390/jcm9082341
38.
HolleJReitmeirRBehrensFSinghDSchindlerDPotapenkoOet al. Gut microbiome alterations precede graft rejection in kidney transplantation patients. Am J Transplant. (2025) 25:1643–56. 10.1016/j.ajt.2025.02.010
39.
HeJYangMQuanXWenJLanYYangHet al. Microbial and metabolic features in renal transplant recipients with post-transplantation diabetes mellitus. Int J Urol. (2023) 30:504–13. 10.1111/iju.15158
40.
DongCTrimignoAJestineJZhaoJO’DayEMKuypersDR. Urinary metabolite signatures to detect and differentiate graft injury in kidney transplant patients. iScience. (2025) 28:113331. 10.1016/j.isci.2025.113331
41.
Post HospersGVisserWJVerhoevenJGHPLagingMBaartSJMertens Zur BorgIRAMet al. Delayed graft function after kidney transplantation: the role of residual diuresis and waste products, as oxalic acid and its precursors. Transpl Int. (2024) 37:13218. 10.3389/ti.2024.13218
Summary
Keywords
clinical research, healthy donors, kidney transplantation, transplant recipients, urinary metabolomics
Citation
Shehab NBN, Lammers-Jannink KCM, Riphagen MH, Eskandari SK, Azzi JR, de Borst MH, Bunte K, Bakker SJL and Heiner-Fokkema MR (2026) Benchmarking the urinary transplant metabolome of kidney transplant recipients using healthy organ donors. Front. Transplant. 5:1833811. doi: 10.3389/frtra.2026.1833811
Received
18 March 2026
Revised
21 April 2026
Accepted
30 April 2026
Published
03 June 2026
Volume
5 - 2026
Edited by
Mercy Rophina, New York University, United States
Reviewed by
Vitaliy Sazonov, Nazarbayev University, Kazakhstan
Cezar Baluta, Grigore T. Popa University of Medicine and Pharmacy, Romania
Updates
Copyright
© 2026 Shehab, Lammers-Jannink, Riphagen, Eskandari, Azzi, de Borst, Bunte, Bakker and Heiner-Fokkema.
This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY). The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.
*Correspondence: Naser B. N. Shehab n.b.n.shehab@umcg.nl
† These authors have contributed equally to this work
Disclaimer
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.