Journal of Hepatology
Volume 41, Issue 6 , Pages 1038-1040, December 2004

Hepatic clearance of advanced glycation end products (AGEs)—myth or truth?

Department of Experimental Pathology, Institute of Medical Biology, University of Tromsø, N-9037 Tromsø, Norway

published online 18 October 2004.

See Article, pages 913‐919

Article Outline

 

The liver sequesters a number of circulating macromolecular soluble and particular waste products. This blood clearance function is carried out by the cells that line the sinusoidal wall: (i) the resident liver macrophages, or the Kupffer cells (KC), and (ii) the liver sinusoidal endothelial cells (LSEC). The KC is tuned to phagocytic uptake of large particles and aggregates, whereas LSEC are specialized on clathrin mediated endocytosis of soluble macromolecules and colloids. Until now four major endocytosis receptors have been observed to mediate waste endocytosis in LSEC. The mannose receptor, the scavenger receptor (several types are expressed by LSEC; the hyaluronan/scavenger receptor, or Stabilin-2 [1] is uniquely expressed and functionally active in LSEC), the collagen α chain receptor, and the Fc-γ receptor. The waste macromolecules that are cleared by these receptors in LSEC include (i) most types of connective tissue molecules that are constantly released to the circulation as a consequence of normal turnover processes throughout the body [2]; (ii) extracellular enzymes and products of platelet-mediated coagulation [3]; (iii) intracellular macromolecules (for instance lysosomal enzymes [4]); (iv) soluble IgG-immune complexes [5]; (v) native macromolecules that have been modified non-enzymatically by for instance oxidation (oxidized low density lipoprotein [6]) or glycation; (vi) foreign molecules (i.e. LPS [7]).

The capacity of endocytosis in LSEC is impressive: some of the waste macromolecules are turned over in quantities of several grams per day in a normal adult human individual. The process of endocytosis in these cells is the most efficient known: endocytic receptors recycle between the plasma membrane and early endosomes with a half-life of a few seconds (most other cell types perform endocytosis with a receptor recycling half-life of minutes). Most of the waste substances that are destined for uptake via receptor-mediated endocytosis in LSEC exist for a very short time in the circulation (less than 1min in rats).

During the past three decades several groups have attempted to design experiments to study the fate of circulating advanced glycation end products (AGEs). Macromolecules modified in this way are present in normal individuals throughout life, and are found in higher concentrations in older people as well as in certain diseases, most typically diabetes. Available data suggest that AGEs are eliminated from the blood mainly by scavenger receptor mediated uptake in KC and LSEC [8]. Alternative hypotheses hold that the uptake is not necessarily in the liver, but in the kidneys. One line of research (hereafter referred to as ‘the in vitro approach’) makes use of in vitro generated AGEs, that can be labelled and chased after administration in vivo. Another approach (hereafter referred to as ‘the in vivo approach’) is based on the notion that chemical analysis of blood samples, with no prior administration of in vitro generated AGEs, represents the key to solve the problem [9], [10]. According to the in vivo approach it would be sufficient to perform sensitive chemical analyses to check if and to what extent AGEs are removed by any given tissue. As discussed below these approaches have distinct strong and weak sides; neither of them represents a perfect approach to determine the anatomical site of uptake of circulating AGEs. In vitro generation of AGEs involves long time (weeks or months) incubation of protein with glucose or other AGEs precursors under aseptic conditions. The in vitro approach offers the possibility of labelling AGEs with high specific radioactivity, enabling very low amounts of AGEs to be chased in vivo after i.v. administration. The weak side of the in vitro approach is that we do not know to what extent AGEs prepared in vitro represent the ‘native’ in vivo generated AGEs. One knows for certain that the AGE adducts that are present in in vitro generated AGEs are also detected on AGEs formed in vivo. But it is unlikely that the extent of AGE modification is as high in the in vivo formed specimen as in the in vitro generated molecules. Using the in vitro approach it was found that i.v. administered AGEs are very rapidly taken up in KC and LSEC [8].

The advantage of the in vivo approach is that only native AGEs are measured. However, using only the in vivo approach one will not be able to detect the most interesting AGEs, namely those that are rapidly cleared from the circulation. Even the most sensitive analytical tools presently available will not be able to show significant differences in the concentration of AGEs in blood samples taken from the portal and hepatic veins, simply because the speed and efficiency of uptake greatly exceed the rate of AGEs formation. The speed of blood circulation must also be considered: in average the recycling time for blood through liver in humans is 3.6min, meaning that all the blood contents are monitored by the liver scavenger receptors every few minutes. This would make it impossible for AGEs modified to a ‘high physiological degree’ to accumulate to a detectable level above the background in the blood. Moreover, the natural formation of AGEs in the blood is certainly slower than a few minutes. One of the authors (P.J.T.) of the presently discussed article previously used a 24h incubation schedule to prepare CML-albumin and methylglyoxal-derived hydroimidazolone-proteins (two AGEs) with minimal degree of modification [11], [12]. Understanding this dynamics is crucial for the appreciation that AGEs modified to a ‘high physiological degree’ escape detection due to (i) their very slow formation, (ii) their very rapid uptake, and (iii) the very efficient blood recycling.

Using the in vivo approach, the high resolution analytical tool LC-MS/MS has been used to determine the presence of specified AGEs in peripheral and hepatic venous blood (control human subjects), or portal venous and hepatic venous blood (cirrhotic subjects) [13]. With this state-of-the-art methodology the authors observed no or only minute differences in the level and type of AGEs that enter and leave the liver. From this, the authors conclude that liver does not contribute to extraction of in vivo formed AGEs from the blood. They also put forward a hypothesis that the kidneys represent the major site of elimination of AGEs from the blood. At first glance, this may seem like a plausible interpretation. However, a closer look at the premises makes it clear that more solid data is required for a shift of the current paradigm of hepatic elimination of AGEs.

The authors state that in vivo formation of proteins highly modified by AGEs is unlikely considering the kinetics of albumin glycation under physiological conditions. If ‘highly modified’ means AGE-modification to the same extent as in vitro modified AGEs prepared by traditional methods [14], this statement by the authors would be agreed upon by most AGE-researchers. There is a general concensus that highly modified AGE-albumin prepared in vitro is not a perfect model for studies of AGEs turnover in vivo. Nevertheless, one has to ask the following question: do in vivo formed AGEs have to be modified to the same extent as the commonly used in vitro highly modified AGE albumin in order to be recognized by receptors for endocytic uptake? In fact, one of the authors (P.J.T.) of the presently discussed article previously reported that HSA minimally modified by methylglyoxal (MGmin-HSA) (1.4–2.4 modified arginine residues per molecule) is taken up by receptor-mediated endocytosis and degraded by the monocytic cell line THP-1 [11]. The degree of AGE-modification was indeed much lower in this MGmin-HSA (1–2 arginines per albumin molecule) compared with conventionally used in vitro formed highly modified AGE-BSA (in a typical batch of highly modified AGE-BSA 37 of 59 lysines and 10 of 23 arginines are modified). Furthermore, it is logical to assume that LSEC, which exhibit a higher endocytic activity and express more scavenger receptors compared to monocytes, would represent a more efficient site of uptake of MGmin-HSA. The authors reported previously that approximately 2% of total HSA contains MG-derived imidazolon (MG-H1) in normal control subjects [9]. Assuming a total content of 250g of albumin in the blood of a normal adult individual, 2% would correspond to 5g of MG-H1-albumin, which will be present in the circulation at any time. There is another important consideration to take into account when comparing AGE-modification of protein in vitro and in vivo: due to the fact that the same individual protein molecules are present during the in vitro generation of AGE-albumin, each of the albumin molecules present will bind AGE-adducts with the same probability. In contrast, since the probability of AGE-modification increases with increasing life time of any individual protein in vivo, the generation of AGEs in vivo will result in differently modified proteins, spanning from ‘young’, newly synthesized proteins containing no or very few AGE-modifications, to ‘old’ proteins containing most of the modifications. On this basis, one would come closer to reality by assuming that only albumin molecules that have existed for more than one half-life will carry all the MG-H1 groups. Using this assumption, along with the result published previously by P.J.T. [11], it follows that 4% of the albumin molecules older than one half-life carry MG-H1. Moreover, applying combinatorics analysis it can be calculated that the probability for any one of these albumin molecules to carry more than 1 MG-H1 is 5.9%. This means that at any time 300mg of albumin molecules in the circulation of normal humans will have more than 1 MG-H1 residues. According to the paper by P.J.T. cited above [11], this degree of modification is sufficient to bring about receptor-mediated uptake in scavenger cells. Of note, this calculation most likely represents an underestimation, since it was based on only one type of AGE-modification. In fact, 12 different AGE-species are presently known, and it is known that formation of AGEs can results in products that contain several AGE-adducts on the same protein molecule. On this basis one can safely assume that the amount of protein sufficiently modified to bring about endocytic uptake and degradation will be significantly higher than that calculated on the basis of only MG-H1-mofication. It should be noted that this calculation applies to healthy humans. In the diabetic state, for example, the protein modification will be much higher than in normals due to the increased levels of glucose, along with increased serum concentrations of many types of AGE-precursors, such as glyoxal, methylglyoxal and 3-deoxyglucosone, that increase 1.2, 3.4 and 3.1 times, respectively [15]. Conceivably, these circumstances when taken together, generate AGEs with a high enough degree of modification for scavenger receptor-mediated uptake in liver. The likelihood for this to happen in the diabetic patient is even greater. But alas, concluding from the considerations discussed above it is practically impossible to detect these AGE-modified proteins in the blood because they will disappear almost immediately after reaching the modification threshold for uptake in the liver RES.

From the above considerations it is conceivable that more solid evidence is needed if the current paradigm of elimination of AGEs in liver scavenger cells be exchanged with a new paradigm stating that AGEs are eliminated mainly in extrahepatic tissues.

Back to Article Outline

References 

  1. Politz O, Gratchev A, McCourt PA, Schledzewski K, Guillot P, Johansson S, et al. Stabilin-1 and -2 constitute a novel family of fasciclin-like hyaluronan receptor homologues. Biochem J. 2002;362:155–164
  2. Smedsrød B, Pertoft H, Gustafson S, Laurent TC. Scavenger functions of the liver endothelial cell. Biochem J. 1990;266:313–327
  3. Hansen B, Melkko J, Smedsrød B. Serum is a rich source of ligands for the scavenger receptor of hepatic scavenger endothelial cells. Mol Cell Biochem. 2002;229:63–72
  4. Smedsrød B, Tollersrud OK. Sinusoidal liver endothelial cells recruit lysosomal enzymes from the circulation by mannose—receptor mediated endocytosis. In:  Wake K,  Wisse E,  Knook DL editor. Cells of the hepatic sinusoid. vol. 5:Leiden, The Netherlands: Kupffer Cell Foundation; 1995;p. 180–183
  5. Lovdal T, Andersen E, Brech A, Berg T. Fc receptor mediated endocytosis of small soluble immunoglobulin G immune complexes in Kupffer and endothelial cells from rat liver. J Cell Sci. 2000;113:3255–3266
  6. De Rijke YB, Biessen EA, Vogelezang CJ, van Berkel TJ. Binding characteristics of scavenger receptors on liver endothelial and Kupffer cells for modified low-density lipoproteins. Biochem J. 1994;15:69–73
  7. van Oosten M, van Amersfoort ES, van Berkel TJ, Kuiper J. Scavenger receptor-like receptors for the binding of lipopolysaccharide and lipoteichoic acid to liver endothelial and Kupffer cells. J Endotoxin Res. 2001;7:381–384
  8. Smedsrod B, Melkko J, Araki N, Sano H, Horiuchi S. Advanced glycation end products are eliminated by scavenger-receptor-mediated endocytosis in hepatic sinusoidal Kupffer and endothelial cells. Biochem J. 1997;322:567–573
  9. Thornalley PJ, Battah S, Ahmed N, Karachalias N, Agalou S, Babaei-Jadidi R, et al. Quantitative screening of advanced glycation endproducts in cellular and extracellular proteins by tandem mass spectrometry. Biochem J. 2003;375:581–592
  10. Thornalley PJ, Argirova M, Ahmed N, Mann VM, Argirov O, Dawnay A. Mass spectrometric monitoring of albumin in uremia. Kidney Int. 2000;58:2228–2234
  11. Westwood ME, Argirov OK, Abordo EA, Thornalley PJ. Methylglyoxal-modified arginine residues—a signal for receptor-mediated endocytosis and degradation of proteins by monocytic THP-1 cells. Biochim Biophys Acta. 1997;1356:84–94
  12. Westwood ME, Thornalley PJ. Molecular characteristics of methylglyoxal-modified bovine and human serum albumins. Comparison with glucose-derived advanced glycation endproduct-modified serum albumins. J Protein Chem. 1995;14:359–372
  13. Ahmed N, Thornalley PJ, Lüthen R, Häussinger D, Sebekova K, Schinzel R, et al. Processing of protein glycation, oxidation and nitrosation adducts in the liver and the effect of cirrhosis. J Hepatol. 2004;41:913–919
  14. Nagai R, Matsumoto K, Ling X, Suzuki H, Araki T, Horiuchi S. Glycolaldehyde, a reactive intermediate for advanced glycation end products, plays an important role in the generation of an active ligand for the macrophage scavenger receptor. Diabetes. 2000;49:1714–1723
  15. Odani H, Shinzato T, Matsumoto Y, Usami J, Maeda K. Increase in three alpha, beta-dicarbonyl compound levels in human uremic plasma: specific in vivo determination of intermediates in advanced Maillard reaction. Biochem Biophys Res Commun. 1999;256:89–93

PII: S0168-8278(04)00462-3

doi:10.1016/j.jhep.2004.10.004

Journal of Hepatology
Volume 41, Issue 6 , Pages 1038-1040, December 2004

Article Tools