INTRODUCTION

Maillard reaction (glycation) is thought to play a role in the pathogenesis of angiopathy in diabetes and aging process (1-3). The advanced stage of this reaction that leads to the formation of advanced glycation end products (AGEs)¹ is very complex due to several possible metabolic pathways. Despite this complexity, the structures of several AGEs have been recently described, such as that of pyrraline (4), pentosidine (5), cross-line (6), and pyrropyridine (7). Carboxymethyllysine is also formed by oxidation of Amadori products (8). Pyrraline is one of the AGEs derived from the reaction of glucose with the lysine amino group on proteins. Several investigators have demonstrated the preferential accumulation of AGEs in diabetic tissues (9-13). Plasma pyrraline levels are also elevated in diabetic rats and humans as determined by ELISA using antibody to pyrraline (14, 15). In addition, Porterootin et al. (16) demonstrated the existence of pyrraline in vivo using high performance liquid chromatography. They also showed that pyrraline in vivo could react with other amino acids on proteins to form cross-links (17). It was also demonstrated that pyrraline is found in vascular lesions of diabetic and elderly subjects using immunohistochemical techniques (15). Among several pathways forming pyrraline, highly reactive dicarbonyl compounds, such as 3-deoxyglucosone, were identified as precursors reacting with free amino groups to form pyrraline (14, 15). We recently reported that plasma 3-deoxyglucosone levels are elevated in diabetic rats using specific high performance liquid chromatography assay (18).

AGEs are also known to alter the structural and functional properties of proteins. Furthermore, a pathological role of the interaction between AGEs-modified proteins and cells for diabetic complications has been recently proposed. Several cell surface proteins are thought to recognize AGEs (19-21). Furthermore, interaction between AGEs and these proteins induces a variety of secondary effects. Vlassara et al. (22) demonstrated that macrophages secrete cytokines, such as tumor necrosis factor and interleukin-1, following the recognition of AGEs through specific receptors. They also showed the induction of insulin-like growth factor in human monocytes by AGEs-modified protein (23). Saishoji et al. (24) also suggested that AGEs-modified bovine serum albumin (BSA) stimulated the activity of urokinase-type plasminogen activator through the scavenger receptor on RAW 246.7 cell line. Schmidt et al. (25) recently reported that the recognition of AGEs by receptors for AGEs, so-called RAGE, leads to the expression of VCAM-1. The interaction between AGEs and cells in the blood or vascular tissue as mentioned above may consequently cause angiopathy. In fact, several lines of evidence have shown the existence of AGEs in vascular lesions (11, 12). However, the exact moiety of AGEs-modified protein recognized by the cells is not yet known.

In the present study, we investigated a pyrraline-modified protein that had been demonstrated to localize at thickened intima of arteriosclerotic lesion in the kidney of diabetic patients (15). We examined the effect of modification of albumin by pyrraline on the degradation of the protein by macrophage-like cell line, P388D₁ cells. This cell line was originally isolated by Dawe and Potter (26) from a methylcholanthrene-induced lymphoid neoplasm of a DBA/2 mouse. Subsequent investigation by Koren et al. (27) revealed that these cells have macrophage-like characteristics. Among these, the P388D₁ cells have phagocytic activity and are rich in lysosomal vesicles.

EXPERIMENTAL PROCEDURES

Murine P388D₁ macrophage-like cells were obtained from Dainippon Pharmaceutical Co. (Osaka, Japan), and maintained in RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 2 mM sodium pyruvate, and 50 µg/ml penicillin/streptomycin at 37 °C in a humidified atmosphere of 5% CO₂. The RPMI 1640 medium and supplements were purchased from Life Technologies, Inc. (Grand Island, NY). Only adherent cells were scraped and used for experiments. Bovine serum albumin (A-0281, Sigma), free of globuline and free fatty acids, was used in the present study. ¹²⁵I-BSA was purchased from ICN Radiochemicals (Irvine, CA). Fluorescamine was purchased from Molecular Probes, Inc. All other chemicals were of analytical grade and purchased from Wako Pure Chemical Industries (Osaka), unless otherwise noted.

Pyrraline-modified BSA (Pyr-BSA) was prepared using the carbodiimide coupling reaction as described previously (14, 15). Briefly, 10 mg of BSA was dissolved in 500 µl of deionized-distilled H₂O and mixed with 2 mg of caproyl pyrraline. To this solution 1.0 mg of N-hydroxysulfosuccinimide (Pierce Chemical Co.) and 30 mg of 1-cyclohexyl-3-(2-morpholinoethyl)carbodiimide metho-p-toluene sulfonate (Sigma) were consecutively added to a final volume of 1.0 ml. Following a reaction for 5 min, the mixture was dialyzed thoroughly against phosphate-buffered saline (PBS, pH 7.4). The extent of modification by pyrraline was estimated using a characteristic UV spectrum. ¹²⁵I-Pyr-BSA was also prepared by conjugating caproyl pyrraline onto ¹²⁵I-BSA, as mentioned above. Control BSA was prepared using the same procedure as above but without adding caproyl pyrraline.

P388D₁ cells were seeded onto 96-well plates at a concentration of 5 × 10⁴/ml in the medium described above, and grown at 37 °C in 5% CO₂ for 3 days to subconfluence. The cell monolayer was washed with ice-cold PBS and then incubated with ¹²⁵I-BSA or ¹²⁵I-Pyr-BSA at a ligand concentration of 0 to 50 µg/ml in PBS. After incubation at 37 °C for 4 h, each medium was saved for assessment of degraded protein. The cell monolayer was washed three times with 200 µl of cold PBS and removed from each well by dissolution in 100 µl of 0.1 N NaOH. The radioactivity of the aliquot of cell suspension was counted to determine the amount of internalized labeled BSA. Another aliquot was used to determine the content of cellular protein using the method of Bradford (28). The determination of the degraded protein was carried out using a modified method of Goldstein and Brown (29), which was designed originally for the estimation of degraded ¹²⁵I-LDL. Briefly, we mixed 100 µl of the saved cell-free medium with the same volume of 40% trichloroacetic acid and kept at 4 °C for 30 min to precipitate undegraded BSA. The precipitated material was removed by centrifugation. An aliquot (180 µl) of trichloroacetic acid-soluble fraction was mixed with 2 µl of 40% potassium iodide and 8 µl of 30% hydrogen peroxide, and kept at room temperature for 5 min. In the next step, we added 400 µl of chloroform and the mixture was kept again at room temperature for another 15 min after thorough mixing. Any free ¹²⁵I that had contaminated the ¹²⁵I-BSA preparation was extracted into chloroform layer at this step. Finally, we counted the radioactivity of the aqueous fraction, derived from the degradation of labeled BSA, using a -counter. The counts were converted to micrograms using specific activity and corrected by cell protein content. Each experiment was performed in duplicate. A representative set of data of each experiment are shown under "Results" (see below) since similar results were obtained for each set of experiments.

The concentration of control and pyrraline-modified BSA was adjusted to 200 µg/ml in 0.1 M phosphate buffer (pH 3.5). Cathepsin D (Sigma) was added to each BSA solution at a final concentration of 1.0 unit/mg BSA. Aliquots of 200 µl from each mixture were incubated at 37 °C for 30, 60, and 120 min. Undigested albumin was precipitated by adding 200 µl of 10% trichloroacetic acid and standing for 10 min at room temperature. After spinning down, the amount of digested albumin in the supernatant was determined by the method of Kirschbaum (30) using bicinchonic acid (Pierce). Digestibility was calculated as the ratio of peptide content in the trichloroacetic acid-soluble fraction to the original total albumin content. The data were expressed as mean ± S.D. of 10 samples. Welch and Student's t tests were used for statistical analysis. A p value less than 0.05 denoted statistical significance.

A lysosome-rich fraction of P388D₁ cells was prepared by sequential centrifugation according to the modified method of de Duve et al. (31). Briefly, cultured P388D₁ cells were washed three times with serum-free RPMI 1640 and adjusted to a concentration of 5 × 10⁷ cells/ml in 0.25 M sucrose containing 1.0 mM EDTA and 0.2 mM phenylmethylsulfonyl fluoride at 4 °C. The following fractionation steps were carried out at 0 °C. About 5 ml of the cell suspension was homogenized using a homogenizer and centrifuged at 500 × g for 12 min to separate crude nuclear fraction. The supernatant was centrifuged at 5,000 × g for 10 min, followed by a final centrifugation of the resultant supernatant at 14,000 × g for 30 min. The pellet containing lysosomes was reconstituted with 0.1% Triton X-100 in 0.1 M phosphate buffer (pH 3.5) containing 1.0 mM EDTA and 0.2 mM phenylmethylsulfonyl fluoride (Buffer A). The concentration of the lysosome-rich fraction was adjusted to 200 µg/ml with Buffer A, and used for the following experiments.

The concentration of control and pyrraline-modified BSA was adjusted to 200 µg/ml in 0.1 M phosphate buffer (pH 3.5). We then added 200 µl of freshly prepared lysosome-rich fraction to 2 ml of each BSA solution and prepared aliquots of 100 µl. These aliquots were divided into three groups and incubated at 37 °C for either 30, 60, or 120 min. Undigested BSA was removed by precipitation with 10% trichloroacetic acid. The concentration of peptides obtained by this digestion was smaller than that with cathepsin D. Therefore, we applied the fluorescamine assay (see below) since it was more sensitive in determining the digested peptide. The calculation method for digestibility was the same as above. The data were expressed as mean ± S.D. of six samples. Welch and Student's t tests were used for statistical analysis. A p value less than 0.05 denoted statistical significance.

Fluorescamine assay was performed as described elsewhere (32, 33). Briefly, 50 µl of each trichloroacetic acid-soluble fraction was placed in a glass tube and mixed with 1.85 ml of 0.5 M sodium borate buffer (pH 8.5). One hundred and fifty microliters of fluorescamine solution in acetone (30%, w/v) was dropped into the tube with vigorous mixing. Fluorescence measurement was carried out with excitation/emission at 390/475 nm. The fluorescence intensity from each sample was compared with that derived from leucine as a standard.

We assessed lysosomal function of the cells incubated with control or pyrraline-modified BSA by determining the activity of two representative lysosomal enzymes, acid phosphatase and -N-acetylglucosaminidase. The activity of acid phosphatase in the cells was assayed using p-nitrophenyl phosphate as a chromogenic substrate. We employed the method of Absolom (34) with some modification by using a microtiter plate as follows. Cultured P388D₁ cells were washed three times with serum-free RPMI 1640. The concentration of the cells was adjusted to 2 × 10⁶ cells/ml in serum-free RPMI 1640 containing control or pyrraline-modified BSA at a final concentration of 100 µg/ml. Aliquots of the cell suspension were incubated at 37 °C for 0, 1, 2, 3, or 4 h. After incubation, the cells were washed three times with serum-free RPMI 1640, and resuspended with 100 µl of the same medium. Each cell suspension was placed onto a well of the microtiter plate for a period of 2 h at 4 °C to allow attachment of the cells. After aspirating the medium, the attached cells were lysed with 30 µl of 0.1% Triton X-100 in 0.15 M NaCl solution. An aliquot of 10 µl was used for determining cell protein content, while another aliquot of 10 µl was transferred to another clean well of the microtiter plate to determine the activity of acid phosphatase. The same amount of authentic acid phosphatase (Boehringer Mannheim, Germany) was also placed on the wells at various concentrations as a standard. We also mixed 20 µl of 0.2 M acetate buffer (pH 5.0) in each well, followed by the addition of 24 mM p-nitrophenyl phosphate (Sigma) solution as a substrate. After incubation at 37 °C for 30 min, color development was induced by the addition of 100 µl of 0.2 M Na₂CO₃ solution. Absorbance of each well at 405 nm was measured by ELISA reader (Bio-Rad). Quantitation was performed in duplicate using a calibration curve obtained using standard solutions.

The activity of -N-acetylglucosaminidase was assayed using the method of Baggiolini (35) with some modification to allow the use of a microtiter plate. Preparation of the cells up to the step of preparing suspension with 30 µl of 0.1% Triton X-100 in 0.15 M NaCl solution was the same as above. An aliquot of 10 µl was used for determining cell protein content, while another aliquot of 10 µl was transferred to a clean well of microtiter plate to determine the activity of -N-acetylglucosaminidase. The same amount of authentic -N-acetylglucosaminidase (Sigma) was also placed on wells at various concentrations as a standard. We added 50 µl of 24 mM -nitrophenyl-N-acetyl--D-glucosamide (Sigma) solution and incubated the mixture at 37 °C for 30 min. Color development was induced by the addition of 100 µl of 0.2 M Na₂CO₃ solution. Absorbance of each well at 405 nm was measured by ELISA reader. Quantitation was performed in duplicate using a calibration curve obtained using standard solutions.

Human monocytes were prepared from blood samples of a healthy volunteer using Ficoll-Paque solution (Pharmacia, Piscataway, NJ) as described by Böyum (36) with slight modification. Briefly, we diluted 120 ml of heparinized whole blood with the same volume of PBS and prepared 8-ml aliquots. Each aliquot was carefully layered onto 4 ml of Ficoll-Paque in a sterile centrifuge tube. After centrifugation at 400 × g for 30 min at room temperature, the mononuclear cell layer was collected and washed with PBS. We then prepared 100 ml of cell suspension with RPMI 1640 containing 10% fetal calf serum and dispensed into 10 dishes, followed by incubation at 37 °C for 2 h under 5% CO₂. After removing floated cells by washing with serum-free RPMI 1640, the attached cells were harvested with PBS using a scraper, followed by centrifugation at 400 × g for 5 min. The cells were resuspended in 25 ml of RPMI 1640 containing 10% fetal calf serum and used for internalization and degradation assay according to the method described for P388D₁ cells.

RESULTS

Pyrraline modification of albumin was confirmed by its characteristic UV spectrum and immunoreactivity to monoclonal antibody against pyrraline with the immuno-dot blotting method (data not shown). We obtained two types of pyrraline-modified BSA, which were termed Pyr-1 and Pyr-2 for convenience. ¹²⁵I-BSA was also conjugated with pyrraline. We estimated the extent of modification induced by pyrraline in each preparation by molar extinction coefficient at 297 nm as summarized in Table I. The extent of pyrraline modification in Pyr-2 was almost twice as that in Pyr-1. Considering that BSA contains an -amino group and 59 -amino groups, 19.7% of amino groups in Pyr-1 were modified by pyrraline while 33.0% were modified in Pyr-2. On the other hand, the extent of pyrraline modification in ¹²⁵I-Pyr-BSA was 17.5%, a level comparable with that of Pyr-1. Control albumin also showed very slight modification from UV spectrum. It is possible that albumin obtained from any animal would be modified by small amounts of AGEs due to exposure to sugars in vivo.

Table I. Extent of pyrraline modification of BSA The extent of pyrraline modification in each preparation was calculated from the characteristic UV spectrum with molar extinction coefficient at 297 nm. BSA BSA:pyrraline ratio mol:mol Control BSA 1 :2.0 Pyr-1 1 :11.8 Pyr-2 1 :19.8 125I-BSA 1 :1.8 125I-Pyr-BSA 1 :10.5

The content of degraded albumin was not different between control and Pyr-BSA at low concentrations. However, when the ligand concentration was equal or exceeded 20 µg/ml, a significant suppression of pyrraline-modified BSA was observed. The degraded content of control BSA was 20.4 µg/mg cell protein at a ligand concentration of 50 µg/ml, while that of Pyr-BSA was 11.7 µg/mg cell protein (Fig. 1). Modification of BSA by pyrraline caused a significant increase in the content of accumulated BSA. The difference was marked at ligand concentrations 20 µg/ml. The internal content of control BSA was 2.4 µg/mg cell protein at a ligand concentration of 50 µg/ml, while that of Pyr-BSA was 4.0 µg/mg cell protein (Fig. 2).

The amount of digested peptide derived from control BSA when incubated with cathepsin D for 120 min was 35.1 ± 1.8% of total albumin content. We compared the digestibility of other samples by expressing the mean value of control BSA at 120 min as 100%. Although albumin from all preparations was digested by cathepsin D with time, pyrraline-modified BSA showed a lower susceptibility to cathepsin D (Fig. 3). The digestibility of Pyr-1 (18.1 ± 2.3%) was significantly lower than control BSA (34.5 ± 7.1%, p < 0.001), even at 30 min incubation, and the difference was still observed at 120 min (56.0 ± 5.0 versus 100.0 ± 5.1%, p < 0.001). Pyr-2 also showed resistance to digestion by cathepsin D. The relative digestibility of Pyr-2 was different from the control at 30-120 min. Although the digestibility of Pyr-2 was significantly different from that of Pyr-1 at 30 and 120 min, the additional suppression from the level of Pyr-1 was not as much as the difference between control and Pyr-1 (Fig. 3).

The digestibility of control BSA when incubated with lysosome-rich fraction derived from P388D₁ cells for 120 min was 11.1 ± 1.5% of original albumin content. This mean value was used as 100% to compare the digestibility under other conditions. The results obtained with cathepsin D were almost reproduced when lysosome-rich fraction was used as shown in Fig. 4. The digestibility of Pyr-1 at 120 min (67.3 ± 14.7%) was significantly lower than the control (p < 0.005). Although Pyr-2 showed a higher resistance compared with control (56.9 ± 7.6% at 120 min, p < 0.001), the difference between Pyr-1 and Pyr-2 was not statistically significant.

We also investigated the activity of acid phosphatase and -N-acetylglucosaminidase in P388D₁ cells to examine the effect of pyrraline-modified BSA on lysosomal function of P388D₁ cells. As shown in Fig. 5, the activity of acid phosphatase in these cells after incubation with pyrraline-modified BSA was not significantly different from that of control BSA even after 4 h of incubation (67.9 versus 65.8 milliunits/mg cell protein at 4 h). Similarly, the difference in activity of -N-acetylglucosaminidase was not significant between cells incubated with control and pyrraline-modified BSA, although they tended to diminish slightly immediately after the commencement of incubation with either type of albumin (Fig. 6). The activities of -N-acetylglucosaminidase in P388D₁ cells incubated for 4 h with control and pyrraline-modified BSA were 65.4 and 63.7 milliunits/mg cell protein, respectively.

Activity of -N-acetylglucosaminidase of P388D₁ cells preincubated with pyrraline-modified BSA.

We examined whether the observed accumulation of pyrraline-modified BSA in P388D₁ cells was reproducible in human monocytes. The degradation of albumin was significantly suppressed in pyrraline-modified BSA compared with control BSA (6.0 versus 12.8 µg/mg cell protein at a ligand concentration of 80 µg/ml, Fig. 7). Consequently, the content of accumulated albumin increased when BSA was modified by pyrraline. The accumulated content of control BSA was 8.3 µg/mg cell protein at a ligand concentration of 80 µg/ml, while that of Pyr-BSA was 19.5 µg/mg cell protein (Fig. 8).

DISCUSSION

The present study was based on the recent demonstration of pyrraline in arteriosclerotic lesions and designed to examine the relationship defining the interaction between pyrraline-modified protein and blood cells, such as macrophages, with the progress of angiopathy. The deposition of pyrraline in arteriosclerotic lesions is somewhat similar to cholesterol deposition in atherosclerotic lesions. The latter is known to be associated with the uptake of chemically modified low density lipoprotein by scavenger receptor of monocytes/macrophages (37). Several receptors, including scavenger receptor and RAGE, are thought to recognize AGEs-modified proteins (19-21). Vlassara et al. (19) reported that the uptake and degradation of AGE-modified BSA is achieved thorough a high-affinity receptor on mouse macrophages. The removal of AGE-modified protein by such mechanism is an attractive hypothesis of tissue remodeling. However, accumulation of AGEs-modified BSA was still present in the cells examined by Vlassara et al. (19), presumably due to an inefficient degradation compared with albumin uptake. Furthermore, these investigators also reported accumulation of AGEs-modified nerve myelin in macrophages and proposed that the interaction between AGEs-modified nerve protein and macrophages may initiate demyelination in diabetes (38). In this regard, the exact structure of AGEs moiety recognized by the receptor is still unknown. Furthermore, the exact effect of AGEs modification is complicated by the presence of heterogeneous AGEs. Therefore, we focused in the present study on pyrraline-modified protein as a ligand, since its structure had been identified. Using pyrraline as a model of AGEs, estimation of the involvement of the modification became more accessible. The P388D₁ used in this study is an established macrophage-like cell line with phagocytic activity. Westwood et al. (39) recently showed the existence of receptor recognizing methylglyoxal-modified protein on P388D₁ cell surface. Although we attempted to identify a high-affinity cell surface receptor against pyrraline, we were unable to find such receptor in the range of ligand concentration examined (data not shown). Since P388D₁ cells are able to phagocyte albumin themselves, pyrraline-modified BSA may not be necessarily recognized at the pyrraline moiety. However, the possible existence of a lower-affinity receptor cannot be excluded completely at present.

Our results showed that pyrraline-modified albumin was resistant to degradation by the cells. This finding suggests that pyrraline-modified albumin may have affected lysosomal proteolytic function of P388D₁ cells. Alternatively, the susceptibility of albumin to proteolysis was modified by pyrraline. To examine the first mechanism, we compared lysosomal function of cells preincubated with control and pyrraline-modified albumin. Our results showed that the activities of two representative lysosomal enzymes, acid phosphatase and -N-acetylglucosaminidase, were not significantly different. On the other hand, the susceptibility of pyrraline-modified albumin to enzymatic proteolysis by cathepsin D was significantly reduced in vitro. Albumin is known to be degraded mainly by cathepsin D and E in the lysosome (40). Since it is difficult to examine the susceptibility of pyrraline-modified albumin to all lysosomal proteolytic enzymes, we fractionated lysosome-rich fraction by sequential centrifugations. These studies showed that pyrraline-modified albumin was resistant to digestion by lysosome-rich fraction as well as cathepsin D alone. These findings are consistent with the well known fact that collagen diminishes its enzymatic digestibility as it is modified by AGEs (41, 42). Furthermore, the degree of pyrraline modification observed at approximately 10 mol/mol BSA was sufficient to reveal the resistance against enzymatic digestion. Although the reduced digestibility was dependent on the degree of modification, the modification induced by an initial 10 mol/mol BSA showed more suppressive effect than an additional 10 mol. Interestingly, the rate of reduction of digestibility of pyrraline-modified BSA by cathepsin D or lysosome-rich fraction in vitro was almost identical to that observed in the degradation of pyrraline-modified albumin in experiments using P388D₁ cells. Thus, we believe that the main cause of suppression of degradation by phagocytes is more likely to be a decrease in the susceptibility to lysosomal proteolytic digestion in P388D₁ cells. It is conceivable that the suppression of degradation leads to accumulation of pyrraline-modified albumin in the cells. In addition, the recent finding of pyrraline accumulation in lesions of Alzheimer disease, such as neurofibrillary tangles and senile plaques, suggests a potential role for AGEs, including pyrraline, in the protease-resistance property of the lesions (43).

The present results also showed accumulation of pyrraline-modified albumin in human monocytes as well as P388D₁ cell line. Plasma pyrraline levels were shown to be elevated in diabetic rats and humans as determined by ELISA using monoclonal antibody to pyrraline (15). Calculating the pyrraline modification per mol of albumin from the findings, some diabetic subject showed about 0.6 mol of pyrraline/mol. Diabetic rats, showing higher blood glucose levels, indicated more extensive modification such as 1.3 mol/mol. Since glycation does not occur uniformly to all albumin molecules, the degree of the modification in relatively longer-lived molecule may be more extensive than the average value described above. In fact, AGEs formed in albumin in vivo is expected to be very heterogeneous. In addition to pyrraline, pentosidine has been reported to increase in diabetic plasma (44, 45). Concerning other AGEs, Makita et al. (13) reported that the levels of AGE-modified serum protein from diabetic patients with renal failure were elevated 8-fold over normal subjects, when immunologically determined by using polyclonal antibody which did not cross-react with either pyrraline or pentosidine. Although structures of the AGEs epitope recognized by the antibody still remains to be identified, their findings support that serum protein is extensively modified by unknown AGEs as well as pyrraline in diabetic patients. Since the decrease in the susceptibility of pyrraline-modified albumin to lysosomal enzymes, observed in the present study, seems attributable to the covalent modification of the substrate albumin, it may be deduced that AGE modifications other than pyrraline also have similar inhibitory effects on enzymatic degradation of albumin. Kato et al. (46) showed that incubation of human serum albumin with glucose under physiological conditions (pH 7.4, 37 °C) in vitro caused the decrease in intact lysine residues by 19%, comparable to our pyrraline-modified albumin. Iberg and Flückiger (47) also indicated that 10 lysine residues were the most susceptible glycation sites by amino acid analysis of tryptic-digested albumin from diabetic patients. Lapolla et al. (48) more directly determined the extent of glycation in serum albumin from diabetic patients, using a matrix-assisted laser desorption ionization method. They proved that the number of modified residues in a mole of albumin was from 1.4 to 14.8. These findings support that the extent of modification of albumin in the present study is in the physiological range. These AGE-modified proteins in the circulation of diabetic patients may accumulate in monocytes/macrophages in vivo. Steinbrecher and Witztum (49) also demonstrated an inhibitory effect of glycation on the degradation of low-density lipoproteins in cultured fibroblasts. They also suggested that the inhibitory effect was due to a modification in lysine residues of apoprotein B. With respect to protein catabolism, the modification of lysine residues by AGEs may be more critical, since the alteration is irreversible. In fact, the levels of AGE in low-density lipoproteins of diabetic patients are also elevated, as reported by Bucala et al. (50). The exact nature of changes affecting the property of these cells following accumulation of AGE-modified protein is under investigation. Considering the fact that chemical modification of low-density lipoproteins is a trigger for atherosclerosis, modification of albumin by AGEs, such as pyrraline, may be also involved in vascular disorders. The present findings may help to elucidate the mechanisms involved in AGE modification of constitutively existing proteins in vivo causing a change in their metabolism, and subsequently, initiation of diabetic complications.