Identification of Amadori-Modified Plasma Proteins in Type 2 Diabetes and the Effect of Short-Term Intensive Insulin Treatment

Eight type 2 diabetic patients with an average diabetes duration of 8.5 years (range 3–19) were studied and compared with eight weight-matched and eight lean nondiabetic subjects without any first-degree relatives with diabetes. The subject data were previously reported in another aspect of this study (31) and are summarized in Table 1. The study protocol was approved by the institutional review board of the Mayo Clinic and Foundation.

Type 2 diabetic patients were studied on two separate occasions, one after the withdrawal of all antidiabetes treatments for 2 weeks (T₂D−) and another occasion following 10 days of four daily injections of regular insulin with an aim to maintain premeal blood glucose values of 80–100 mg/dl (T₂D+). The subjects were randomly assigned to the order of the treatment regimen, with a washout period of 8–10 weeks between regimens. The subjects were admitted to the general clinical research center for studies described elsewhere (31). During the T₂D− phase, the subjects were given an overnight intravenous saline infusion, while during the T₂D+ phase, they were infused with insulin intravenously overnight to maintain their blood glucose between 80 and 100 mg/dl. While on either insulin or saline infusion, blood samples were drawn at 7:00 a.m. after an overnight fast.

Glucose was measured using a Beckman Glucose Analyzer 2 (Beckman Instruments, Fullerton, CA). The plasma concentration of insulin was measured using a two-site immunoenzymatic assay (Linco Research, St. Louis, MO). The serum fructosamine was measured at the central clinical laboratory of the Mayo Clinic using a Hitachi 911 fructosamine calorimetric rate reaction. The intravenous glucose tolerance test was performed for the determination of insulin sensitivity and acute response to glucose (32,33).

A total of 0.25 μl of plasma from each subject was loaded onto 10% SDS-PAGE and run under reducing conditions in a Criterion gel apparatus (Biorad Laboratories, Hercules, CA) at 100 V constant until the dye front reached the bottom of the gel. The gels were stained with Coomassie blue for proteomic analysis. For the Western blot, the plasma proteins in the gel were transferred to a polyvinylidine fluoride membrane using a semidry blot (BioRad Laboratories) at 17 V constant for 70 min. After the transfer, the membrane was blocked with 5% defatted milk. The membrane was incubated in a 1:10 dilution culture filtrate of monoclonal IgM antibody (HB-9644; American Type Culture Collection [ATCC], Manassas, VA) raised against glucose-derived Amadori-modified lysine residues (1-deoxyfructosyl lysine). The membrane was subsequently incubated with a rabbit anti-mouse peroxidase-conjugated secondary antibody at a concentration of 1:10,000. Detection was accomplished using an enhanced chemoluminescence plus reagent kit (Amersham Bioscience, Piscataway, NJ). All incubations were performed for 60 min on a rocker in 0.05M Tris-buffered saline pH 7.5 with 0.05% Tween 20.

Plasma samples from the T₂D− and T₂D+ phases of a type 2 diabetic subject were subjected to depletion of albumin and globulin, respectively, using SwellGel Blue and NAb Protein-A Spin Chromatography kits (both from Pierce Biotechnology, Rockford, IL). All steps were performed according to the manufacturer’s instructions at equal volumes and conditions.

Protein samples of 0.5 μl of total plasma and 10 μl of depleted plasma from the T₂D− and T₂D+ phases of a diabetic patient were dissolved in 125 μl immobilized pH gradient (IPG) rehydration buffer containing 9M urea, 4% CHAPS (3-[(3-cholamidopropyl)dimethylammonio]-1-propane sulfonate), 65 mmol/l dithiothreitol, 0.2% 3–10 pH Biolyte (Ampholyte solution from Biorad), and a trace of Bromophenol blue. IPG strips that were 7 cm and 3–10 pH and 4–7 pH were allowed to rehydrate with the above sample mixture for 12–16 h at room temperature in a rehydration/reswelling tray. The rehydrated IPG strips were subjected to isoelectric focusing in a protean isoelectrofocusing (IEF) cell (Biorad Laboratories, Hercules, CA) using a three-step protocol with a maximum of 4,000 V increased from 300 V for 3 h in the second step and a third step continuing at 4,000 V for 24,000 V hr. The cell temperature was kept at 20°C, with a maximum current of 50 μA/strip. The IEF-completed IPG strips were equilibrated for the SDS-PAGE in a two-step equilibration using 2.5 ml of equilibration buffer (6M urea, 2% SDS, 0.375 mol/l Tris-HCl pH 8.8, and 20% glycerol), with 130 mmol/l dithiothreitol in the first step and 135 mmol/l iodoacetamide in the second step. The equilibration steps were done in a rehydration tray for 10 min each on a rotatory shaker at room temperature. The second-dimension separation by sub unit molecular weight was performed by vertical SDS-PAGE. We used the mini-PROTEAN 3 cells and 10% two-dimensional prep precast gels (both from BioRad) for the second dimension of the 7-cm IPG strips. The strips were loaded onto the IPG well of the gel with the help of molten agarose and run at 100 V constant until the dye front reached the bottom of the gel. The gels were either stained with Coomassie blue for proteomic analysis or subjected to Western blot analysis, as described previously.

The gel bands and spots that reacted with the antibody HB-9644 were excised for mass spectrometry analysis. The gel pieces were destained, subjected to reduction and alkylation, and trypsin digested, and the resulting peptides were extracted. The peptides were analyzed by nano-LC/MS/MS (liquid chromatography/mass spectrometry/mass spectrometry) using an ultimate cap-LC system (LC Packings, San Francisco, CA) directly coupled with an API QSTAR PulsarTM hybrid quadrupole time-of-flight LC/MS/MS mass spectrometer (Applied Biosystems, MDS Sciex, Foster City, CA). Data were processed with the software program Pro ID (Applied Biosystems) for protein identification. Pro ID operates by first preparing a formatted database from a protein or DNA FASTA file. At run time, the precursor molecular weight (MW) and fragment masses from each experimental MS/MS spectrum were compared with the database fragment masses for all peptides with the same precursor MW. For this study, the SWISS-PROT database was searched.

Values are expressed as means ± SE. A two-tailed unpaired t test was performed for comparisons among the groups. A paired t test was performed to determine the effect of insulin treatment on nonenzymatic glycation of plasma proteins in people with type 2 diabetes.

The lean control subjects had lower BMIs and total abdominal fat than the type 2 diabetic patients and the weight-matched nondiabetic control subjects, whereas the latter two groups had similar body composition and were similar in all other aspects (Table 1).

The plasma glucose levels were higher in the T₂D− (Table 1) than T₂D+ group as well as the control groups. Both nondiabetic control groups had similar glucose values. Plasma insulin was significantly higher in the T₂D+ group than in the T₂D− group and the two control groups. The weight-matched nondiabetic control subjects tended to have higher insulin levels than the lean control and T₂D− subjects, but the differences did not reach statistical significance. The plasma insulin levels in the T₂D− and lean control groups were similar. The insulin sensitivity of the control groups was significantly higher than the T₂D− group, as previously reported (31). The serum fructosamine levels were significantly higher in the T₂D− than the control groups and showed a significant reduction (13%) after 10 days of insulin treatment (Table 1).

The proteins that showed a reduction in the reaction with the anti-Amadori antibody upon insulin treatment were identified. From the single dimension gel (Fig. 1), the identified proteins were transferrin, serum albumin, Ig heavy chain constant region, fibrin, and Ig light chains. The additional proteins identified from the two-dimensional gel (Fig. 2) were α-fibrinogen precursor, β-fibrinogen precursor, fibrinogen γ-B chain precursor, transferrin, hemopexin, vitamin D binding protein, and serine protease inhibitor (α-1 antitrypsin precursor).

The protein identification from each band or spot was based on the highest score obtained with the highest confidence and maximum number of unique peptides identified for each protein. Mass spectrometry results from the single dimension SDS-PAGE gel bands showed other proteins besides the major proteins identified from each band. For example, the Ig heavy chain band showed the presence of serum albumin as well. To confirm that it was Ig heavy chain that was glycated, we performed Western blot analysis of purified globulins from respective plasma samples (Fig. 3). The results were similar to those obtained when the Western blot was conducted on total plasma samples. The mass spectrometry results of the two-dimensional gel spots identified single proteins.

A typical immunoblot of representative samples from each group is shown in Fig. 1. An image analysis of the Western blots showed (Fig. 4) a significantly higher glycation of albumin, fibrin, transferrin, and Ig heavy chain in T₂D− compared with nondiabetic control groups, though there were no significant differences in glycation of plasma proteins between weight-matched control subjects and lean control subjects. All of these proteins showed a significant reduction in glycation after 10 days of insulin treatment. Serum albumin (Fig. 4A) showed a 23.2% (P < 0.02) reduction in glycation after insulin treatment, fibrin (Fig. 4B) 34.6% (P < 0.003), Ig heavy chain constant (Fig. 4C) 20.7% (P < 0.02), transferrin (Fig. 4D) 25.4% (P < 0.04), and Ig light chain (Fig. 4E) 13.0% (P < 0.02).

Western blot analysis of the two-dimensional gel was performed only for T₂D− and T₂D+ samples to determine whether additional proteins would react with the HB-9644 antibody and to determine the effect of insulin treatment on glycation. We observed a remarkable reduction in glycation (antibody reaction) for β-fibrinogen precursor, hemopexin, vitamin D binding protein, and serine protease inhibitor after insulin treatment. Complete absence of antibody reaction was observed for α-fibrinogen precursor, fibrinogen γ-B chain precursor, and transferrin after 11 days of insulin treatment (Fig. 2).

The current study applied Western blot analysis using an antibody that recognizes specific Amadori formation, 1-deoxyfructosyl lysine, to determine the extent of glycation in plasma proteins of type 2 diabetic patients on poor glycemic control. We found a significant difference in the glycation of many plasma proteins from type 2 diabetic patients following 2 weeks of poor glycemic control, when compared with weight-matched and lean control subjects. The Amadori product, the first stable product of the Maillard reactions, is quantitatively the most abundant glycated protein found in the plasma. Accumulated Amadori products may lead not only to AGE deposition in tissues but also to other pathophysiological consequences, such as activation of oxidative stress, protein aggregation and cross-linking, loss of receptor binding, and enzymatic or other physiological activities by glycated proteins (6,34). The prevention of formation of Amadori products or the removal of these modified proteins is potentially the most effective way to interrupt the glycation cascade and prevent the potential pathological consequences of glycation. The current study also demonstrated that this glycation is significantly reduced by 10 days of short-term intensive insulin treatment and examined this approach to identify molecules to use as a short-term index of glycemia.

HbA_1c, an Amadori product, has been used extensively since the 1970s as an index of blood glucose control in clinical practice and research (35). A correlation between glycated albumin, another Amadori-modified plasma protein, and diabetic nephropathy and other complications has been recently reported (23,24). The current study demonstrates that glycated albumin is increased in type 2 diabetic patients after 2 weeks of poor glycemic control. Besides glycated albumin, we also found a significant increase in the glycation of several plasma proteins including transferrin, Ig heavy chain constant, and fibrin in type 2 diabetic patients during poor glycemic control. Previous studies have demonstrated that nonenzymatic glycation of Igs leads to impairment in immunoreactivity due to glycation of the Fc fragment of Igs (36,37). Our data are in agreement with these previous results demonstrating significant glycation of these important proteins. A recent study by MALDI/MS (matrix-assisted laser description ionization/mass spectrometry) analysis conducted on plasma samples of 60 type 2 diabetic patients shows more glycation on the Fab fragments than in the Fc portion of the Igs (20). Though our results show no significant increase in the glycation of Ig light chains in the T₂D− group when compared with control subjects, a significant reduction in glycation was observed after 10 days of intensive insulin treatment. It is reported that nonenzymatic glycation reduces the susceptibility of fibrin to degradation by plasmin (38). Our results show that the glycation of fibrin is significantly increased during 2 weeks of poor glycemic control and showed maximum improvement on intensive insulin treatment. It remains to be determined whether nonenzymatic glycation could result in an increased half-life of fibrin, thus contributing to the hemostatic changes that occur in metabolic syndrome (39). The current study also shows a similar increase of glycation of transferrin during poor glycemic control. An in vitro study (40) demonstrated that nonenzymatic glycation of transferrin decreases the iron binding capacity by binding the iron ions loosely on the protein and also increasing the production of free oxygen radicals including O₂− and OH. The current study indicates potential pathological changes that occur in diabetes. The fructosamine assay produced results comparable to those of the results for the individual proteins. However, the percentage of the reduction of glycation observed in the fructosamine assay was much less than what is observed for the individual proteins, except Ig light chains. Fructosamine represents the glycation of the entire plasma protein. The current report shows the potential value of these newly identified plasma proteins for monitoring the effect of short-term glycemic changes in diabetic patients. In addition, the current approach will help in identifying new molecules that may be responsible for the pathogenesis of the complications of diabetes. In the current study, we performed Western blot analysis on the two-dimensional gel electrophoresis of depleted plasma samples to determine whether additional proteins could be identified as glycated during poor glycemic control in diabetes. We identified glycated allo-albumin, γ-fibrinogen precursor, α-fibrinogen precursor, hemopexin, vitamin D binding protein, β-fibrinogen precursor, and α-1 anti-trypsin, all of which have specific physiological functions and may contribute to the pathogenesis of complications in diabetes. We have also shown that glycation reactions in these proteins are either absent or markedly reduced by short-term intensive insulin treatment. Our data show that a combination of Western blot analysis using specific antibodies and two-dimensional gel electrophoresis approaches yields important information on protein modification related to pathophysiology that could be useful for the treatment of diabetes. A similar proteomic approach to profile glycated proteins can be used to identify molecules that may offer additional prognostic markers other than HbA_1c, specifically to detect the short-term effect of diabetes treatment. The disappearance of protein glycation depends on the in vivo half-life of individual proteins. It is also possible that nonenzymatic glycation may alter the half-lives of serum proteins. Hence, proteins with wide variability in their half-lives may be useful tools as markers of glycemic index.

The Diabetes Control and Complications Trial (41) has confirmed that intensive insulin treatment will significantly reduce the HbA_1c level of type 1 diabetic patients. That trial also demonstrated conclusively that intensive insulin treatment is associated with a substantial reduction in microvascular complications, such as retinopathy and nephropathy. Others have also shown that short-term glycemic control by aggressive intensive insulin therapy significantly improves insulin action and insulin secretion in patients with type 2 diabetes with secondary failure to oral agents (42,43). The current study demonstrated the potential for utilizing many proteins as markers of glycemic status. Glycated proteins with a long half-life may continue to exert a pathogenic effect even after normoglycemia is restored. Thus, it is important to determine the glycated proteins with close correlations to the pathogenesis of diabetes.

The current study represents a novel attempt to identify and quantify Amadori- modified plasma proteins in a group of very well-defined study subjects by using Western blot analysis in combination with two-dimensional gel electrophoresis and mass spectrometry. This approach may provide an opportunity for more focused investigations that may identify protein modifications responsible for pathogenesis of complications related to type 2 diabetes as well as to monitor insulin treatment.

We acknowledge the support of Maureen Bigelow, RN, and the nursing and nutrition staff of the Mayo General Clinical Research Center. The study is supported by Public Health Service grant R01DK41973, MO1RR00585, and the Mayo Foundation and David Murdock Dole Professorship (to K.S.N.).