Nonenzymatic glycation of human blood platelet proteins

THROMBOSIS RESEARCH 55; 341-349, 1989 0049-3848/89 $3.00 t .OO Printed in the USA. Copyright (c) 1989 Pergamon Press plc. All rights reserved.

NONENZYMATIC GLYCATION OF HUMAN BLOODPLATELETPROTEINS

I. Cohen, D. Burk, R. J. Fullerton, A. Veis and D. Green Atherosclerosis Program, Rehabilitation Institute of Chicago and Departments of Molecular Biology, Oral Biology, and Medicine Northwestern University Medical School Chicago, Illinois, USA

(Received 8.3.1989; accepted in revised form 15.5.1989 by Editor N.U. Bang)

ABSTRACT all of uhom had severe athercthrombotic We studied 11 diabetic patients, disease, and 11 normal controls. Overall glycation was assessed by the extent of incorporation of C3H]-NaBH, into fructosyl lysine separated from whole platelet proteins following aminoacid analysis. Fructosyl lysine represented 5.7% + 1 .O S.D. of the total radioactivity in the normal whole platelet samples. Increased glycation was observed in platelets from 5 of the 11 diabetics. Platelet glycation did not correlate with glycation of hemoglobin or albumin. The pattern of glycation of various platelet proteins in whole platelets, as determined by the incorporation of [‘HI-NaBH, into electrophoretically separated proteins did not display selectivity, although myosin and glycoproteins IIb and IIIa showed relatively increased levels of [‘HI-NaBH, incorporation. Artificially glycated platelet membranes exhibited glycation mainly in proteins corresponding to the electrophoretic mobility of myosin, glycoproteins IIb and IIIa. INTRODUCTION Atherosclerosis and associated thromboembolic disorders are prevalent in These major complications are thought to be related, in diabetes mellitus (1). to enhanced platelet function. The increased sensitivity of platelets part, from poorly-controlled diabetics to aggregating agents and the increased binding of fibrinogen to activated platelets from diabetic patients have been correlated with thromboxane A2 formation and blood glucose levels (2, 3). In this report, we investigated the nonenzymation glycation of platelet proteins from diabetic platelets.

Key Words:

Platelets,

Diabetes,

Nonenzymatic Clycation 341

342

NONENZYMATIC PLATELET GLYCATION

Vol. 55, No. 3

Nonenzymatic glycation of proteins is a process whereby glucose reacts reversibly with protein amino groups from NH,-terminal residues or c-NH,-groups of lysine residues to yield a Schiff base. The aldimine subsequently rearranges in a slow process to a more stable ketoimine known as an Amadori product (4). Amadori products can then undergo further slow modifications resulting in the formation of proteins cross-linked by fluorescent 2-(2-furoyl)-4(5)-2-furanyl)1 H-imidazole molecules (5). The rate and intensity of these processes is dependent on glUCOSe levels and is amplified in patients with poorly-controlled diabetes mellitus. The discovery of hyperglycated hemoglobin as well as the glycation of other proteins such as albumin and collagen may have implications for the pathogenesis of complications in diabetes. The glycation of diabetic platelets has been reported (6). no correlation was found with glycated hemoglobin or albumin. made to identify the glycated platelet proteins.

In our study, An attempt was

MATERIALSANDMETHODS Patients

and Normal Donors

Patients were diabetics hospitalized at the Rehabilitation Institute of Every patient had a history of either a cerebral infarct or amputation Chicago. Controls were non-diabetic laboratory volunfor peripheral vascular disease. Informed consent was obtained from all subjects. teers. Preparation

of Platelet

Proteins

Fifty ml of venous blood was collected from the antecubital vein of normal or diabetic donors and mixed with 1 vol of 84 mM Na, citrate, 64.7 mM citric All platelet washing steps were acid, 111 mM dextrose for 6 vol of blood (7). Centrifugation at 125 g for 15 min yielded carried out at room temperature. The platelet suspension was centrifuged platelet-rich plasma with a pH of 6.5. at 1,000 x g and the pellet washed three times in a pH 6.5 buffer containing 107 mM NaCl, 3.75 mM KCl, I .6 mMNaHCO,, 21.2 mM sodium citrate, 27.7 mM at 120 x g interdextrose, 1 mMMgCl, and 1 mMEDTA. Low speed centrifugation calated at each step enabled us to obtain platelet pellets virtually free of red The washed platelets (5 x 10’ - 10”‘) were suspended blood cell contamination. 0.2% sodium dodecyl sulfate and 50 mM 20 pM leupeptin, in 1 ml buffer containing Na phosphate buffer pH 7.4. It was then sonicated for 10 pulses at 30 watts output (50% pulse) in a Heat Systems Ultrasonics Cell Disruptor, Model U225R. The disrupted platelet suspension was then clarified by centrifuging for 5 min at 125 x g. Control experiments exposure of platelets the washing procedure Measurement of Platelet

with platelets from non-diabetic volunteers showed that the to 27.7 mMdextrose for 40 min at room temperature during did not influence the degree of glycation. Clycation

The method described by Miller et al (8) was modified. 100 mCi [‘HI-NaBH, (New England Nuclear, 300-500 mCi/mmol) was dissolved immediately before use in 1 ml 0.01 N NaOH. It was added to 1 ml of sonicated platelets at a final concentration of 19 mMNaHH,. The final pH was 9.0 which is optimal for this reaction. A few glycation experiments performed in duplicate with the use of

Vol. 55, No. 3

NONENZYMATIC PLATELET GLYCATION

343

40 mM [“HI-NaBH, yielded the same degree of incorporation as 19 mM. The latter was therefore used in our experiments and considered as a large excess. Following 1 h incirbation at room temperature, the samples were extensively dialyzed The samples were then divided in two against 50 mMNa phosphate buffer pH 6.8. equal aliquots, A and B. One-fifth volume of 9% sodium dodecyl sulfate, 15% glycerol, 3 mMEDTA, 187.5 ?M ? Tris-HCl pH 6.8 was added to aliquot A a,nd the mixture was heated for 5 min at 1OO’C. Protein was measured by the bicinchoninic acid procedure (9). Samples were reduced with 25 mM dithiothreitol upon heating for 5 min at 100°C and then alkylated with 62.5 mMiodoacetamide upon incubation for 10 min at 60°C (10). Proteins (200 ug) were fractionated by SDSpolyacrylamide gel electrophoresis (SDS-PAGE) according to Laemmli’s procedure (11) and stained with Coomassie Brilliant Blue. One set of gels was “enhanced” for f luorography ( 12) then dried and exposed to Kodak XRP x-ray film at -7O’C for 2-4 weeks. Another set of gels was sliced. The slices, dried overnight at room temperature, were then dissolved in 0.5 ml of 30% H,O, at 60°C (13) and counted for ‘H following addition of 5 ml of RPI 3a70B complete counting cocktail (Mt. Prospect, IL). Similar counts were obtained in duplicate unstained gels. Aliquot B was hydrolyzed with 6N HCl for 22 h at 106’C. Following evaporation to dryness, each sample was diluted in an appropriate volume of 0.01 N HCl, depending on the protein concentration. 0.8 ml of each sample was then submitted to analysis in a JEOL Model JLC-6AH amino acid analyzer. A single column system with citrate elution buffers was used. Before ninhydrin reaction, the stream was split 50/50. The unreacted eluate was collected for counting. The standard of fructosyl lysine was prepared by incubating 50 mg of poly-Llysine (Sigma, MW-- 200,000) with 120 mM D-glucose for 6 days at 37’C in a buffer containing 0.02 g% NaN,, 0.135 M NaCl, and 5 mM Na phosphate buffer pH 7.4. Following extensive dialysis, glucose-reacted and glucose-free control samples were treated with CSHl-NaBH,, similarly to the procedure described for sonicated platelets. A sample of authentic fructosyl lysine, a gift from Dr. P. A. Finot (Nestle’, La Tour de Peilz, Switzerland), was also analyzed under the same conditions. Clycation

of Hemoglobin and Plasma Albumin

Clycated hemoglobin and plasma albumin were estimated by affinity binding to aminophenylboronic acid agarose gel (14, 15). 1 ml m-aminophenylbororiic acid agarose columns (Glycogel, Pierce Chemical Co., Rockford, IL), equilibrated with wash buffer (250 mM ammonium acetate, 50 mMMgCl,, 0.02 g% NaN,, pH 8.0), were used. 0.1 ml of 5% red cell hemolysate or 0.2 ml plasma was then applied. The non-bound hemoglobin and albumin were washed from the column and the bound hemoglobin was eluted with 5 ml of a buffer containing 200 mMsorbitol, 100 mMTris, 0.02% NaN, pH 8.5 and bound albumin was eluted with 3 ml sodium citrate pH 4.5. Hemoglobin was determined in the non-bound and bound fractions by measuring the absorbance at 414 nm. Albumin was measured at 630 nm following reaction with bromcresol green.

RESULTS The incorporation of [‘HI-NaBH, into the glucosyl-ketoamine linkage was to evaluate nonenzymatic glycation of platelet proteins in normal and diabetic patients. In order to validate this technique, glucose was incubated

used

344

NONENZYMATIC PLATELET GLYCATION

Vol. 55, No. 3

with poly-L-lysine. Amino acid analysis showed that the radioactive component of the hydrolyzed incubation mixture eluted at position 79 similarly to the authentic fructosyl lysine standard. When platelet sonicates were incubated with [‘HI-NaBH,, three major radioactive peaks were observed. The first major peak eluted at a position similar to Pructosyl lysine and represented 5.7% + 1 .O S.D. of the total radioactivity in normal samples. In a diabetic sample, this peak (fraction 76-82) represented 8.4% of the total sample radioactivity (Figure 1). The two unidentified major peaks (fractions 85-95) eluting after 7’ 6-

7 c

5-

i

4-

I

76-

82

s i 0

32-

20

40

79

60

FRACTION

88

100

120

NUMBER

FIGURE1 Incorporation of [‘HI-NaBH, into fructosyl lysine isolated by aminoacid analysis from hydrolysed platelets of a diabetic patient (patient 2, Table 1).

fructosyl lysine may not be due to glycation since they were labeled to the same These peaks may be related to the extent in the normal and diabetic samples. The reducing ability of BH, towards disulfide bridges (16) and lipids (8). Therefore, for minor peaks could represent glycated N-terminal aminoacids. only the first major peak coranalysis of the glycation of platelet proteins, Table 1 shows the individual responding to fructosyl lysine was considered. quantitative estimates of platelet protein glycation for 11 diabetic patients. Glycation was clearly increased in 5 (patients 1, 2, 7-91 of the 11 diabetics. We compared the_ values for glycated platelet proteins with the levels of glyCated hemoglobin and albumin, and found that although the values for the latter were clearly elevated, they did not correlate with the levels of the glycated platelet proteins. Next, we performed gel electrophoresis of the glycated platelet proteins to determine whether glycation preferentially occurred in specific prOteinS. Figure 2 shows the increased incorporation of [‘HI-NaBH, in proteins from two diabetic patients. Samples from diabetics with increased platelet glyCatiOn were selected for further study. Quantitation of the incorporation of [‘HIas compared with normals, showed NaBH, into the gel slices of these diabetics, an overall increased incorporation into platelet proteins by factors of 1.38 to It I.3 Figure 3 illustrates a representative pattern. 1.97, diabetic/normal.

TABLE 1.

PATIENTS

2 3 4 5 6 7 8 9 10 11

345

NONENZYMATIC PLATELET GLYCATION

Vol. 55, No. 3

DETERMINATION OF BLOOD GLUCOSE AND GLYCATION OF HEMOGLOBIN, ALBUMIN AND PLATELET PROTEINS FROM DIABETIC PATIENTS AND NORMAL CONTROLS.

BLOOD GLUCOSE,mg% 170 135 109 115 74 95 173 i20 287 280 119

Ccntrols(l1) 82.6 2 7.4* ** P < 0.02

GLYCATED Hb,% 13.9 11.2 10.5 10.0 9.8 9.6 8.9

GLYCATED ALBUMIN,%

['HI FRUCTOSYL LYSINE ‘$ OF TOTAL

5.7 2.6

7.9 8.4 5.1 6.8 5.8 5.6 7.9 7.8 7.8

2.2 3.6

2.6 1.6 1.2

8.1 8.0 7.2 6.7

::A 3.2 4.11

4.4 + 0.6"

0.9 + 0.2'

( 0.0001

< 0.0002

6.0 6.0 5.7

+ 1.0'"

< 0.03

*Standard Deviation **Using the student t test, P estimates are for all values of patients as compared to controls.

MTSP -

IIbma -

A-

L-

FIGURE 2 Incorporation of ['HI-NaBH, into electrophoreticallyseparated platelet proteins of normal controls and diabetic patients. 1 and 4 (1 and 2, respectively, Table 1) are from diabetic patients. Left panel, Coomassie Blue staining; Right panel, fluorography. M, myosin; TSP, Thrombospondin; IIb and IIIa, glycoproteins; A, actin; L, lipid.

346

NONENZYMATIC PLATELET GLYCATION

Vol. 55, No. 3

FIGURE3 Incorporation of [“HI-NaBH, into gel slices of electrophoretically separated platelet proteins of a normal control (M) and a diabetic patient (patient 2, Table 1) (e--o). Overall incorporation diabetic/normal is 1.97. although the bands which corresdifficult to see selectivity in these patterns, pond to the relative electrophoretic mobility of purified myosin and glycoproteins IIb and IIIa appear to have relatively increased levels of incorporation. When isolated platelet membranes from normal individuals were incubated with incorporation of [ ‘H] -NaBH, 500 mg% glucose for 3 days at 37”C, increased occurred mainly in proteins corresponding to the relative electrophoretic The glycoprotein mobility of myosin, glycoproteins IIb and IIIa (Figure 4). bands were identified by the characteristic mobility of glycoproteins IIb and IIIa in non-reduced and reduced conditions (17). Under reduced conditions, glycoprotein IIb showed increased mobility due to the cleavage of the fraction IIbB, but IIIa showed decreased mobility due to cleavage of intramolecular disulfide bonds. DISCUSSION Excessive nonenzymatic glycation of a variety of proteins, either derived from diabetic patients or following incubation with glucose has been described Since abnormalities of platelet function have been regularly observed in (18). diabetic patients, we quantitated and characterized platelet protein glycation. The incorporation of [‘HI-NaBH, into proteins under conditions favoring reductive alkylatlon has been., used for monitoring nonenzymatic glycatfon (19). However, in view of the reducing ability of NaBH, towards disulfide bridges (16) and lipids (8). the specific incorporation of [‘HI-NaBH, into fructosyl lysine, separated by aminoacid analysis, was used in our studies. Overall glycation as measured by the level of [‘HI-NaBH, incorporation into fructosyl lysine, was increased, and glycation of specific platelet protein3 examined by gel electrophoresis of whole platelets and platelet membranes showed that membrane glycoproteins IIb and IIIa participated in this glycation.

Vol. 55, No. 3

NONENZYMATIC PLATELET GLYCATION

347

Increased glycation of platelet membrane proteins in all diabetic patients studied was reported (6). The data presented, however, was not conclusive in relation to correlation between platelet, hemoglobin and plasma protein glycation. Moreover , the membrane fraction was only partially isolated with the triton-insoluble residue since no proteolytic inhibitors were used, as reported by Fox et al (20). In our study, five of eleven patients had increased glycabut there was no correlation between platelet, tion of their platelet proteins, Correlation was also not found with the hemoglobin, and albumin glycation. There are a number of potential severity of the atherosclerotic complications. explanations for these findings of varying platelet glycation. Platelets, L

ABP

+? ABP

+!,a Ilb

FIGURE4 Densitometry of fluorographic patterns following incorporation of [‘HI-NaBH, into platelet membranes isolated according to Jennings and Phillips (22) and incubated for 3 days at 37’C in the absence (continuous line) and presence of 500 mg$ of ABP , glucose (dashed line). Coomassie Blue pattern at bottom. actin-binding protein; M, myosin; IIb-IIIa, glycoproteins IIbIIIa; A, actin; L, lipid. albumin and hemoglobin all have different half-lives, of 10, 20 and 60 days, respectively, and therefore may show different degrees of glycation. Another explanation may be that the most glycated platelets are incorporated into thrombi. A shortened survival of diabetic platelets has indeed been reported In this case the platelets analyzed may represent only the least glycated (21). population. In this regard the successful 3-day --in vitro glycation of platelet membrane proteins (Figure 4) is noteworthy. Their glycation indicates that exposure of proteins to high ambient glucose concentrations for relatively brief time periods may result in easily-measurable increases in glycation. It appears that the nature of the plal;elet population must be taken into account for a correct interpretation of the results. A degree of selectivity was apparent in the -in vitro glycation of platelet membranes. Proteins corresponding to the relative electrophoretic mobility of and myosin were found to incorporate substantial glycoproteins IIb and IIIa, This may be related to the number of lysine residues in amounts of [ ‘HI-NaBH,.

348

NONENZYMATIC PLATELET GLYCATION

Vol. 55, No. 3

these proteins availablefor glycation. GlycoproteinsIIb and IIIa are known to form the fibrinogen receptor of activated platelets, and as such their increased glycation is of special interest. Even in a non-selective glycation pattern,the glycationof a particularproteinmay still result in altered function. This could involve increased fibrinogenbindingresultingin increased aggregation,and/or alterationin membranedeformability. Correlation between these platelet functions and the extent of plateletglycationis an important subjectfor future investigation.

ACKNOWLEDGEMENT We thank Dr. P. A. Finot (Nestle, La Tour de Peilz,Switzerland)for the authenticfructosyllysine. REFERENCES 1.

RUDERMAN,N. B. and HAUDENSCHILD, C. Diabetesas an atherogenicfactor. Prog. Cardiovas.Dis. 26, 373-411,1984

2. HALUSHKA,P. V., ROGERS,R. C., LOADHOLT,C. 9. and COLWELL,J. A. Increasedplateletthromboxanesynthesisin diabetesmellitus. J. Lab. Clin. Med. 97, 87-96, 1981 3.

DIMINNO,G., SILVER,M. J., CERBONE,A. M., RICCARDI,G., RIVELLESE,A., MANCINI,M. and THIAGARAJAN,P. Increasedbindingof fibrinogento plateand thromboxane. Blood 65, lets in diabetes: the role of prostaglandins 156-162,1985

4.

HIGGINS,P. J. and BUNN, F. H. Kineticanalysisof the nonenzymaticglycosylationof hemoglobin. J. Biol. Chem. 256, 5204-5208,1981

5. CHANG, J. C. F., ULRICH,P. C., BUCALA,R. and CERAMI,A. Detectionof an advancedglycosylationproductbound to Proteinin situ. J. Biol. Chem. 260, 7970-7974,1985 6. SAMPIETRO,T., LENZI,S., CECCHETI,P., GIAMPIETRO,O., CRUSCHELLI,L. and NAVALESI,R. Nonenzymaticglycationof human plateletmembraneproteinsin vitro and-in vivo. Clin. Chem. 32, 1328-1331,1986 7. ASTER, R. H. and JANDL, J. H. Plateletsequestrationin man. I. Methods. J. Clin. Invest.43, 843-855,1964 8. MILLER,J. A., GRAVALLESE,E. and BUNN, H. F. Nonenzymaticglycosylation of erythrocytemembraneproteins. J. Clin. Invest.65, 896-901,1980 9. SMITH, P. K., KROHN, R. I., HERMANSON,G. T., MALLIA,A. K., GARTNER,F. H., PROVENZANO,M. D., FUJIMOTO,E. K., GOEKE, N. M., OLSON, 8. J. and KLENK,D. C. Measurementof proteinusing bicinchoninicacid. Anal. Biochem.150, 76-85, 1985 10. LANE, L. C. A simplemethod for stabilizingprotein-sulfhydryl groups during SDS-gelelectrophoresis.Anal. Biochem.86, 655-664,1978

Vol. 55, No. 3

NONENZYMATIC PLATELET GLYCATION

349

11.

LAEMMLI, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophase T,. Nature 227, 680-685, 1970

12.

BONNER, W. M. and LASKEY, R. A. A film detection method for tritiumlabelled proteins and nucleic acids in polyacrylamide gels. J. Biochem. 46, 83-08, 1974

13. TISHLER, P. V. and EPSTEIN, C. J. A convenient method of preparing polyacrylamide gels for liquid scintillation spectrometry. Anal. Biochem. 22, 89-98, 1968 14.

KLENK, D. C., HERMANSON, G. T., KROHN, R. I., F'JJLMOTO,A. K., MALLIA, P. K., SMITH, P. K., ENGLAND, J. D., WIEDMEYER, H. M., LITTLE, R. R. and GOLDSTEIN, D. E. Determination of glycosylated hemoglobin by affinity chromatography: comparison with calorimetric and ion exchange methods, and effects of common interferences. Clin. Chem. 28, 2088-2094, 1982

15.

RENDELL, M., KAO, G., MECHERIKUNNEL, ?., PETERSEN, B., DUHANEY, J., NIERENBERG, K., RASBOLD, D., KLENK. D. and SMITH, P. K. Use cf aminophenylboronic acid affinity chromatography to measure glycosylated albumin levels. J. Lab. Clin. Med. 105, 63-69, 1985

16.

KRESS, L. F. and LASKOWSKI, H. The basic trypsin inhibitor of bovine pancreas. VII. Reduction with borohydride of disulfide bond linking half cystine residues 14 and 38. J. 3iol. Chem. 242, 4925-4929, 1967

17.

PHILLIPS, D. R. and AGIN, P. P. Platelet plasma membrane glyccproteins: evidence for the presence of nonequivalent disulfide bonds using nonreduced-reduced two-dimensional gel electrophoresis. J. Bicl. Chem. 252, 2121-2126, 1977

18.

BROWNLEE, M. and VLASSARA, H. Nonenzymatic glycosylation and the pathogenesis of diabetic complications. Ann. Int. Med. 101, 527-537, 1984

19.

MEANS, G. E. and FEENEY, R. E. Reductive alkylation of amine groups in proteins. Biochemistry 7, 2192-2201, 1968

20.

FOX, J. E. B., BOYLES, J. K., BERNDT, M. C., STEFFEN, P. K. and ANDERSON, L. K. Identification of a membrane skeleton in platelets, J. Cell Biol. 106, 1525-1538, 1988

21.

FERGUSON, J. C., MACKAY, N., PHILIP, A. D. and SUMNER, D. J. Determination of platelet and fibrinogen half-life with ("SE) selenomethionine: studies in normal and diabetic subjects. Clin. Sci. Mol. Med. 49, 115-120, 1975

22.

JENNINGS, L. K. and PHILLIPS, D. R. Purification of glycoproteins IIb and III from human platelet plasma membranes and characterization of a Calciumdependent glycoprotein IIb-III complex. J. Biol. Chem. 257, 10458-10466, 1982