Oleg Mayboroda (Presenter)
LUMC
Authorship: S. Kostidis(1), D. Soonawala(2), E. Nevedomskaya(1), C. van Kooten(2), J.W. de Fijter(2), O.A. Mayboroda(1)
(1) Center for Proteomics and Metabolomics, Leiden University Medical Center (LUMC), Leiden, The Netherlands; (2) Department of Nephrology, Leiden, Leiden University Medical Center (LUMC), The Netherlands
| Short Abstract Confronted with organ shortage for transplantation, many countries have been readdressing donation after cardiac death (DCD) donors as an alternative source to expand the organ pool. DCD grafts, however, are exposed to significantly more ischemia-reperfusion injury than organs derived from donation after brain death. Here, we demonstrate that the urinary metabolome can be used to monitor the process physiological stabilization of the kidney and resolution of ischemic injury after transplantation. We show that metabolic profiles at a single time point, combined with the clinical predictor of ECD can be applied for the practical assessment of the delayed graft function duration. |
Long Abstract
Introduction.
Confronted with organ shortage for transplantation, many countries have been readdressing donation after cardiac death (DCD) donors as an alternative source to expand the organ pool. In the Netherlands for example, from 1997 to 2008, the relative contribution of DCD to the deceased donor pool increased from 7% to more than 40% (source: Dutch Transplantation Foundation, Annual Reports 1997– 2008). DCD grafts, however, are exposed to significantly more ischemia-reperfusion injury than organs derived from donation after brain death. Previous studies have reported a three- to four-fold higher risk of delayed graft function (DGF) with DCD kidneys as compared to kidneys from donation after brain death donors (DBD)[1]. To this end an accurate, minimally invasive, prediction of duration of DGF is an important element in the post-transplantation monitoring of the patients. Current clinical routine for assessment of the graft condition is a complex process and includes invasive procedure such as renal biopsy. Here we have investigated an alternative or auxiliary approach, namely – metabolomics. Metabolomics is a post-genomic discipline aiming at studying the metabolites, i.e., the end points and the intermediate products of metabolism. The totality of the metabolites in a biological system is referred to as the metabolome. Metabolome of body fluids is the closest approximation of the physiological phenotype of an organism and, as such, it represents an important, so far largely undervalued, source of clinical information. Taking advantage of the material collected within a single center study (PROTECT) which included 92 DCD kidney transplant recipients [1], we present an exploratory analysis of the urinary metabolome at different time points after transplantation and an outline of a prediction model for DGF resolution.
Material and Methods
The study included 92 patients scheduled to receive a DCD kidney transplant at Leiden University Medical Center between August 2005 and December 2009. Kidneys were allocated according to the allocation algorithm and matching criteria of Eurotransplant. The study was approved by the Ethical Committee of the Leiden University Medical Center. Urine samples were collected at 10, 42, 180 and 360 days after transplantation. . 1H-NMR data of all samples were collected in a random order using a Bruker 600 MHz AVANCE II spectrometer (Bruker BioSpin, Karlsruhe, Germany) equipped with a 5-mm TCI cryo-probe and a z-gradient system; a Bruker SampleJet sample changer system was used for sample transfer; samples were kept at 4 °C while queued for acquisition. The technical details of the NMR experiments are reported elsewhere [2]. The data analysis was performed using “R” (cran.r-project.org), version 3.2.4 for Windows. Exploratory data analysis was performed using “Rcpm” (http://cpm.lumc.nl/R/) and “rrcov” packages. For variable section and subset validation “caret”, “glmnet” and “vsurf” packages were used. For data visualization ggplot2 package was used.
Results
An exploratory analysis of the entire data showed no significant trends associated with the common confounding factors such as age and gender. The main variance in the data was explained by the time after transplantation. Thus, we applied a time trajectory analysis method [3], which showed that “recovery trajectories” of patients without DGF were clearly different from the ones with DGF. Moreover, the same approach revealed differences between patients with a DGF duration up to 21 day and those who experienced DGF for more than 21 days. Our data also indicated that the main metabolic remodeling occurred within 42 days after transplantation and that the current sampling density was not sufficient for in depth analysis of the process within that time period. Therefore, we concentrated on the samples collected at 10 day after transplantation and questioned whether urinary metabolic profiles can be used for the prediction of DGF duration, e.g. whether DGF is going to be resolved before or after the 21 day cut off point. Using a combination of variable selection methods and cross-validation tools we succeeded in selecting a subset of three metabolites. Passing a selected subset of variables into logistic regression we obtain a model with an AUC = 0.78. When we combined our model of metabolic predictors with the clinical definition for expanded-criteria donors (ECD), the combined model had a strong predictive value (AUC=0.92, chi square of 29.35 and chi square probability of 6.63e-06).
Conclusions.
We have demonstrated that the urinary metabolome can be used to monitor the process physiological stabilization of the kidney and resolution of ischemic injury after transplantation. We show that metabolic profiles at a single time point, combined with the clinical predictor of ECD can be applied for the practical assessment of the DGF duration. To extend and take full advantage of the possible options offered by metabolic profiling a cohort with more frequent sampling in the period between 10 and 42 days is required.
References & Acknowledgements:
1. Aydin Z, Mallat MJK, Schaapherder AFM, van Zonneveld AJ, van Kooten C, Rabelink TJ, de Fijter JW: Randomized Trial of Short-Course High-Dose Erythropoietin in Donation After Cardiac Death Kidney Transplant Recipients. Am J Transplant 2012, 12(7):1793-1800.
2. Nevedomskaya E, Pacchiarotta T, Artemov A, Meissner A, van Nieuwkoop C, van Dissel JT, Mayboroda OA, Deelder AM: H-1 NMR-based metabolic profiling of urinary tract infection: combining multiple statistical models and clinical data. Metabolomics 2012, 8(6):1227-1235.
3. Keun HC, Ebbels TMD, Bollard ME, Beckonert O, Antti H, Holmes E, Lindon JC, Nicholson JK: Geometric trajectory analysis of metabolic responses to toxicity can define treatment specific profiles. Chem Res Toxicol 2004, 17(5):579-587.
| Description | Y/N | Source |
| Grants | no | |
| Salary | no | |
| Board Member | no | |
| Stock | no | |
| Expenses | no |
IP Royalty: no
| Planning to mention or discuss specific products or technology of the company(ies) listed above: | no |