Abstract
Objective
Glucose degradation products (GDPs) are important for the outcome of peritoneal dialysis (PD) treatment. The most cytotoxic GDP found in conventionally manufactured fluids, 3,4-dideoxyglucosone-3-ene (3,4-DGE), may in addition be recruited from 3-deoxyglucosone (3-DG). What happens with the GDPs in the fluid infused into patients during PD is not known. We investigated whether 3,4-DGE and 3-DG in PD fluid can be found in plasma during treatment.
Design
Patients on PD were dialyzed with a conventional PD fluid containing 43 μmol/L 3,4-DGE and 281 μmol/L 3-DG. Parallel experiments were performed in rats and in vitro with human plasma. The rats were dialyzed with a PD fluid containing 100 μmol/L 3,4-DGE and 200 μmol/L 3-DG.
Results
The 3,4-DGE concentration in the peritoneum declined at a much higher rate during the dwell than did the 3-DG concentration. However, 3,4-DGE was not detected in the plasma of patients or of rats during dialysis. The 3-DG concentration in plasma peaked shortly after infusion of fluid into the peritoneal cavity. The 3,4-DGE concentration during experimental incubation in plasma declined rapidly; the 3-DG concentration declined only 10% as rapidly (or less).
Conclusion
During dialysis, 3,4-DGE could not be detected in plasma of either PD patients or rats, presumably because of its high reactivity. On the other hand, 3-DG may pass through the membrane and be detected in the blood.
Keywords
During sterilization of peritoneal dialysis (PD) fluids, glucose degrades into various compounds—that is, glucose degradation products (GDPs). Many hundreds of different GDPs are probably formed, and to date, approximately 10 compounds (constituting a preponderance of the GDPs created) have been identified in PD fluids. Some of these hundreds are dicarbonyls such as 3,4-dideoxyglucosone-3-ene (3,4-DGE) and 3-deoxyglucosone (3-DG).
Several clinical studies demonstrated a negative effect of the presence of GDPs in PD fluids. Markers such as cancer antigen 125, procollagens, and hyaluronan in PD effluent indicate superior biocompatibility of fluids with low levels of GDPs. In patient dialysate, GDPs also form advanced oxidation protein products and advanced glycation end-products (AGEs) reflecting high intraperitoneal oxidative stress. The presence of GDPs has recently been linked to decline in residual renal function (RRF) (1), and a retrospective observational study indicated superior patient survival among patients treated with low-GDP fluids (2).
The α,β–unsaturated carbonyl 3,4-DGE has been identified in PD fluids. Substances in this class, being extremely reactive, are known to modify cellular processes and to exert a variety of toxic effects. The concentration of 3,4-DGE in conventional PD fluids is normally 10 – 50 μmol/L, but can be considerably higher directly after the sterilization process. However, after 2 months of storage at room temperature, this concentration declines by approximately 90% (3) because of a temperature-dependent equilibrium in which 3,4-DGE exists in balance with 3-deoxyaldose-2-ene (3-DA) and 3-DG. Furthermore, 3,4-DGE is a precursor in the irreversible formation of 5-hydroxymethylfuraldehyde.
Even though 3,4-DGE is the most reactive and toxic GDP in PD fluid, it is not known whether 3,4-DGE or another GDP is responsible for side effects in PD patients. Very few studies have investigated what happens with GDPs after PD fluid is infused into a patient. It has been demonstrated that cytotoxicity and GDPs in PD fluids disappear during a dwell in a time-dependent manner. However, the GDPs entering the peritoneal cavity might have different fates. They could react with proteins and cells in the dialysate. They could also react with constituents of the membrane such as mesothelial cells and fibroblasts, or they might be transported from the peritoneal cavity into the circulation.
Only a small fraction of the GDPs in PD fluid appear to be transported into the circulation. In rats, only about 3% of the “lost” 3-DG in dialysate is recovered in plasma (4). The actual appearance of 3,4-DGE and 3-DG in patient plasma during PD treatment has so far not been investigated. Thus, it is not known whether these GDPs might be related to observed clinical side effects such as decline in RRF.
We therefore decided to investigate whether 3,4-DGE or 3-DG could be identified in plasma after infusion of conventional PD fluids.
Materials and Methods
Gdps in PD Patients
Before, during, and after a dwell, blood and dialysate samples were taken from 4 continuous ambulatory peritoneal dialysis patients who used a conventional PD fluid, Dianeal 3.86% (Baxter Healthcare, Castlebar, Ireland), that contains relatively high concentrations of 3,4-DGE. During the treatment, 2-mL blood samples were collected at 0, 15, 30, 60, and 240 minutes. The blood was centrifuged, and the plasma (approximately 1 mL) was used for analysis. Before and after the dialysis treatment, 3,4-DGE and 3-DG in plasma and in the dialysis fluid or effluent were analyzed using high-performance liquid chromatography (HPLC).
Gdps in Rats
Dialysis was performed on 9 male Wistar rats (Møllegaard, Copenhagen, Denmark) having an average weight of 265 g (± 21 g standard deviation) with 20 mL 2.5% Gambrosol Trio dialysis fluid (Gambro Lundia AB, Lund, Sweden), to which had been added 100 μmol/L 3,4-DGE and 200 μmol/L 3-DG. The experiment was set up as previously described (5). Blood samples (500 μL) were taken at 0, 10, 20, 40, 60, 90, and 120 minutes of the dwell. Dialysate samples (250 μL) were taken at 0, 10, 20, 40, 60, 90, and 120 minutes. All samples were immediately put on ice. Concentrations of 3,4-DGE and 3-DG in plasma and dialysate were immediately analyzed by HPLC.
In Vitro Incubation of 3,4-Dge and 3-DG with Plasma
Plasma from healthy donors was incubated with 35 μmol/L 3,4-DGE standard at 37°C for various times under 2 hours. The same procedure was performed using 35 μmol/L 3-DG. The 3,4-DGE and 3-DG concentrations were later determined by HPLC, and the fluorescence was analyzed.
Analysis of 3,4-DGE in Plasma
In the plasma samples, 3,4-DGE was analyzed using a HPLC system (Agilent Technologies, Waldbronn, Germany) equipped with ultraviolet detection at 230 nm. Plasma (100 μL) was diluted with 400 μL water.
To remove ionic and non-polar analytes, the sample was filtered through an Isolute SPE mixed-mode column [International Sorbent Technology, Hengoed, U.K. (part no. 904-0010-B)]. The packing material (100 mg and 3-mL reservoir volume) consisted of silica particles covered with 18C, -SO–3, and -NR+3 as functional groups. The particle size was 30 – 90 μm, and the specific surface area was 521 m2/g.
To extract 3,4-DGE from the filtered sample after removal of ionic and non-polar analytes, it was passed through an Isolute SPE ENV+ column [International Sorbent Technology (part no. 915-0010-A)]. The packing material (100 mg and 1-mL reservoir volume) of this column consisted of a hyper–cross-linked polystyrene–divinylbenzene copolymer that is efficient for extraction of polar analytes. The particle size was 30 – 160 μm, and the specific surface area was 1000 m2/g. After the sample passed through the column, it was washed with 1 mL water. The 3,4-DGE bound to the column was then extracted with 1 mL 20% ethanol in water solution into a Vivaspin 2 filter [Vivascience Sartorius, Stonehouse, U.K. (part no. VS0201)] with a 10-kDa molecular weight cut-off. The filtered extract was centrifuged (5700 rpm) for 1 hour, frozen (–20°C), and freeze-dried. The freeze-dried extract was dissolved in 1 mL water before HPLC analysis.
The 3,4-DGE standard was extracted from a heat-sterilized glucose-containing PD fluid as described in previous work.
Analysis of 3-DG in Plasma
Plasma (100 μL) was diluted with 400 μL water. The sample was placed in a Vivaspin 2 filter (10-kDa cut-off) and centrifuged (5700 rpm) for 1 hour before HPLC analysis as previously described (3). A 3-DG standard was purchased from Toronto Research Chemicals (North York, ON, Canada).
Fluorescence in Plasma
The formation of AGEs during incubation of 35 μmol/L 3,4-DGE and 3-DG with plasma was measured as fluorescence at 430 nm (emission) and 350 nm (excitation) respectively as described by Lamb et al. (6). The plasma samples were diluted 10 times with water before analysis. The instrument was equipped with a fluorescence spectrophotometer (Hitachi High-Technologies Corporation, Tokyo, Japan).
Detection Limit of 3,4-DGE in Plasma
The detection limit for 3,4-DGE in plasma from healthy donors was determined after adding 0.35, 0.7, 1.4, 3.5, 7.0, and 14 μmol/L 3,4-DGE standard to the plasma. Thereafter 3,4-DGE was analyzed by HPLC.
Statistical Analysis
All data are expressed as mean ± standard error of the mean. Statistical evaluations were performed using oneway ANOVA for multiple comparisons. The chosen level of significance was p < 0.05.
Results
Gdps in PD Patients
In plasma samples taken from patients during the dwell, 3,4-DGE was below the detection limit (1.4 μmol/L). The concentration of 3,4-DGE in the PD fluid before infusion was 43 μmol/L; it could not be detected in the dialysate after the 4-hour dwell. The concentration of 3-DG in plasma peaked at 15 minutes at a concentration of 1.5 μmol/L. Thereafter, the concentration of 3-DG declined slowly during the dwell. The concentration of 3-DG in the dialysis fluid was 281 μmol/L before dialysis and 76 μmol/L in the effluent at the end of the treatment (240 minutes).
Gdps in the Rats
After infusion of the PD fluid containing 100 μmol/L 3,4-DGE and 200 μmol/L 3-DG into the peritoneal cavity, the concentration of 3,4-DGE in the effluent rapidly declined. The concentration of 3-DG declined at a slower rate, similar to that seen when glucose was transported over the peritoneal membrane (5). No detectable amount of 3,4-DGE was found in plasma at any time during the dwell. The concentration of 3-DG in plasma increased during the first 20 minutes of the dwell and peaked at 4.5 μmol/L. Thereafter, it declined slowly during the rest of the dwell.
In Vitro Incubation of 3,4-DGE and 3-DG in Plasma
Incubation of 35 μmol/L 3,4-DGE at 37°C in plasma resulted in a decline in 3,4-DGE to 6 μmol/L after 240 minutes. After 1 day of incubation, the concentration of 3,4-DGE was undetectable. The concentration of 3-DG declined slowly during the incubation from 35 μmol/L at the beginning to 32 μmol/L after 240 minutes. After 1 day of incubation, the concentration of 3-DG had declined further, to 22 μmol/L.
Fluorescence
After 4 hours of incubation, the level of fluorescence was significantly higher in the samples in which 3,4-DGE and 3-DG had been added separately (p < 0.05). Fluorescence in the plasma to which 3,4-DGE had been added increased at a considerably higher rate than it did in the plasma solution to which 3-DG had been added.
Discussion
Despite a detection limit of 1.4 μmol/L, 3,4-DGE could not be detected in the plasma of PD patients during a PD treatment. In contrast, 3-DG could be detected during the entire dwell. We conclude that 3,4-DGE does not reach the blood in free form during PD treatment, but that 3-DG does. This observation was further strengthened by rat experiments, in which 3,4-DGE was not detectable in blood, but 3-DG was, despite a rather high 3,4-DGE concentration added to the PD fluid.
The expected concentration of 3,4-DGE in the rat blood was about 2.3 μmol/L, assuming 3,4-DGE to have transport characteristics and reactivity similar to those of 3-DG. However, 3,4-DGE is a far more reactive molecule than 3-DG is.
When 3,4-DGE was infused into the peritoneal cavity, it disappeared. This disappearance is probably a result of a rapid reaction with cells and proteins in the dialysate and with the cells lining the peritoneal cavity. The 3,4-DGE was therefore unable to reach the bloodstream in a detectable amount. It is not known if the pool of 3-DG transported into the blood produces 3,4-DGE there. The likelihood of detecting 3,4-DGE in conjugate (such as with glutathione) or protein-bound form may be greater than the likelihood of detecting 3,4-DGE per se. This hypothesis has to be further investigated.
Only a slow decline in 3-DG concentration was observed during the 4 hours of the in vitro incubation with plasma. After 1 day, the decline was more pronounced. It is not clear if this decline was a result of the production of 3,4-DGE or of a slow direct reaction of 3-DG with plasma proteins. We cannot exclude the possibility that 3-DG may be metabolized in the blood.
The GDPs promote the formation of AGEs, detectable as fluorescence, in patient plasma and in effluent during PD. However, it is not clearly known how the AGEs are formed in the blood. There are three possibilities:
Unreactive 3-DG and 3-DA may be transported from the dialysate into the circulation, there producing 3,4-DGE that later reacts with proteins, forming AGEs.
The 3,4-DGE may react with proteins locally in the dialysate, producing AGEs that are, in turn, transported into the circulation.
The 3,4-DGE may be directly transported into the circulation. (This last possibility is not likely, considering the high reactivity of the molecule.)
We conclude that 3,4-DGE could not be detected in plasma in either PD patients or rats during dialysis. This observation is probably a result of a rapid reaction of 3,4-DGE with tissues, cells, and proteins in the peritoneal cavity. On the other hand, 3-DG, which is not as reactive as 3,4-DGE (and probably 3-DA), may reach the blood. The 3,4-DGE present in the PD fluid is therefore probably not directly responsible for systemic effects in PD patients. Because 3-DG (and possibly 3-DA) are present in the patient, 3,4-DGE may be recruited from that pool, causing systemic effects by subsequently binding to proteins.
Footnotes
Acknowledgments
This study was financed by Gambro AB, Lund, Sweden.