N epsilon-(carboxymethyl)lysine is a dominant advanced glycation end product (AGE) antigen in tissue proteins

Advanced glycation end products (AGEs) and glycoxidation products are formed during Maillard or browning reactions between sugars and proteins and are implicated in the pathophysiology of aging and the complications of diabetes. To determine the structure of AGEs, antibodies were prepared to protein browned by incubation with glucose and used in ELISA assays to measure AGEs formed in model reactions between bovine serum albumin (BSA) or N alpha-acetyllysine and glucose, fructose, or glyoxal. AGEs were formed from glucose and fructose only under oxidative conditions, but from glyoxal under both oxidative and antioxidative conditions. Gel permeation chromatographic analysis indicated that a similar AGE was formed in reactions of N alpha-acetyllysine with glucose, fructose, and glyoxal and that this AGE co-eluted with authentic N alpha-acetyl-N epsilon-(carboxymethyl)lysine. Amino acid analysis of AGE proteins revealed a significant content of N epsilon-(carboxymethyl)lysine (CML). In ELISA assays using polyclonal antibodies against AGE proteins, CML-BSA (approximately 25 mol of CML/mol of BSA), prepared by chemical modification of BSA, was a potent inhibitor of the recognition of AGE proteins and of AGEs in human lens proteins. We conclude that AGEs are largely glycoxidation products and that CML is a major AGE recognized in tissue proteins by polyclonal antibodies to AGE proteins.

Effects of radiolabeling monoclonal antibodies with a residualizing iodine radiolabel on the accretion of radioisotope in tumors

The effect of using a "residualizing" iodine radiolabel, dilactitol-iodotyramine, for radioimmunolocalization of antibodies to tumors was investigated. This tracer is designed to be lysosomally trapped after catabolism of the labeled antibody. mAbs RS7 and RS11 were used for in vivo and in vitro studies on the uptake and retention of radioisotope into tumor cells. Both are murine IgG1 mAbs with pancarcinoma reactivity, which react with integral membrane glycoproteins. mAb RS7 has been shown to be relatively rapidly catabolized by the antigen-bearing cell line Calu-3, whereas RS11 is catabolized more slowly in the same cells. An 111In- or 88Y-p-isothiocyanatobenzyl-diethylenetriamine pentaacetic acid conjugate was also tested because these radiometals are known to be lysosomally trapped, and iodination via chloramine T was used to provide a baseline. In vitro, a substantial increase in retention of the label by cells was observed when the dilactitol-tyramine DLT- or 111In-labeled mAbs were used, and the improvement gained by the use of these residualizing labels was greater with the use of the rapidly catabolized mAb (RS7) than it was with the more slowly catabolized mAb (RS11). In biodistribution studies in nude mice bearing Calu-3 tumor xenografts, a dramatic improvement in the tumor accretion of the radiolabel was seen with the use of the 131I-labeled DLT- or 88Y-labeled mAbs. For example, at day 7 the percentage of injected dose/g in the tumor was 5.54 +/- 1.47% (SD), 38.06 +/- 8.04%, and 43.18 +/- 19.50% for the conventionally iodinated, DLT- and 88Y-labeled RS7, respectively. Dosimetry calculations performed on the biodistribution data predict increases of approximately 8- and 4-fold in the absorbed dose to tumor with the use of 131I-labeled DLT- and 90Y-labeled mAbs, respectively, compared to the conventional 131I. In contrast to in vitro findings, these results were similar for both RS7 and RS11, suggesting that the use of DLT may be advantageous for most of the mAbs binding to the cell surface, including antibodies that are catabolized relatively slowly. The advantage of 131I-labeled DLT over 90Y is due to the longer physical half-life of the 131I.

Identification of liver endothelial cells as the primary site of IgM catabolism in the rat

Rat IgMs, both monoclonal protein from ascites fluid and total serum IgM, were purified by sequential gel filtration and metal chelate affinity chromatography on immobilized zinc-iminodiacetate. Two monoclonal IgMs, IR202 and IR968, chromatographed identically on gel filtration, but required different pHs for elution from the zinc affinity column. IR202 behaved like a euglobulin, being readily precipitated in low-ionic-strength buffers, while IR968 remained soluble under these conditions. IgM was isolated from serum in 30-50% yield by chromatographic procedures similar to those used for the monoclonal proteins, and 20-30% of the isolated serum IgM was precipitable as a euglobulin. The half-life of both monoclonal and serum euglobulin IgMs was 0.8 days, while the polyclonal globulin and IR968 had half-lives of 1.8 and 2.8 days, respectively, in the rat circulation. The tissue and cellular sites of catabolism of the monoclonal IgMs were determined after labeling with the residualizing label, dilactitol-[125I]tyramine. For both proteins the liver was identified as the major tissue site of catabolism, accounting for 60-80% of degraded protein in the body. When liver was fractionated into parenchymal and nonparenchymal cells (NPC), the NPC were found to account for 86 and 69% of protein recovered in liver, for IR202 and IR968, respectively. Separation of NPC into endothelial (EC) and Kupffer cell populations by elutriation centrifugation revealed that EC contained the majority, approximately 70% of total NPC radioactivity from either IgM. Based on the ratios of endocytic indices (microliter of plasma/10(6) cells/day) for each cell type, the EC also had a higher efficiency for uptake of both IgMs, approximately threefold greater, than for the fluid phase marker, polyvinylpyrrolidone, or for rat serum albumin. We conclude that hepatic EC are a major site of IgM catabolism, regardless of the heterogeneity in physical and biological properties of various IgM populations.

Mechanism of autoxidative glycosylation: identification of glyoxal and arabinose as intermediates in the autoxidative modification of proteins by glucose

Glycation and oxidation reactions contribute to protein modification in aging and diabetes. Formation of dicarbonyl sugars during autoxidation of glucose is the hypothetical first step in the autoxidative glycosylation and subsequent browning of proteins by glucose [Wolff, S. P., & Dean, R. T. (1987) Biochem. J. 245, 243-250]. In order to identify the dicarbonyl sugar(s) formed during autoxidation of glucose under physiological conditions, glucose was incubated in phosphate buffer (pH 7.4) at 37 degrees C under air (oxidative conditions) or nitrogen with transition metal chelators (antioxidative conditions). Dicarbonyl compounds were analyzed spectrophotometrically and by HPLC after reaction with Girard-T reagent. Carbohydrates were analyzed by gas chromatography-mass spectrometry. Both dicarbonyl sugar and arabinose concentrations increased with time and glucose concentration in incubations conducted under oxidative conditions; only trace amounts of these products were detected in glucose incubated under antioxidative conditions. HPLC analysis of adducts formed with Girard-T reagent indicated that glyoxal was the only alpha-dicarbonyl sugar formed on autoxidation of glucose. Glyoxal and arabinose accounted for > or = 50% of the glucose lost during a 21 day incubation. Neither glucosone nor its degradation product, ribulose, was detectable. Reaction of glyoxal with RNase yielded the glycoxidation product, N epsilon-(carboxymethyl)lysine, while arabinose is a source of pentosidine. Our results implicate glyoxal and arabinose as intermediates in the browning and crosslinking of proteins by glucose under oxidative conditions. They also provide a mechanism by which antioxidants and dicarbonyl trapping reagents, such as aminoguanidine, limit glycoxidation reactions and support further evaluation of these types of compounds for inhibition of chemical modification and crosslinking of proteins during aging and diabetes.

Formation of reactive intermediates from Amadori compounds under physiological conditions

The Maillard or browning reaction between reducing sugars and proteins contributes to the chemical aging of tissue proteins in vivo and to the accelerated aging of proteins in diabetes. To identify reactive carbohydrate intermediates formed in the Maillard reaction under physiological conditions, we studied the decomposition of the model Amadori compound, N alpha-formyl-N epsilon-fructoselysine (fFL) and of Amadori compounds on glycated collagen at pH 7.4 and 37 degrees C. Because of effects of buffer and oxidative conditions on the decomposition of Amadori compounds, the kinetics and products of decomposition were studied in varying phosphate concentrations and in N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes) buffer under both aerobic and anaerobic conditions. The half-life of fFL was significantly shorter in phosphate, compared to Hepes buffer, and under aerobic, compared to anaerobic, conditions. The decomposition of both fFL and Amadori adducts on glycated collagen was accelerated by increasing the phosphate concentration and/or pH. Glucose and mannose were identified as major products formed by reversal of the Amadori rearrangement, along with tetroses, pentoses, and 3-deoxyglucosone, formed by reverse aldol, rearrangement, and hydrolysis reactions. The tetrose and pentose products included both aldose and ketose sugars. These same products were also formed in similar yields on decomposition of Amadori adducts on glycated collagen in vitro. The spontaneous decomposition of Amadori compounds to more reactive sugars in vivo, including tetroses, pentoses, and 3-deoxyglucosone, provides a mechanism for generating reactive intermediates under physiological conditions and for propagating damage to protein as a result of glycation of proteins by glucose in vivo.

Chemistry of the fructosamine assay: D-glucosone is the product of oxidation of Amadori compounds

The chemistry of the fructosamine assay was studied by using the Amadori compound, N alpha-formyl-N epsilon-fructose-lysine (fFL), an analog of glycated lysine residues in protein. Previously (Clin Chem 1993;39:2460-5), we reported that free lysine was formed from fFL at 70% yield during incubation with alkaline nitroblue tetrazolium (NBT) under the conditions routinely used for the fructosamine assay (sodium carbonate buffer, pH 10.35 at 37 degrees C). Here, we show that D-glucosone is the primary carbohydrate oxidation product formed from Amadori compounds in the fructosamine assay. Glucosone, which decomposes under alkaline assay conditions with a half-life of < 30 min, reaches a maximum concentration of approximately 50% of the initial fFL concentration after 10 min of incubation. Like fFL, glucosone reduces NBT to the purple monoformazan dye, but its decomposition is not accelerated by the presence of NBT. The dicarbonyl-trapping reagent, aminoguanidine, inhibits the fructosamine assay by approximately 25% when fFL is the substrate, but by nearly 100% with glucosone as substrate. Studies with serum samples from diabetics and nondiabetics indicate that glucosone formation does not have a significant effect on the clinical usefulness of the fructosamine assay; however, corrections for glucosone formation may be required when the assay is used for estimating the extent of glycation of proteins.

Chemistry of the fructosamine assay: D-glucosone is the product of oxidation of Amadori compounds

The chemistry of the fructosamine assay was studied by using the Amadori compound, N alpha-formyl-N epsilon-fructose-lysine (fFL), an analog of glycated lysine residues in protein. Previously (Clin Chem 1993;39:2460-5), we reported that free lysine was formed from fFL at 70% yield during incubation with alkaline nitroblue tetrazolium (NBT) under the conditions routinely used for the fructosamine assay (sodium carbonate buffer, pH 10.35 at 37 degrees C). Here, we show that D-glucosone is the primary carbohydrate oxidation product formed from Amadori compounds in the fructosamine assay. Glucosone, which decomposes under alkaline assay conditions with a half-life of &lt; 30 min, reaches a maximum concentration of approximately 50% of the initial fFL concentration after 10 min of incubation. Like fFL, glucosone reduces NBT to the purple monoformazan dye, but its decomposition is not accelerated by the presence of NBT. The dicarbonyl-trapping reagent, aminoguanidine, inhibits the fructosamine assay by approximately 25% when fFL is the substrate, but by nearly 100% with glucosone as substrate. Studies with serum samples from diabetics and nondiabetics indicate that glucosone formation does not have a significant effect on the clinical usefulness of the fructosamine assay; however, corrections for glucosone formation may be required when the assay is used for estimating the extent of glycation of proteins.

3-Deoxyfructose concentrations are increased in human plasma and urine in diabetes

3-Deoxyglucosone (3-DG) is a reactive dicarbonyl sugar thought to be a key intermediate in the nonenzymatic polymerization and browning of proteins by glucose. 3-DG may be formed in vivo from fructose, fructose 3-phosphate, or Amadori adducts to protein, such as N epsilon-fructoselysine (FL), all of which are known to be elevated in body fluids or tissues in diabetes. Modification of proteins by 3-DG formed in vivo is thought to be limited by enzymatic reduction of 3-DG to less reactive species, such as 3-deoxyfructose (3-DF). In this study, we have measured 3-DF, as a metabolic fingerprint of 3-DG, in plasma and urine from a group of diabetic patients and control subjects. Plasma and urinary 3-DF concentrations were significantly increased in the diabetic compared with the control population (0.853 +/- 0.189 vs. 0.494 +/- 0.072 microM, P < 0.001, and 69.9 +/- 44.2 vs. 38.7 +/- 16.1 nmol/mg creatinine, P < 0.001, respectively). Plasma and urinary 3-DF concentrations correlated strongly with one another, with HbA1c (P < 0.005 in all cases), and with urinary FL (P < 0.02 and P = 0.005, respectively). The overall increase in 3-DF concentrations in plasma and urine in diabetes and their correlation with other indexes of glycemic control suggest that increased amounts of 3-DG are formed in the body during hyperglycemia in diabetes and then metabolized to 3-DF. These observations are consistent with a role for increased formation of the dicarbonyl sugar 3-DG in the accelerated browning of tissue proteins in diabetes.

The processing and fate of antibodies and their radiolabels bound to the surface of tumor cells in vitro: a comparison of nine radiolabels

Processing radiolabeled degradation products is the key factor affecting retention of antibodies within the cell. In this study, we have analyzed the processing of antibodies labeled in nine different ways.

METHODS

Antibodies were labeled with three different radioisotopes and seven different forms of 125I. Eight of the radiolabels (except 188Re) were conjugated to the same antibody, MA103, and tested on the renal carcinoma cell line SK-RC-18 and/or the ovarian carcinoma cell line SK-OV-6. Rhenium conjugation utilized the antibody RS7, the target cell line ME180 and three of the other radiolabels were also tested with this antibody-target cell combination for comparison.

RESULTS

Iodine conjugated to antibodies by conventional methods was rapidly released from the cell after antibody catabolism. In contrast, iodinated moieties, such as dilactitol-tyramine and inulin-tyramine were retained within cells four to five times longer.

CONCLUSIONS

The use of radiolabels that are trapped within cells after antibody catabolism can potentially increase the dose of radiation delivered to the tumor, from the same amount of radioactivity deposited by a factor of four or five. The prolonged retention of 111In relative to 125I is not due to deiodination of iodine conjugates, but rather to intracellular retention of catabolic products containing 111In, perhaps within lysosomes.

Glycation, glycoxidation, and cross-linking of collagen by glucose. Kinetics, mechanisms, and inhibition of late stages of the Maillard reaction

The Maillard or browning reaction between sugar and protein contributes to the increased chemical modification and cross-linking of long-lived tissue proteins in diabetes. To evaluate the role of glycation and oxidation in these reactions, we have studied the effects of oxidative and antioxidative conditions and various types of inhibitors on the reaction of glucose with rat tail tendon collagen in phosphate buffer at physiological pH and temperature. The chemical modifications of collagen that were measured included fructoselysine, the glycoxidation products N epsilon-(carboxymethyl)lysine and pentosidine and fluorescence. Collagen cross-linking was evaluated by analysis of cyanogen bromide peptides using sodium dodecyl sulfate-polyacrylamide gel electrophoresis and by changes in collagen solubilization on treatment with pepsin or sodium dodecylsulfate. Although glycation was unaffected, formation of glycoxidation products and cross-linking of collagen were inhibited by antioxidative conditions. The kinetics of formation of glycoxidation products proceeded with a short lag phase and were independent of the amount of Amadori adduct on the protein, suggesting that autoxidative degradation of glucose was a major contributor to glycoxidation and cross-linking reactions. Chelators, sulfhydryl compounds, antioxidants, and aminoguanidine also inhibited formation of glycoxidation products, generation of fluorescence, and cross-linking of collagen without significant effect on the extent of glycation of the protein. We conclude that autoxidation of glucose or Amadori compounds on protein plays a major role in the formation of glycoxidation products and cross-liking of collagen by glucose in vitro and that chelators, sulfhydryl compounds, antioxidants, and aminoguanidine act as uncouplers of glycation from subsequent glycoxidation and cross-linking reactions.

Catabolism of hirudin and thrombin-hirudin complexes in the rat

The metabolic fate of the anticoagulant protein, hirudin, and its complex with thrombin are presently unknown. Therefore we have labelled hirudin and human thrombin-hirudin complex with the residualizing label dilactitol-125I-tyramine (*I-DLT) in order to identify their tissue sites of catabolism in the rat. The rapid plasma clearance of hirudin after intravenous injection was unaffected by *I-DLT labelling, and by 2 h 6% or less of the injected dose remained in the blood. The majority (80.3 +/- 4.0%, n = 2) of *I-DLT-hirudin radioactivity recovered in tissues was found in kidney, and kidney was also at least 150 times more active in taking up hirudin, on a weight basis, than any other tissue examined (liver, spleen, skin, muscle, intestine, fat, lung). *I-DLT-hirudin which bound to thrombin was isolated by chromatography on concanavalin A-Sepharose; hirudin itself does not bind to concanavalin A. Radioactivity from thrombin-*I-DLT-hirudin was precipitable by anti-thrombin antibody and *I-DLT-thrombin-hirudin was precipitable by anti-hirudin antibody. By 1 h after injection of labelled thrombin-hirudin complexes, the recoveries of radioactivity from hirudin and thrombin in liver were comparable (38.6 +/- 3.0 and 36.4 +/- 4.1%, n = 3), whereas more radioactivity was recovered in kidney from hirudin than from thrombin (27.6 +/- 8.7 compared with 13.6 +/- 4.5%) and less was recovered in lung (0.4 +/- 0.2 compared with 17.7 +/- 2.9%). We conclude that hirudin is catabolized predominantly in kidney, whereas the thrombin-hirudin complex is catabolized by both liver and kidney.

Mechanism of fructosamine assay: evidence against role of superoxide as intermediate in nitroblue tetrazolium reduction

We studied the chemistry of the fructosamine assay for glycated serum proteins by using the model Amadori compound N alpha-formyl-N epsilon-fructoselysine (fFL), an analog of glycated lysine residues in protein. Free lysine was formed at approximately 70% yield during a standard 20-min incubation of fFL with alkaline nitroblue tetrazolium (NBT) at 37 degrees C. Although superoxide dismutase (SOD; EC 1.15.1.1) and catalase (EC 1.11.1.6) decreased the yield of the product, monoformazan dye (MF+), the yield of MF+ was slightly greater under anaerobic than aerobic conditions, excluding a role for superoxide as an intermediate in the reduction of NBT during the fructosamine assay. SOD added to diabetic patients' sera at physiological concentrations also caused a significant (approximately 50%) inhibition of MF+ formation. This inhibition was reduced by addition of nonionic detergents, which contain organic peroxide inhibitors of SOD, to the fructosamine reagent. Overall, these data indicate that the Amadori compound is the direct reductant of NBT in the fructosamine assay and that superoxide is not an intermediate in the reaction. The inhibitory effects of SOD and catalase are most likely the result of oxygen regeneration in the assay mixture.

Mechanism of fructosamine assay: evidence against role of superoxide as intermediate in nitroblue tetrazolium reduction

We studied the chemistry of the fructosamine assay for glycated serum proteins by using the model Amadori compound N alpha-formyl-N epsilon-fructoselysine (fFL), an analog of glycated lysine residues in protein. Free lysine was formed at approximately 70% yield during a standard 20-min incubation of fFL with alkaline nitroblue tetrazolium (NBT) at 37 degrees C. Although superoxide dismutase (SOD; EC 1.15.1.1) and catalase (EC 1.11.1.6) decreased the yield of the product, monoformazan dye (MF+), the yield of MF+ was slightly greater under anaerobic than aerobic conditions, excluding a role for superoxide as an intermediate in the reduction of NBT during the fructosamine assay. SOD added to diabetic patients' sera at physiological concentrations also caused a significant (approximately 50%) inhibition of MF+ formation. This inhibition was reduced by addition of nonionic detergents, which contain organic peroxide inhibitors of SOD, to the fructosamine reagent. Overall, these data indicate that the Amadori compound is the direct reductant of NBT in the fructosamine assay and that superoxide is not an intermediate in the reaction. The inhibitory effects of SOD and catalase are most likely the result of oxygen regeneration in the assay mixture.

Formation of o-tyrosine and dityrosine in proteins during radiolytic and metal-catalyzed oxidation

To evaluate their usefulness as chemical indicators of cumulative oxidative damage to proteins, we studied the kinetics and extent of formation of ortho-tyrosine (o-Tyr), dityrosine (DT), and dityrosine-like fluorescence (Ex = 317 nm, Em = 407 nm) in the model proteins RNase and lysozyme exposed to radiolytic and metal-catalyzed (H2O2/Cu2+) oxidation (MCO). Although there were protein-dependent differences, o-Tyr, DT, and fluorescence increased coordinately during oxidation of the proteins in both oxidation systems. The contribution of DT to total dityrosine-like fluorescence in oxidized proteins varied from 2-100%, depending on the protein, type of oxidation, and extent of oxidative damage. In proteins exposed to MCO, DT typically accounted for > 50% of the fluorescence at DT wavelengths. These studies indicate that o-Tyr and DT should be useful chemical markers of cumulative exposure of proteins to MCO in vitro and in vivo.

Oxidized amino acids in lens protein with age. Measurement of o-tyrosine and dityrosine in the aging human lens

The concentrations of ortho-tyrosine (o-Tyr) and dityrosine (DT) were measured in noncataractous human lenses in order to assess the role of protein oxidation reactions in the aging of lens proteins. The measurements were conducted by selected ion monitoring-gas chromatography/mass spectrometry using deuterium-labeled internal standards, which provided both high sensitivity and specificity for the quantitation of o-Tyr and DT. Between ages 1 and 78 years, the o-Tyr concentration in lens proteins varied from 0.3 to 0.9 mmol of o-Tyr/mol of Phe (n = 19), while DT ranged from 1 to 3 mumol of DT/mol of Tyr (n = 30). There were no significant changes in levels of o-Tyr with lens age. There was a statistically significant, but only slight, increase in DT in lens proteins with age (approximately 33% increases between ages 1 and 78, r = 0.5, p < 0.01). At the same time, total protein fluorescence, measured at DT wavelengths (Ex = 317 nm, Em = 407 nm), increased 11-fold between ages 1 and 78 and correlated strongly with age (r = 0.82, p < 0.0001). Although the fluorescence maxima of lens proteins were similar to those of DT, DT accounted for less than 1% of the DT-like fluorescence in lens protein at all ages. These observations indicate that oxidation of Phe and Tyr plays a limited role in the normal aging of lens proteins in vivo.

Maillard reaction products and their relation to complications in insulin-dependent diabetes mellitus

Glycation, oxidation, and browning of proteins have all been implicated in the development of diabetic complications. We measured the initial Amadori adduct, fructoselysine (FL); two Maillard products, N epsilon-(carboxymethyl) lysine (CML) and pentosidine; and fluorescence (excitation = 328 nm, emission = 378 nm) in skin collagen from 39 type 1 diabetic patients (aged 41.5 +/- 15.3 [17-73] yr; duration of diabetes 17.9 +/- 11.5 [0-46] yr, [mean +/- SD, range]). The measurements were related to the presence of background (n = 9) or proliferative (n = 16) retinopathy; early nephropathy (24-h albumin excretion rate [AER24] > or = 20 micrograms/min; n = 9); and limited joint mobility (LJM; n = 20). FL, CML, pentosidine, and fluorescence increased progressively across diabetic retinopathy (P < 0.05, P < 0.001, P < 0.05, P < 0.01, respectively). FL, CML, pentosidine, and fluorescence were also elevated in patients with early nephropathy (P < 0.05, P < 0.001, P < 0.01, P < 0.01, respectively). There was no association with LJM. Controlling for age, sex, and duration of diabetes using logistic regression, FL and CML were independently associated with retinopathy (FL odds ratio (OR) = 1.06, 95% confidence interval (CI) = 1.01-1.12, P < 0.05; CML OR = 6.77, 95% CI = 1.33-34.56, P < 0.05) and with early nephropathy (FL OR = 1.05, 95% CI = 1.01-1.10, P < 0.05; CML OR = 13.44, 95% CI = 2.00-93.30, P < 0.01). The associations between fluorescence and retinopathy and between pentosidine and nephropathy approached significance (P = 0.05). These data show that FL and Maillard products in skin correlate with functional abnormalities in other tissues and suggest that protein glycation and oxidation (glycoxidation) may be implicated in the development of diabetic retinopathy and early nephropathy.

Accumulation of Maillard reaction products in skin collagen in diabetes and aging

To investigate the contribution of glycation and oxidation reactions to the modification of insoluble collagen in aging and diabetes, Maillard reaction products were measured in skin collagen from 39 type 1 diabetic patients and 52 nondiabetic control subjects. Compounds studied included fructoselysine (FL), the initial glycation product, and the glycoxidation products, N epsilon-(carboxymethyl) lysine (CML) and pentosidine, formed during later Maillard reactions. Collagen-linked fluorescence was also studied. In nondiabetic subjects, glycation of collagen (FL content) increased only 33% between 20 and 85 yr of age. In contrast, CML, pentosidine and fluorescence increased five-fold, correlating strongly with age. In diabetic patients, collagen FL was increased threefold compared with nondiabetic subjects, correlating strongly with glycated hemoglobin but not with age. Collagen CML, pentosidine and fluorescence were increased up to twofold in diabetic compared with control patients: this could be explained by the increase in glycation alone, without invoking increased oxidative stress. There were strong correlations among CML, pentosidine and fluorescence in both groups, providing evidence for age-dependent chemical modification of collagen via the Maillard reaction, and acceleration of this process in diabetes. These results support the description of diabetes as a disease characterized by accelerated chemical aging of long-lived tissue proteins.

The design and application of residualizing labels for studies of protein catabolism

Residualizing labels (R-labels) are chemical tags for proteins, originally designed for studies of the sites and mechanisms of plasma protein catabolism. The labels consist of oligosaccharides derivatized with radioactive, fluorescent, nuclear magnetic resonance (NMR), or positron emission tomography (PET) active reporter molecules. Because these glycoconjugates generally have molecular masses in excess of 500 daltons and are hydrophilic, they are relatively membrane impermeant. They are also designed to be resistant to lysosomal hydrolases and are therefore retained inside cells with half-lives of 2-5 days after endocytosis and degradation of the carrier protein. The R-labels thus provide a convenient means for following the cumulative uptake and catabolism of proteins by cells in vivo or in vitro. This review summarizes how R-labels have provided insights into the sites and regulation of the turnover of circulating proteins, and pathways for intracellular transport and degradation of endocytosed proteins. The potential use of R-labels for noninvasive studies of the distribution of protein pharmaceuticals in vivo is also discussed.

In vivo fate of hemopexin and heme-hemopexin complexes in the rat

The disposition in the rat of the plasma heme-binding protein hemopexin (Hx), as the native apoprotein and as its heme complex (HHx), has been studied using the residualizing protein label dilactitol-125I-tyramine (*I-DLT). The aim of this work was to identify the tissue sites of Hx uptake and catabolism, independent of heme binding, and to evaluate how heme loading affects Hx catabolism at these sites. *I-DLT-Hx had a circulating half-life of approximately 1.2 days and was recovered in degraded form in comparable amounts in visceral (liver, kidney, spleen) and peripheral (skin, muscle) tissues, indicating a generalized diffuse catabolism of the protein throughout the body. The plasma half-life of *I-DLT-Hx injected as a preformed heme-Hx complex was the same as that of the apoprotein; however, injection of the complex resulted in about a twofold increase in hepatic degradation of Hx. The lack of an effect of heme on overall catabolism of the preformed HHx complex was consistent with the approximately 1-h half-life of heme, injected as 14C-heme-Hx, in the circulation; however, as much as 20-fold more 14C-heme than Hx protein was recovered in liver from 14C-heme-Hx. The absolute amount of *I-DLT-Hx degraded in liver was significantly increased when heme was injected in excess of the heme binding capacity of circulating Hx, while 131I-DLT-albumin catabolism in liver was unaffected. Thus, depending on the physiological conditions studied, the data are consistent with a model in which, following hepatic uptake of heme from HHx, varying proportions of the protein are either returned to the circulation or degraded in the liver.

Role of oxygen in cross-linking and chemical modification of collagen by glucose

The role of oxygen in chemical modification and cross-linking of rat tail collagen by glucose was studied at physiological pH and temperature in vitro. Cross-linking of collagen under air depended on glucose concentration, but was inhibited under antioxidative conditions (nitrogen atmosphere with transition metal chelators). The cross-linking reaction under air depended on phosphate buffer concentration, but this effect was eliminated by addition of chelators, identifying trace metal ions in the buffer as catalysts of oxidative cross-linking reaction. Antioxidative conditions had no effect on glycation, that is, formation of fructose lysine, but inhibited formation of the glycoxidation products N epsilon-(carboxymethyl)lysine and pentosidine as well as the development of fluorescence in glycated collagen. Glycation itself decreased during continued incubation of the collagen without glucose; however, cross-linking and concentrations of glycoxidation products and fluorescence in collagen were not reversible under either oxidative or antioxidative conditions. These observations are consistent with recent studies in vivo on the reversibility of collagen glycation, the irreversibility of formation of glycoxidation products and fluorescence, and the strong correlations between glycoxidation products and fluorescence in collagen (1). These results indicate that oxidation reactions play a critical role in the extended chemical modification and cross-linking of collagen by glucose and suggest that measurement of glycoxidation products should be useful for assessing cumulative chemical modification of collagen by glucose in vivo.

Quantification of the accumulation and degradation of beta-very-low-density lipoproteins in vivo using a 19F-containing residualizing label and n.m.r. spectroscopy

beta-Very-low-density lipoproteins (beta-VLDL)O were conjugated to the 19F-containing residualizing label, NN-dilactitol-3,5-bis(trifluoromethyl)benzylamine (DLBA), to determine whether the metabolism of this lipoprotein fraction could be characterized in vivo with n.m.r. spectroscopy. Solution state 19F high-resolution n.m.r. spectroscopy of DLBA-beta-VLDL, containing either intact apoproteins or selectively enzymically digested products, demonstrated that the extent of degradation could be distinguished by differences in spin-spin relaxation times (T2 times). DLBA-beta-VLDL was injected intravenously into rabbits, and accumulation of 19F in hepatic tissue was quantified non-invasively by n.m.r. spectroscopy 5 and 30 h after injection. In addition to quantifying the accumulation of DLBA-beta-VLDL in hepatic tissue, a marked decrease (approx. 100 Hz) in the linewidth of 19F resonance from labelled lipoproteins was observed at 30 h compared with the 5 h interval in continuously monitored animals. The change in linewidth was consistent with a decrease in molecular size that occurred during protein degradation, resulting in increased T2 times. To demonstrate that T2 times can be used as an index to quantify apoprotein degradation in vivo, relaxation measurements were performed on livers excised 20 h after injection of DLBA-beta-VLDL into rabbits. Two molecular motional fractions were revealed by relaxation profiles representing either an intact or an extensively degraded form of apoprotein. The amplitudes of each component were compared with results from trichloroacetic acid precipitation of liver homogenates acquired from rabbits 20 h after injection of beta-VLDL labelled with the radioiodinated analogue of DLBA, dilactitol-125I-tyramine. The amount of degraded apoprotein determined by n.m.r. spectroscopy and acid precipitation was 68.6 +/- 7.0% and 58.7 +/- 7.5% (n = 4) respectively. The results of this study demonstrate that 19F n.m.r. spectroscopy can be used to define the temporal characteristics of the hepatic metabolism of lipoproteins in vivo by quantifying both the tissue-specific accumulation and extent of apoprotein degradation. The methodology developed offers promise for the non-invasive, sequential and longitudinal evaluation of lipoprotein metabolism in vivo.

Quantitative assessment of lipoprotein metabolism by positron emission tomography with an 18F-containing residualizing label

Residualizing labels for proteins are designed to remain entrapped within cells following uptake and degradation of the carrier protein. In the present work we report the synthesis of a novel residualizing label, N-lactitol-S-([18F]fluorophenacyl)-cysteamine ([18F]LCSH, and its use for quantifying the accumulation of low density lipoprotein in tissues in vivo by positron emission tomography (PET). The retention of degradation products in tissues from lipoprotein or from other rapidly catabolized protein pharmaceuticals tagged with [18F]LCSH reduces leakage of tracer into the plasma compartment. Thus, residualizing labels provide a valuable tool for enhancing signal-to-noise ratios, even during the relatively short interval of PET studies.

Imaging of thrombi with tissue-type plasminogen activator rendered enzymatically inactive and conjugated to a residualizing label

BACKGROUND

Contemporary cardiovascular practice relies increasingly on thrombolysis as a therapeutic modality. Its optimal use requires prompt, noninvasive delineation of thrombotic occlusion in arterial beds and rapid detection of reocclusion after initially successful thrombolysis.

METHODS AND RESULTS

We have been developing an approach to noninvasively image thrombi in which plasminogen-activating properties of tissue-type plasminogen activator (t-PA) are attenuated by treatment with D-Phe-L-Pro-L-Arg-chloromethyl ketone (PPACK) and have shown that the inactive t-PA avidly and promptly binds to clots in vitro. In the present study, we conjugated this material to a residualizing label, radioiodinated dilactitol tyramine (*I-DLT), and characterized the potential use of the inactivated, conjugated t-PA as a radiopharmaceutical for imaging thrombi in vivo. The approach developed requires not only avid binding of the tracer to thrombi but also rapid clearance from plasma and a lack of prompt release of radiolabeled degradation products from the liver. The rapid clearance of unaltered or PPACK-treated t-PA was not influenced by conjugation to *I-DLT, but the release of radioiodinated degradation products into plasma after injection of *I-DLT-conjugated t-PA was markedly less than release of degradation products of directly radioiodinated t-PA. When 131I-DLT-PPACK-t-PA was infused for 15 minutes intravenously after a bolus injection of 20% in dogs with coronary, pulmonary, or carotid artery thrombi, clearance was rapid. Mean +/- SEM thrombus-to-blood ratios of radioactivity were high, ranging from 37 +/- 9:1 and 2.8 +/- 0.6:1 with carotid thrombi formed concomitantly or approximately 30 minutes before infusion of tracer, respectively, to 35:1 for concomitantly formed coronary thrombi, 42 +/- 7:1 and 8.1 +/- 0.8:1 for concomitantly formed and preformed pulmonary thrombi, respectively, and 18:1 for a preformed femoral artery thrombus. Thrombi were detectable by planar gamma scintigraphy even though image quality was affected adversely by low concentrations of radioactivity that in aggregate composed a relatively large amount of radioactivity in underlying and overlying tissues. This limitation was overcome by tomographic imaging, which was used to detect both femoral and pulmonary thrombi.

CONCLUSIONS

Use of enzymatically inactivated t-PA coupled to a residualizing label permits rapid detection and localization of thrombi in vivo.

Formation of pentosidine during nonenzymatic browning of proteins by glucose. Identification of glucose and other carbohydrates as possible precursors of pentosidine in vivo

A fluorescent compound has been detected in proteins browned during Maillard reactions with glucose in vitro and shown to be identical to pentosidine, a pentose-derived fluorescent cross-link formed between arginine and lysine residues in collagen (Sell, D. R., and Monnier, V. M. (1989) J. Biol. Chem. 264, 21597-21602). Pentosidine was the major fluorophore formed during nonenzymatic browning of ribonuclease and lysozyme by glucose, but accounted for less than 1% of non-disulfide cross-links in protein dimers formed during the reaction. Pentosidine was formed in greatest yields in reactions of pentoses with lysine and arginine in model systems but was also formed from glucose, fructose, ascorbate, Amadori compounds, 3-deoxyglucosone, and other sugars. Pentosidine was not formed from peroxidized polyunsaturated fatty acids or malondialdehyde. Its formation from carbohydrates was inhibited under nitrogen or anaerobic conditions and by aminoguanidine, an inhibitor of advanced glycation and browning reactions. Pentosidine was detected in human lens proteins, where its concentration increased gradually with age, but it did not exceed trace concentrations (less than or equal to 5 mumol/mol lysine), even in the 80-year-old lens. Although its precise carbohydrate source in vivo is uncertain and it is present in only trace concentrations in tissue proteins, pentosidine appears to be a useful biomarker for assessing cumulative damage to proteins by nonenzymatic browning reactions with carbohydrates.

Age-dependent accumulation of N epsilon-(carboxymethyl)lysine and N epsilon-(carboxymethyl)hydroxylysine in human skin collagen

N epsilon-(Carboxymethyl)lysine (CML) is formed on oxidative cleavage of carbohydrate adducts to lysine residues in glycated proteins in vitro [Ahmed et al. (1988) J. Biol. Chem. 263, 8816-8821; Dunn et al. (1990) Biochemistry 29, 10964-10970]. We have shown that, in human lens proteins in vivo, the concentration of fructose-lysine (FL), the Amadori adduct of glucose to lysine, is constant with age, while the concentration of the oxidation product, CML, increases significantly with age [Dunn et al. (1989) Biochemistry 28, 9464-9468]. In this work we extend our studies to the analysis of human skin collagen. The extent of glycation of insoluble skin collagen was greater than that of lens proteins (4-6 mmol of FL/mol of lysine in collagen versus 1-2 mmol of FL/mol of lysine in lens proteins), consistent with the lower concentration of glucose in lens, compared to plasma. In contrast to lens, there was a slight but significant age-dependent increase in glycation of skin collagen, 33% between ages 20 and 80. As in lens protein, CML, present at only trace levels in neonatal collagen, increased significantly with age, although the amount of CML in collagen at 80 years of age, approximately 1.5 mmol of CML/mol of lysine, was less than that found in lens protein, approximately 7 mmol of CML/mol of lysine. The concentration of N epsilon-(carboxymethyl)hydroxylysine (CMhL), the product of oxidation of glycated hydroxylysine, also increased with age in collagen, in parallel with the increase in CML, from trace levels at infancy to approximately 5 mmol of CMhL/mol of hydroxylysine at age 80.(ABSTRACT TRUNCATED AT 250 WORDS)

Effect of diabetes and aging on carboxymethyllysine levels in human urine

Carboxymethyllysine (CML) has been identified as a modified amino acid that accumulates with age in human lens proteins and collagen. CML may be formed by oxidation of fructoselysine (FL), the Amadori adduct formed on nonenzymatic glycosylation of lysine residues in protein, or by reaction of ascorbate with protein under autoxidizing conditions. We proposed that measurements of tissue and urinary CML may be useful as indices of oxidative stress or damage to proteins in vivo. To determine the extent to which oxidation of nonenzymatically glycosylated proteins contributes to urinary CML, we measured the urinary concentrations of FL and CML in diabetic (n = 26) and control (n = 28) patients. The urinary concentration of FL correlated strongly with HbA1 measurements and was significantly higher in diabetic compared with control samples (9.2 +/- 6.5 and 4.0 +/- 2.8 micrograms/mg creatinine, respectively; P less than 0.0001). There was also a strong correlation between the concentrations of CML and FL in both diabetic and control urine (r = 0.67, P less than 0.0001) but only a weakly significant increase in the CML concentration in diabetic compared with control urine (1.2 +/- 0.5 and 1.0 +/- 0.3 micrograms/mg creatinine, respectively; P = 0.05). The molar ratio of CML to FL was significantly lower in diabetic compared with control patients (0.25 +/- 0.12 and 0.43 +/- 0.16, respectively; P less than 0.0001).(ABSTRACT TRUNCATED AT 250 WORDS)

The Maillard reaction in vivo

The Maillard or browning reaction between reducing sugars and protein contributes to the chemical deterioration and loss of nutritional value of proteins during food processing and storage. This article presents and discusses evidence that the Maillard reaction is also involved in the chemical aging of long-lived proteins in human tissues. While the concentration of the Amadori adduct of glucose to lens protein and skin collagen is relatively constant with age, products of sequential glycation and oxidation of protein, termed glycoxidation products, accumulate in these long-lived proteins with advancing age and at an accelerated rate in diabetes. Among these products are the chemically modified amino acids, N epsilon-(carboxymethyl)lysine (CML), N epsilon-(carboxymethyl)hydroxylysine (CMhL), and the fluorescent crosslink, pentosidine. While these glycoxidation products are present at only trace levels in tissue proteins, there is strong evidence for the presence of other browning products which remain to be characterized. Mechanisms for detoxifying reactive intermediates in the Maillard reaction and catabolism of extensively browned proteins are also discussed, along with recent approaches for therapeutic modulation of advanced stages of the Maillard reaction.

Differences between the catabolism and tumour distribution of intact monoclonal antibody (791T/36) and its Fab/c fragment in mice with tumour xenografts revealed by the use of a residualizing radiolabel (dilactitol-125I-tyramine) and autoradiography

Radioiodine-labelled 791T/36 monoclonal antibody (mAb) and its Fab/c fragment, consisting of one Fab arm and the Fc portion, have identical whole-body survival curves in BALB/c mice (t1/2 = 3.75 days). Therefore, these two forms of this antibody provide a suitable model for studying the role of valency in the targeting efficiency of antibodies to tumours in vivo. 791/T36 antibody and its Fab/c fragment were labelled either by direct iodination using the iodogen method (125I) or by dilactitol-125I-tyramine (125I-DLT), a residualizing label, which accumulates in the cells involved in degradation of the carrier protein. In tumour-bearing nude mice, the percentage of injected dose of mAb or Fab/c fragment reaching the specific 791T tumour was similar, and these proteins appeared to be catabolized at a similar rate in this tissue. mAb, but not the Fab/c fragment, was found to be very actively catabolized by the liver and spleen of tumour-bearing mice compared to control nude mice, this probably resulting from clearance of immune complexes. This effect was most pronounced when the mAb was labelled with 125I-DLT, the percentage of injected dose of mAb reaching the spleen and liver being higher than the percentage of injected dose reaching the tumour. This effect was not seen with the Fab/c fragment. Autoradiographic studies on tumour sections, which exhibit antigenic sites throughout the tumour mass, showed that the Fab/c fragment was already homogeneously distributed in the tumour 12 h after injection whereas the whole antibody was mainly localized at the periphery of the tumour. Those results suggest a "binding site barrier" effect. Overall, these results indicate that the highest valency and affinity may not be the optimal choice for mAb to be used for therapeutic purposes.

Reaction of ascorbate with lysine and protein under autoxidizing conditions: formation of N epsilon-(carboxymethyl)lysine by reaction between lysine and products of autoxidation of ascorbate

N epsilon-(Carboxymethyl)lysine (CML) has been identified as a product of oxidation of glucose adducts to protein in vitro and has been detected in human tissue proteins and urine [Ahmed, M. U., Thorpe, S. R., & Baynes, J. W. (1986) J. Biol. Chem. 261, 4889-4894; Dunn, J. A., Patrick, J. S., Thorpe, S. R., & Baynes, J. W. (1989) Biochemistry 28, 9464-9468]. In the present study we show that CML is also formed in reactions between ascorbate and lysine residues in model compounds and protein in vitro. The formation of CML from ascorbate and lysine proceeds spontaneously at physiological pH and temperature under air. Kinetic studies indicate that oxidation of ascorbic acid to dehydroascorbate is required. Threose and N epsilon-threuloselysine, the Amadori adduct of threose to lysine, were identified in the ascorbate reaction mixtures, suggesting that CML was formed by oxidative cleavage of N epsilon-threuloselysine. Support for this mechanism was obtained by identifying CML as a product of reaction between threose and lysine and by analysis of the relative rates of formation of threuloselysine and CML in reactions of ascorbate or threose with lysine. The detection of CML as a product of reaction of ascorbate and threose with lysine suggests that other sugars, in addition to glucose, may be sources of CML in proteins in vivo. The proposed mechanism for formation of CML from ascorbate is an example of autoxidative glycosylation of protein and suggests that CML may also be an indicator of autoxidative glycosylation of proteins in vivo.

Site of catabolism of autologous and heterologous IgA in non-human primates

Because of similarities between the human and monkey immune systems, we considered the monkey a suitable model for studies on the catabolism of various molecular forms of IgA, for which little information is available. The residualizing label dilactitol-[125I]tyramine was coupled to monkey (Macaca fuscata) IgA and IgG, as well as to human monomeric and polymeric myeloma IgA1 and IgA2 proteins. When labelled proteins were injected intravenously into monkeys, the non-metabolizable radioiodinated tracer accumulated at the cellular site of protein degradation, allowing identification of the catabolic sites. To determine the uptake of injected proteins by various tissues, monkeys were sacrificed 6-7 days after injection of labelled proteins, when blood-associated radioactivity was less than or equal to 10% of the injected dose, as measured by plasma clearance. When monkey or human monomeric IgA, as well as human polymeric IgA, irrespective of subclass, was administered to monkeys, the liver showed the greatest tissue uptake relative to total dose injected and to organ weight, and the highest acid soluble radioactivity (degraded protein). Although both hepatocytes and non-parenchymal liver cells were involved in IgA uptake, the hepatocytes were more active. Therefore, it appears that the liver is the major site of uptake and catabolism of IgA in monkeys and possibly in humans.

A fluorescent residualizing label for studies on protein uptake and catabolism in vivo and in vitro

Residualizing labels are tracers which remain in lysosomes after uptake and catabolism of the carrier protein and have been especially useful for studies on the sites of plasma protein degradation. Thus far these labels have contained radioactive reporters such as 3H or 125I. In the present paper we describe a fluorescent residualizing label, NN-dilactitol-N'-fluoresceinylethylenediamine (DLF). Modification of asialofetuin (ASF) or rat serum albumin (RSA) with DLF affected neither their normal kinetics of clearance from the rat circulation nor their normal tissue sites of uptake and degradation. After injection of DLF-ASF, fluorescent degradation products were recovered nearly quantitatively in liver and retained with a half-life of about 2 days. Fluorescent degradation products from DLF-RSA were recovered in skin and muscle, and were localized in fibroblasts by fluorescence microscopy. These results confirm previous studies with radioactive residualizing labels in which fibroblasts in peripheral tissues were identified as primary sites of albumin degradation. Fluorescent catabolites also accumulated in fibroblasts incubated with DLF-RSA in vitro, and residualized with a half-life of about 2 days. Overall, the data establish that DLF functions efficiently as a fluorescent residualizing label both in vivo and in vitro. The advantages of fluorescent, compared with radioactive, residualizing labels should make them valuable tools for studies on protein uptake and catabolism in biological systems.

Intracellular processing of residualizing labels in different cell types in vitro

In previous autoradiographic studies on the sites of catabolism of rat serum albumin (RSA) in the rat, fibroblasts in skin and muscle were shown to accumulate degradation product from RSA labeled with the residualizing label dilactitol-125I-tyramine (125I-DLT) (Strobel et al., 1986 J. Biol. Chem., 261:7989-7994). Residualizing labels remain at the cellular site of degradation of the carrier protein because of their size, hydrophilicity, and resistance to lysosomal hydrolases. This study was designed to evaluate whether fibroblasts might retain labeled degradation products more efficiently than other cell types. The uptake of 125I-DLT-RSA and release of its degradation products and of a second non-biodegradable probe, fluorescein isothiocyanate (FITC)-dextran, were studied in fibroblasts, endothelial cells, and macrophages, all cell types previously implicated in the catabolism of albumin in vivo. The rates of uptake of labeled protein and dextran were comparable in all cell types and consistent with fluid phase endocytosis. The rate of release of both intact protein (30-35% of total radioactivity released) and radioactively labeled degradation products followed similar kinetics and had half-lives ranging from 26 to 37 hr. The rate of release of FITC-dextran was slower than that of radioactivity, with a half-life of 42-125 hr. Thus, although there were differences between the rates of release of the fluorescent and radioactive materials in vitro, there were no significant differences in the disposition of protein-derived catabolites among these three cell types.

Nonenzymatic glycosylation of protein does not increase with age in normal human lenses

Nonenzymatic glycosylation or glycation is a posttranslational modification of proteins which has been implicated in the aging of lens proteins and the development of senile and diabetic cataracts. The extent of glycation of normal human lens proteins was measured by reduction of the protein with [3H]NaBH4, acid hydrolysis and quantitation of radioactive hexitol-amino acids by phenylboronic acid (PBA) affinity chromatography. Hexitollysine (HL) accounted for greater than or equal to 90% of total radioactivity recovered as hexitol-amino acids (HAA). In lenses in the age range (1-79) years (n = 26) there was no significant age-dependent increase in glycation of lens proteins (p greater than .10). The average extent of glycation was 2.3 +/- 0.3 mmol glycated lysine/mol lysine, or approximately 0.8 nmol hexitollysine/mg lens protein. These results indicate that the extent of glycation of lysine residues in lens proteins is comparable to that of lysine residues in soluble proteins, such as hemoglobin and albumin, and that the extent of glycation of lens proteins does not increase with age. Thus, glycation, per se, is not an age-dependent chemical modification of human lens protein.

Non-invasive detection of protein metabolism in vivo by n.m.r. spectroscopy. Application of a novel 19F-containing residualizing label

Protein residualizing labels facilitate localization of tissue sites of protein catabolism and the quantification of protein accumulation because of their prolonged intracellular retention of protein accumulation because of their prolonged intracellular retention times. Radioiodinated residualizing labels have been used to define the metabolism of a wide variety of proteins, but this has necessitated destructive analysis. Here we describe the implementation and validation of a novel 19F-containing residualizing label for protein, NN-dilactitol-3,5-bis(trifluoromethyl)benzylamine (DLBA), that permits the non-invasive assessment of protein accumulation and catabolism by n.m.r. spectroscopy in vivo. DLBA comprises a reporter molecule containing six equivalent 19F atoms. 19F is strongly n.m.r.-active, has 100% natural abundance, and is present in minimal background concentrations in soft tissues. We validated the use of DLBA as a protein-labelling compound by coupling to asialofetuin (ASF), a protein that is recognized exclusively by hepatic tissue via a saturable receptor-mediated process. Coupling of DLBA to ASF by reductive amination had no effect on the physiological receptor-mediated uptake of the protein in rat liver in vivo. The 19F-n.m.r. spectrum of DLBA exhibited a single peak that was subject to a small chemical-shift change and broadening after coupling to ASF. Pronase digestion of DLBA-ASF was performed to simulate intracellular degradation products, and resulted in a narrower set of resonances, with chemical shifts intermediate between those of uncoupled DLBA and DLBA-ASF. Intravenous administration of DLBA-ASF to rats followed by quantification of 19F in homogenates of liver tissue indicated that the half-life of residence time of degradation products from DLBA-ASF in liver was approx. 2 days. This intracellular half-life was comparable with that described for similar residualizing labels that contain radioiodide as a reporter. Similar results for the half-life of retention were obtained non-destructively and non-invasively in situ with the use of a whole-body radio-frequency antenna to acquire sequential spectra over 80 h after intravenous administration of DLBA-ASF. Quantification of these spectra demonstrated an initial accumulation of DLBA-ASF in liver followed by an expected gradual loss of 19F-labelled degradation products. The approach developed offers promise for the sequential and longitudinal characterization of metabolism of specific proteins in individual experimental animals and ultimately in human subjects.

Oxidation of glycated proteins: age-dependent accumulation of N epsilon-(carboxymethyl)lysine in lens proteins

N epsilon-(Carboxymethyl)lysine (CML) has been identified as a product of oxidation of fructoselysine (FL) in glycated (nonenzymatically glycosylated) proteins in vitro and has also been detected in human tissues and urine [Ahmed et al. (1986) J. Biol. Chem. 261, 4889-4894]. In this study, we compare the amounts of CML and FL in normal human lens proteins, aged 0-79 years, using specific and sensitive assays based on selected ion monitoring gas chromatography-mass spectrometry. Our results indicate that the lens content of FL increases significantly between infancy and about age 5 but that there is only a slight, statistically insignificant increase in FL between age 5 and 80 (mean +/- SD = 1.4 +/- 0.4 mmol of FL/mol of Lys). In contrast, the lens content of the oxidation product, CML, increased linearly with age, ranging from trace levels at infancy up to 8 mmol of CML/mol of lysine at age 79. The ratio of CML to FL also increased linearly from 0.5 to 5 mol of CML/mol of FL between age 1 and 79, respectively. These results indicate that CML, rather than FL, is the major product of glycation detectable in adult human lens protein. The age-dependent accumulation of CML in lens protein indicates that products of both glycation and oxidation accumulate in the lens with age, while the constant rate of accumulation of CML in lens with age argues against an age-dependent decline in free radical defense mechanisms in this tissue.

The role of the liver in catabolism of mouse and human IgA

The fate of intravascular IgA which is produced in large quantities in humans and many animal species was investigated in vivo and in vitro with emphasis on the monomeric form of IgA. The site(s) of the catabolism of intravenously injected mouse monomeric IgA labeled with a residualizing tracer (dilactitol - 125I tyramine) was studied in mice. The greatest in vivo uptake of monomeric IgA was observed in the liver. In contrast to identically labeled IgG, liver accounted for more internal catabolism of monomeric IgA than all other tissues (spleen, muscle, skin, and kidney) combined. Although both parenchymal and nonparenchymal liver cells internalized monomeric IgA, hepatocytes were far more active. The uptake of monomeric IgA was primarily mediated by the asialoglycoprotein receptor. In humans, the particulate fraction of liver homogenates and a human hepatoma cell line (Hep G2) bound human IgA proteins of various molecular forms. Inhibition of the binding by desialylated glycoproteins, requirement for the presence of calcium, and the molecular properties of the IgA-binding protein from the plasma membrane of Hep G2 cells indicated that the binding was primarily mediated by the asialoglycoprotein receptor. IgA proteins bound by Hep/G2 cells were internalized and catabolized to low molecular weight fragments.

Inulin-125I-tyramine, an improved residualizing label for studies on sites of catabolism of circulating proteins

Residualizing labels for protein, such as dilactitol-125I-tyramine (125I-DLT) and cellobiitol-125I-tyramine, have been used to identify the tissue and cellular sites of catabolism of long-lived plasma proteins, such as albumin, immunoglobulins, and lipoproteins. The radioactive degradation products formed from labeled proteins are relatively large, hydrophilic, resistant to lysosomal hydrolases, and accumulate in lysosomes in the cells involved in degradation of the carrier protein. However, the gradual loss of the catabolites from cells (t1/2 approximately 2 days) has limited the usefulness of residualizing labels in studies on longer lived proteins. We describe here a higher molecular weight (Mr approximately 5000), more efficient residualizing glycoconjugate label, inulin-125I-tyramine (125I-InTn). Attachment of 125I-InTn had no effect on the plasma half-life or tissue sites of catabolism of asialofetuin, fetuin, or rat serum albumin in the rat. The half-life for hepatic retention of degradation products from 125I-InTn-labeled asialofetuin was 5 days, compared to 2.3 days for 125I-DLT-labeled asialofetuin. The whole body half-lives for radioactivity from 125I-InTn-, 125I-DLT-, and 125I-labeled rat serum albumin were 7.5, 4.3, and 2.2 days, respectively. The tissue distribution of degradation products from 125I-InTn-labeled proteins agreed with results of previous studies using 125I-DLT, except that a greater fraction of total degradation products was recovered in tissues. Kinetic analyses indicated that the average half-life for retention of 125I-InTn degradation products in tissues is approximately 5 days and suggested that in vivo there are both slow and rapid routes for release of degradation products from cells. Overall, these experiments indicate that 125I-InTn should provide greater sensitivity and more accurate quantitative information on the sites of catabolism of long-lived circulating proteins in vivo.

The sites of catabolism of murine monomeric IgA

The tissue sites of monomeric IgA (mIgA) catabolism were determined in a BALB/c mouse model. Mouse mIgA myeloma proteins were labeled either by direct iodination or by coupling the residualizing label, dilactitol-125I-tyramine (125I-DLT) to the proteins; catabolites from protein labeled with 125I-DLT accumulate at the site of protein degradation, allowing identification of the tissue and cellular sites involved in catabolism of the protein. The circulating half-lives of 125I- and 125I-DLT-mIgA were the same. The distribution of radioactivity in tissues was measured at 1, 3, 24, and 96 h after iv. injection of 125I-DLT-labeled mIgA, dimeric IgA (dIgA), IgG, or mouse serum albumin. The greatest uptake of 125I-DLT-mIgA was attributable to the liver. This organ accounted for more internal catabolism of mIgA than all other tissues combined. In contrast, 125I-DLT-IgG was catabolized equally in skin, muscle, and liver. These data indicate that, in mice, the liver is the major site of mIgA catabolism. To determine the cell types involved, collagenase digestion was used to isolate parenchymal and non-parenchymal cells from perfused liver of animals injected with 125-DLT-mIgA. Most of the radioactivity was associated with the hepatocyte fraction, even though both cell types showed uptake of 125I-DLT-mIgA. Inhibition studies, with asialofetuin and mouse IgA demonstrated that the uptake of mIgA by liver cells was mediated primarily by the asialoglycoprotein receptor.

The sites of catabolism of murine monomeric IgA

The tissue sites of monomeric IgA (mIgA) catabolism were determined in a BALB/c mouse model. Mouse mIgA myeloma proteins were labeled either by direct iodination or by coupling the residualizing label, dilactitol-125I-tyramine (125I-DLT) to the proteins; catabolites from protein labeled with 125I-DLT accumulate at the site of protein degradation, allowing identification of the tissue and cellular sites involved in catabolism of the protein. The circulating half-lives of 125I- and 125I-DLT-mIgA were the same. The distribution of radioactivity in tissues was measured at 1, 3, 24, and 96 h after iv. injection of 125I-DLT-labeled mIgA, dimeric IgA (dIgA), IgG, or mouse serum albumin. The greatest uptake of 125I-DLT-mIgA was attributable to the liver. This organ accounted for more internal catabolism of mIgA than all other tissues combined. In contrast, 125I-DLT-IgG was catabolized equally in skin, muscle, and liver. These data indicate that, in mice, the liver is the major site of mIgA catabolism. To determine the cell types involved, collagenase digestion was used to isolate parenchymal and non-parenchymal cells from perfused liver of animals injected with 125-DLT-mIgA. Most of the radioactivity was associated with the hepatocyte fraction, even though both cell types showed uptake of 125I-DLT-mIgA. Inhibition studies, with asialofetuin and mouse IgA demonstrated that the uptake of mIgA by liver cells was mediated primarily by the asialoglycoprotein receptor.

Oxidative degradation of glucose adducts to protein. Formation of 3-(N epsilon-lysino)-lactic acid from model compounds and glycated proteins

The chemistry of Maillard or browning reactions of glycated proteins is being studied in model systems in vitro in order to characterize potential reaction pathways and products in biological systems. In previous work with the Amadori rearrangement product N alpha-formyl-N epsilon-fructoselysine (fFL), an analog of glycated lysine residues in proteins, we showed that fFL was oxidatively cleaved between C-2 and C-3 of the carbohydrate chain to yield N epsilon-carboxymethyllysine (CML) and D-erythronic acid. We then detected CML in proteins glycated in vitro, as well as in human lens proteins and collagen in vivo (Ahmed, M. U., Thorpe, S. R., and Baynes, J. W. (1986) J. Biol. Chem. 261, 4889-4894). This work provided an explanation for the origin of CML in human urine and evidence for non-browning pathways of the Maillard reaction in vivo. In this report we describe the identification of a second set of products resulting from oxidative cleavage of fFL between C-3 and C-4 of the sugar chain, i.e. 3-(N epsilon-lysino)-lactic acid (LL) and D-glyceric acid. The formation of LL from fFL was increased at slightly acid pH, representing about 30% of the yield of CML at pH 6.4, compared with 4% at pH 7.4 in phosphate buffer. By gas chromatography-mass spectroscopy, LL was detected in proteins glycated in vitro and then identified as a natural product in human lens proteins and urine. Our results indicate that oxidative degradation of Amadori adducts to proteins occurs in vivo, leading to formation and excretion of CML and LL. These non-browning pathways for reaction of Amadori compounds may be physiologically relevant mechanisms for averting potentially damaging consequences of the Maillard reaction.

Purification of residualizing glycoconjugate labels for protein by reversed-phase high-pressure liquid chromatography

Residualizing labels are radioactive or fluorescent tracers used for identifying the tissue and cellular sites in which circulating proteins are catabolized in the body. When attached to protein the labels do not affect normal mechanisms of protein catabolism, but remain at the cellular site of protein uptake, after the carrier protein itself is degraded to diffusible catabolites. Until recently these labels consisted of biologically indigestible carbohydrates attached to a radioactive reporter molecule. In this report we describe the synthesis and purification of a new fluorescent residualizing label, N,N-dilactitol-N'-fluoresceinyl-ethylenediamine. The label is prepared by first derivatizing ethylenediamine 1:1 with fluorescein isothiocyanate and then coupling lactose to the remaining primary amino group by reductive amination. A rapid one step purification of this and other glycoconjugate labels by reversed-phase high-pressure liquid chromatography is described.

Effect of phosphate on the kinetics and specificity of glycation of protein

The glycation (nonenzymatic glycosylation) of several proteins was studied in various buffers in order to assess the effects of buffering ions on the kinetics and specificity of glycation of protein. Incubation of RNase with glucose in phosphate buffer resulted in inactivation of the enzyme because of preferential modification of lysine residues in or near the active site. In contrast, in the cationic buffers, 3-(N-morpholino)propane-sulfonic acid and 3-(N-tris(hydroxymethyl)methyl-amino)-2-hydroxypropanesulfonic acid, the kinetics of glycation of RNase were decreased 2- to 3-fold, there was a decrease in glycation of active site versus peripheral lysines, and the enzyme was resistant to inactivation by glucose. The extent of Schiff base formation on RNAse was comparable in the three buffers, suggesting that phosphate, bound in the active site of RNase, catalyzed the Amadori rearrangement at active site lysines, leading to the enhanced rate of inactivation of the enzyme. Phosphate catalysis of glycation was concentration-dependent and could be mimicked by arsenate. Phosphate also stimulated the rate of glycation of other proteins, such as lysozyme, cytochrome c, albumin, and hemoglobin. As with RNase, phosphate affected the specificity of glycation of hemoglobin, resulting in increased glycation of amino-terminal valine versus intrachain lysine residues. 2,3-Diphosphoglycerate exerted similar effects on the glycation of hemoglobin, suggesting that inorganic and organic phosphates may play an important role in determining the kinetics and specificity of glycation of hemoglobin in the red cell. Overall, these studies establish that buffering ions or ligands can exert significant effects on the kinetics and specificity of glycation of proteins.

Identification of fibroblasts as a major site of albumin catabolism in peripheral tissues

Rat serum albumin has been labeled with dilactitol-125I-tyramine, (125I-DLT) a radioactive tracer which remains entrapped within lysosomes following cellular uptake and degradation of the carrier protein. Similar kinetics of clearance from the rat circulation were observed for albumin labeled conventionally with 125I or 125I-DLT-albumin, both proteins having circulating half-lives of approximately 2.2 days. In contrast, the recovery of whole body radioactivity had half-lives of approximately 2.2 and 5.1 days, respectively, for the two protein preparations, indicating substantial retention of degradation products derived from catabolism of 125I-DLT-albumin. Measurement of total and acid-soluble radioactivity in tissues 2 or 4 days after injection of 125I-DLT-albumin revealed that skin and muscle accounted for the largest fraction (50-60%) of degradation products in the body. Fibroblasts were identified by autoradiography as the major cell type containing radioactive degradation products in skin and muscle. Fibroblasts were isolated from skin by collagenase digestion, followed by density gradient centrifugation. The amount of acid-soluble radioactivity recovered in these cells was in excellent agreement with that predicted based on acid precipitation of solubilized whole skin preparations. These studies demonstrate for the first time that fibroblasts are a major cell type involved in the degradation of albumin in vivo.