AGE Content of a Protein Load Is Responsible for Renal Performances: A Pilot Study

Chronic kidney disease is associated with higher morbidity and mortality in patients with diabetes because of an increased risk of cardiovascular complications (1). Low-protein diets are an additional measure to slow the progression of diabetic nephropathy (2,3) because each protein load leads to renal functional reserve (RFR) mobilization, characterized by an increase of renal blood flow, followed by an increase in the glomerular filtration rate (GFR) in healthy subjects (4,5). Nevertheless, RFR repeated mobilizations, as seen in the first stages of diabetic nephropathy, lead to RFR loss, followed by diabetic nephropathy progression (6). The precise mechanisms responsible for kidney hemodynamic changes induced by a protein load remain unknown, but Uribarri and Tuttle (7) suggested that advanced glycation end products (AGE) could be responsible. Using the new positron emission tomography (PET) and MRI device (BioGraph mMR; Siemens, Berlin, Germany) to simultaneously assess renal perfusion using [¹⁵O]H₂O PET imaging, renal oxidative metabolism using [¹¹C]-acetate PET imaging, and renal oxygen content using blood oxygenation–level dependent (BOLD)-MRI, we wanted to evaluate renal performance before and 120 min after high-AGE and low-AGE protein loads.

The study was a single-center, prospective, randomized, cross-over designed trial, registered at ClinicalTrials.gov (NCT02695251). Our study was accepted by the local ethics committee (Comité de Protection des Personnes Sud-Est II, no. 2015-52-2) on 30 December 2015 and by the French National Drugs Agency (ANSM, no. 1515451-11) on 24 January 2016 and was performed according to the principles outlined in the Declaration of Helsinki. Informed signed consent was obtained for all participants after information was given.

Inclusion criteria were male sex, age between 18 and 30 years, no diabetes, no hypertension, and no chronic kidney disease. Each intervention (low- and high-AGE load) was performed on 2 different days, each separated by a minimum 1-week interval, in a crossover design after an overnight fast.

Our primary end point was the increase in renal perfusion assessed by PET using [¹⁵O]H₂O, and our secondary outcomes were oxidative metabolism assessed by PET using [¹¹C]-acetate and renal oxygen content measured by BOLD-MRI. PET and BOLD measurements were performed at baseline and 120 min after each meal.

Patients had to follow a low-protein (0.8 g/kg/day) and low-AGE diet for 48 h before each intervention.

The high-protein (1 g/kg) high-AGE meal consisted of mixed Chicken McNuggets (McDonald’s, 4 nuggets = 11 g protein, 0.83 g sodium chloride, 12 g carbohydrate, and 9.3 g lipid, >30,000 kU AGE [8]). The high-protein low-AGE meal consisted of mixed 10 min–cooked large-sized eggs (2 “caliber L” eggs = 12 g protein, 0.29 g sodium chloride, 0.8 g carbohydrate, and 10.6 g lipid, <3,000 kU AGE per serving).

PET-MRI data were blinded for the type of intervention performed, and analyzed by G.N.

Randomization (high vs. low AGE first) was achieved using the RAND function of an Excel file (Microsoft Excel for Mac 2011, version 14.6.3). Quantitative variables were expressed as the mean ± SD. The Mann-Whitney test was used to compare two means. The paired t test or paired Wilcoxon matched signed rank test were used for matched data analysis, depending on the sample size. Statistical significance was defined as P < 0.05. Statistical analysis was performed using GraphPad software (GraphPad Software, La Jolla, CA).

We included 10 healthy subjects from February to June 2016, with a mean age of 22 ± 3.7 years. Renal function was normal, with a mean estimated GFR of 121 ± 13 mL/min 1.73 m² and no proteinuria (0.1 ± 0.1 g/L). Fasting blood glucose at inclusion was normal at 5.0 ± 0.3 g/L. Nine patients completed the study (both meals). AGE per serving was 36,144 ± 4,030 kU AGE for the high-AGE meal and 2,321 ± 197 kU AGE for the low-AGE meal (P = 0.004), based on tables in Uribarri et al. (8).

Renal perfusion, assessed by PET using [¹⁵O]H₂O, increased significantly after the high-AGE meal (3.16 ± 0.5 to 3.8 ± 0.42 mL/g/min, +27.2%; P = 0.0002, for both cortices), whereas there was no significant change after the low-AGE meal (3.35 ± 0.65 to 3.38 ± 0.53 mL/g/min, +3%; P = 0.88) (Fig. 1).

Oxygen consumption increased significantly after the high-AGE meal (0.30 ± 0.02 to 0.36 ± 0.08 min⁻¹, +20.8%; P = 0.005) compared with the low-AGE meal (0.30 ± 0.04 to 0.31 ± 0.07 min⁻¹, +0.5%; P = 0.8) for both cortices.

There was no change in BOLD medullary R2* measurements at 120 ± 5 min compared with the baseline values after the low-AGE meal (medullary R2* values from 26.4 ± 2.22 to 28.8 ± 2.99, P = 0.12) or the high-AGE meal (from 29.5 ± 3.32 to 32.3 ± 6.91, P = 0.25).

Using functional imaging in vivo and in humans, this study suggests that only a high-AGE high-protein load induces a significant mobilization of RFR. These findings could represent a paradigm shift.

Renal effects of proteins have been evidenced in humans and animals (4,9) using meat meal and amino acid perfusion, but those protein loads were all rich in AGE because high-temperature heating, used for meat meal cooking and amino acid preparation, is associated with AGE generation. Those “protein loads” were, therefore, high-AGE protein loads (8).

Molecular mechanisms involved in the increase of renal blood flow and hyperfiltration after a “protein load” are currently not known, but AGE hemodynamic effect has recently been highlighted. First, Uribarri et al. (10) demonstrated that a single oral AGE challenge could acutely impair brachial artery endothelial function. Second, Lin et al. (11) showed that methylglyoxal, an AGE agent, induces cyclooxygenase-2 expression in synovial cells, known to increase prostaglandin local concentration, which could explain an increase in renal blood flow, followed by RFR mobilization. Indeed, patients who eat a “high-protein” diet exhibit higher prostaglandin plasmatic concentrations, followed by RFR mobilization, whereas a pretreatment with cyclooxygenase-2 inhibitors completely blocks the protein-induced renal hemodynamic variations (12).

Our study presents some limitations. First, we decided to lead a clinical study of young healthy subjects, with the aim of reproducing initial studies’ experimental conditions. This study design did not allow us to determine which AGE agent was responsible for the hemodynamic effect observed. After our study, this could be performed with specific intravenous injection of animals. Second, despite a different lipid and carbohydrates content between our low- and high-AGE loads, those two interventions bring especially a different salt content (mean difference for each subject: 3.4 ± 0.45 g). Studies have already demonstrated that lipid and carbohydrate are not responsible for RFR mobilization (13,14), whereas an acute sodium load is not responsible for acute intrarenal hemodynamic changes compared with an amino acid perfusion (15). Finally, we hypothesized that an increase in kidney oxidative metabolism would lead to a decrease in the oxygen content of kidneys. Unfortunately, postprandial breath holding (20 s), required for BOLD acquisition, was difficult, and six subjects exhibited movements that precluded precise BOLD measurements.

To our knowledge, this is the first study to support the deleterious effect of the AGE content of a protein load on renal hemodynamics. This study suggests that prevention of diabetic nephropathy progression could aim predominantly at reducing protein AGE content more than protein.

Acknowledgments. The authors thank Maxime Paturel (senior dietitian, Hôpital Edouart Herriot); Nans Florens (Department of Nephrology, Hôpital Edouard Herriot); Gérard Gimenez (Centre d'Etude et de Recherche Multimodal et Pluridisciplinaire); Elise Mistretta, Jamila Lagha, Véronique Berthier, Fréderic Bonnefoi, Thibaut Iecker, and Christian Tourvieille (Centre d’Etude et de Recherche Multimodal et Pluridisciplinaire) for pertinent advice and technical assistance; Annie Varennes (Laboratoire de Biochimie, Hôpital Edouard Herriot); Monica Sigovan (Centre d’Etude et de Recherche Multimodal et Pluridisciplinaire) for providing in-house software support; Catherine Cereser, Valérie Plattner, and Cécile Riera (Hospices Civils de Lyon Research Center) for their administrative assistance and support; Carole Dhelens and Damien Salmon (Department of Pharmacology, Hôpital Edouart Herriot) for their help in building the initial project; and Jean-Jacques and Benoit Bertin (Lyon, France) for their technical support.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. G.N., S.L., and N.C. participated in conception and design of the study and in analysis and interpretation of data and drafted and revised the manuscript. M.V. participated in conception and design of the study and in analysis and interpretation of data and revised the draft of the manuscript. D.L.B. participated in conception of the study, provided intellectual content of critical importance to the work described, and revised the draft of the manuscript. I.M. and Z.I. participated in analysis and interpretation of data and provided intellectual content of critical importance to the work described. T.T. participated in conception and design of the study, mostly the MRI protocol conception part. L.J. participated in conception and design of the study and in analysis and interpretation of data, drafted and revised the manuscript, and provided intellectual content of critical importance to the work described. G.N., S.L., M.V., and D.L.B. accepted the final version of the manuscript to be published. I.M., Z.I., T.T., N.C., and L.J. gave final approval of the version to be published. G.N. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.