Original article

SHORT-TERM VARIATIONS OF SERUM GLYCATED APOLIPOPROTEIN B

A.J.C. STAHL, A. RIMA, J.F. BLICKLE, J.M. BROGARD

SUMMARY-  

Serum glycated apolipoprotein B (apo B-G) levels were determined in31 non-insulin-dependent diabetic patients and 13 control subjects by anenzyme-linked immunosorbent assay. Apo B-G was increased in diabetic patientscompared to non-diabetic subjects, whether expressed as a serum concentration(57.7 vs 36.1 mg/l) or a percentage of total apolipoprotein B (4.42 vs3.14 %). Apo B-G, together with other markers (mean daily plasma glucose, serumfructosamines, triglycerides, total cholesterol, glycated haemoglobin), was measuredbefore and after 5 days of therapeutic adjustment in diabetic patients. In20 patients with a favourable course of glycaemic control, the mean decrease of apoB-G concentration reached nearly 16 %. In 11 patients with an unfavourablecourse, the increase of apo B-G concentration was about 14 %. Therefore, variation ofserum apo B-G concentration could serve as an additional short-term marker forglycaemic control, although possible concomitant variations of serum triglycerides ortotal apolipoprotein B concentrations should also be considered. 151-155.

apolipoprotein B, glycation, non-insulin-dependent diabetes mellitus, glycaemiccontrol markers.

R+SUM+  -  

Variations à court terme de l'apolipoprotéine B glyquée sérique.Laconcentration sérique de l'apolipoprotéine B glyquée (apo B-G) a été dosée chez31 patients diabétiques non insulinodépendants et chez 13 sujets nondiabétiques, à l'aide d'une méthode immuno-enzymatique. Les concentrations de l'apo-Bsérique sont plus élevées que chez les patients diabétiques que chez les sujetscontrôles, qu'elles soient exprimées en concentration moyenne (57,7 vs36,1 mg/l : p < 0,02) ou en pourcentage de glycation del'apolipoprotéine B totale (4,42 vs 3,14 % ; p < 0,05).L'apo B-G, ainsi que d'autres marqueurs (glycémie moyenne, fructosamines, triglycérides,cholestérol total, hémoglobine glyquée) a été mesurée chez les patients diabétiquesavant, puis après 5 jours d'ajustement thérapeutique. Chez 20 patients dontl'évolution du contrôle glycémique s'est avérée favorable, la concentration sériquedes apo B-G s'est abaissée en moyenne d'environ 16 %. Chez les 11 patients,dont l'évolution glycémique s'est montrée défavorable, la concentration sériquemoyenne des apo B-G s'est élevée d'environ 14 %. La variation de l'apo B-Gsérique pourrait constituer un marqueur supplémentaire du contrôle glycémique à courtterme, mais il importe de tenir compte alors d'une éventuelle modification concomitantedes concentrations sériques des triglycérides ou des apolipoprotéines B totales. 151-155.

apolipoprotéines B, glycation, diabète non insulinodépendant, marqueurs ducontrôle glycémique.

    &writer; : A.J.C. Stahl,Laboratoirede la Clinique Médicale B, Service de Médecine Interne, Hôpital Civil, 1, place del'Hôpital, 67091 Strasbourg Cedex (France).Fax : (33)(0)3 88 11 62 62. Received : February 20, 1997 ; revised :December 20, 1997.

L'yperglycaemia in diabetes leads to non-enzymatic glycation of cellular and bloodproteins [1]. Serum beta-lipoprotein [2] or low-density lipoprotein (LDL) glycation occursin diabetic patients [3]. Serum î-lipoproteins, separated by electrophoresis and stainedby tetrazolium salts, show an increased glycation rate and a high atherogenic index inthese patients [4]. A higher degree of glycation of apolipoprotein B (apo B),isolated by affinity chromatography and assayed colorimetrically, has been noted indiabetic patients as compared to non-diabetic subjects [5]. A positive correlation hasbeen found in diabetes between plasma glycated LDL and glucose [6], glycatedî-lipoproteins and fructosamines [7], serum glycated apo B and fasting plasmaglucose or glycated haemoglobin [8].

The methods used to date to determine glycated LDL or glycated apo B have beentechnically arduous and time-consuming, without being specific. Several immunochemicalassays have been described. Monoclonal antibodies binding specifically to reduced glucoseconjugates and covalently linked to the &epsiv;-amine of lysine, glucitol-lysine, havebeen used after isolation of lipoproteins from plasma by sequential ultracentrifugation[9]. A sandwich immunoradiometric assay, using polyclonal antiserum towards LDL and ¹²⁵I-labelledmonoclonal antibodies towards apolipoprotein B100, has been applied to glycated serumfractions recovered from an affinity column [8]. Specificity was improved, but thedetermination was still long and cumbersome.

More recently, an enzyme-linked immunosorbent assay using a new antibody, specific forglycated apo B, has been described [10]. This prompted us to measure apo B-G inpopulations of control subjects and diabetic patients. As apolipoprotein B has aplasma residence time of only 3 days [11], we measured its glycation level atadmission and 5 days after therapeutic adjustment in the hospital.

PATIENTS AND METHODS

The control subjects in the study (9 men and 4 women,70.6 Ý 14.8 years of age) were admitted to hospital for evaluation and hadno disease severity or history of diabetes. Their fasting plasma glucose was lower than6.5 mmol/l in two successive measurements, and their body mass index was24.5 Ý 2.3 kg/m². The diabetic patients in the study(15 men and 16 women, 67.5 Ý 11.6 years of age) were admittedfor therapeutic adjustment because of poor metabolic control. All had NIDDM and were on adiet and taking oral hypoglycaemic drugs (sulphonylurea), supplemented in 10 caseswith insulin. Twenty-two of them presented complications. Their body mass index was30.0 Ý 3.1 kg/m Ý. After five days of hospital stay, the patientswere divided into two subgroups : those with a favourable course (D. fav.,n = 20 : 9 men, 11 women ; 12 with complications)exhibiting a 15 % or more decrease in their mean daily plasma glucoseconcentration ; and those with no significant change (or even a worsening) of meanplasma glucose (D. unfav., n = 11 : 6 men, 5 women ;8 with complications). Control subjects were on standard hospital diet. Informedconsent was obtained from all subjects, and investigations were performed in accordancewith the principles of the Helsinki Declaration.

Mean daily plasma glucose (GLY) was determined at 8 a.m. (fasting), 11 a.m.,5 p.m. and 12 p.m., using the glucose-oxidase-peroxidase method [12] on anHitachi 737. Serum fructosamine (FRU) was assayed by the method of Baker et al.[13] on an Hitachi 704. Glycated haemoglobin (Hb A1c) was measured byion-exchange chromatography on a Diamat (Bio-Rad). Total serum apolipoprotein B(apo B) was assayed by immuno-nephelemetry on a Beckman-Array 360. Serumtriglycerides were obtained by the lipase-glycerokinase method [14], and serum totalcholesterol with the full enzymatic method [15] on an Hitachi 737. Glycatedapolipoprotein B (apo B-G) was determined in serum by immunocompetition ELISA onmicroplates (Glycacor, Exocell, Philadelphia, PA, USA) using mouse monoclonal antibodyES-12 which recognizes a specific glycated epitope on apo B in LDL particles[10]. At the level of 34 mg/l of apo B-G, intra-assay variation was 5.8 % andinter-assay reproducibility was 8.1 %. Results (mean Ý standard deviation)were analysed by the Kolmogerow-Smirnov non-parametric test for paired or unpaired data.Statistical calculations were performed using Stat View (Cricket Software, Inc.,Philadelphia, PA, USA).

RESULTS

Serum apo B-G in control subjects and diabetic patients - Totalapo B was slightly but not significantly higher in diabetic patients than in controlsubjects. Serum apo B-G levels, expressed either as a concentration or as a percentage oftotal apo B glycation, were significantly increased in diabetic patients as comparedto controls (Table I). Significance was greater for the concentration(57.7 Ý 16.0 vs 36.1 Ý 21.1 ; p < 0.02) thanthe percentage of apo B glycation (4.42 Ý 1.74 vs3.14 Ý 1.75 ; p < 0.05).

Serum parameters in diabetic patients with a favourable or unfavourablecourse - Table II shows the values for markers of glycaemic controlat hospital admission (D0). Mean daily plasma glucose, serum fructosamines andHbA1c were increased in both subgroups of diabetic patients as compared to non-diabeticcontrol subjects. Five days later (D5), the values of these three markersdecreased as expected in the D. fav. subgroup but were maintained or increasedslightly in the D. unfav. subgroup. The initial glycaemic status of the D. fav.subgroup was poorer than that of the D. unfav. subgroup.

Table III shows the values for three markers of lipid metabolism (serumtriglycerides, total cholesterol and total apo B) that were only slightly increasedin both subgroups of diabetic patients as compared to control subjects, except for serumtriglycerides which were significantly higher at D0 in patients with asubsequent favourable course (p < 0.02) consistent with their poorerglycaemic control. After 5 days of therapeutic adjustment, all parameters were lowerin patients with a favourable course, although the decrease in total cholesterol was notsignificant. In the subgroup with an unfavourable course, the three parameters increasedslightly but not significantly.

Serum glycated apolipoprotein B in diabetic patients with a favourable orunfavourable course - Serum apo B-G concentration and the rate ofapo B glycation were higher at D0 in both diabetic subgroups as comparedto controls. As for the other markers, the rise was greater in the D. fav. than theD. unfav. subgroup (Table IV). After 5 days of therapeuticadjustment, the mean apo B-G concentration (- 15.6 %) and the percentage ofapo B glycation (- 11.9 %) decreased significantly in the D. fav.subgroup without reaching the values found in control subjects. Conversely, theseparameters did not improve, and even worsened, in the D. unfav. subgroup(respectively + 14.3 % and + 4.3 %).

Rate of variation of glycated apolipoprotein B as compared to othermarkers - In the D. fav. subgroup, decreases in mean serumapo B-G concentration and apo B glycation rate were markedly higher than forHb A1c (- 4 %), but less than for mean plasma glucose (- 23 %)and fructosamines (- 15 %). In the subgroup of diabetic patients with anunfavorable course, serum apo B-G concentration increased slightly less than meanplasma glucose (+ 22 %) but more than fructosamines (+ 5.5 %) andHb A1c (+ 5 %). In both subgroups, no significant correlation was foundbetween variations in apo B-G concentration and those of markers of glycaemiccontrol, except for mean plasma glucose (r = 0.520 ; p = 0.02).This result seems to be in accordance with the plasma residence half-time of the variousproteins.

DISCUSSION

Glycation of apo B results in a deviation of LDL metabolism favouring capture bymacrophage scavenger receptors, which leads to foam cells and yjr onset of atherosclerosis[16-18]. As apo B, essentially apo B 100, has a plasma residence half-timeof 3 days [11], measurement of its glycation could serve as a marker of short-termglycaemic control in diabetic patients.

Previous methods of measuring glycated LDL and glycated apo B have beencomplicated and unreliable. They required at least two steps consisting of separation byultracentrifugation, electrophoresis or chromatography, and colorimetric measurement aftereither acid hydrolysis or tetrazolium salt reduction [3-5]. The purpose of our study wasto test a recently described direct and rapid immunochemical assay of glycated apo Bby measuring apo B-G changes in diabetic patients in conjunction with variations ofblood glucose levels.

As indicated in previous studies [2, 4, 5, 8, 9], the serum concentration of glycatedapo B was found to be higher in diabetic patients than control subjects. In ourstudy, the mean increase reached 60 %, and the percentage of total serum apo Bglycation was 40 %. Using the same method, Cohen et al. [19] obtained similarresults in a smaller group of patients, with the increase in serum apo B-Gconcentration reaching 87 % for some Type 2 diabetic patients. Withaffinity chromatography and immunonephelemetry, Panteghini et al. [20] found a meanincrease of 33 % for glycated apo B in Type 2 patients. The antibodies usedin our assay were specific for apo B 100 [10], which represents more than95 % of total LDL apolipoproteins. Little is currently known about possible changesin the environment of the epitope [21] identified by the apo B antibody whenapo B is glycated. Although apo B-G seems to be recognized by the antibody, anunderevaluation of the results cannot be excluded.

Our results show that changes in apo B glycation occur within five days when thetreatment is adapted towards better glycaemic control. Serum glycated apo Bconcentration decreased by 16 % and the apo B glycation rate by 12 % in thesubgroup of diabetic patients with a favourable course, as assessed by the decrease ofmean serum glucose, fructosamine concentration and glycated haemoglobin. It is noteworthythat the patients of this subgroup had a very poor glycaemic status on the day ofadmission. The subgroup of patients with an unfavourable course during their 5-dayhospital stay showed a mean increase in serum glycated apo B concentration of 14 %and an increase in the total apo B glycation rate of 4 %. On the basis of theseresults, the serum glycated apo B concentration would seem suitable as a routineshort-term marker (1 week) of glycaemic control in diabetic patients. It would beeasier to assess than glycated fibrinogen [22] or serum immunoglobulin M, which decreasedby 13 % in a group of 28 Type 2 patients who improved their glycaemiccontrol after 5 days of therapeutic adjustment [23].

Though changes in apo B glycation may be mainly due to variations in serum glucoseconcentrations, they are also under the influence of the concentration, size andcomposition of apo B-carrying lipoprotein particles. Table III shows thatserum triglyceride concentrations decreased by nearly 37 % within 5 days indiabetic patients with a favourable course, but increased by 6 % in those with anunfavourable course. No close correlations were found between the 5-day variations inserum triglyceride concentrations and serum apo B-G concentrations(R = 0.32 ; p > 0.3) or apo B glycation rates(R = 0.18 ; p > 0.3). This lack of correlation could have beendue to the more rapid metabolism of triglycerides than of apo B-G. Triglycerides areessentially carried by VLDL particles, and to a lesser extent by LDL particles, both ofwhich contain apo B. VLDL half-life is very short (5 to 6 h) in comparisonwith LDL particles, so that day-to-day changes in apo B-G could be accelerated by the mainvariations of VLDL particles.

The changes in serum apo B-G concentrations were greater than those of apo Bglycation rates (Table IV). In the diabetic subgroup with a favourable course,mean total serum apo B concentration decreased only by about 6 % in 5 days,which suggests that the catabolism of glycated apo B was slower than that ofnon-glycated apo B, or that the day-to-day changes in total apo B were morerapid due to the fast metabolism of VLDL particles.

In searching for an influence relating to the kind of treatment (dietary change alone,oral hypoglycaemic agents, insulin addition), we found no significant differencescorresponding to the 5-day variations in apo B-G. The presence of complications indiabetic patients had no significant influence on serum apo B-G variations.

In conclusion, our study showed a rapid and significant variation in serum apo Bglycation in non-insulin-dependent diabetic patients after 5 days of therapeuticadaptation. A decrease in apo B glycation occurred in the subgroup of patients whoimproved their glycaemic control, while an increase was observed in the subgroup in whichglycaemic control became poorer. Although our results for serum glycated apo B cannotbe taken as definitive because of the limited number of patients included in this study,they open up a new area of interest for future clinical investigations in the field ofatherogenesis in Type 2 diabetic patients.

TABLE I.    Serum totalapolipoprotein B concentration (Apo B tot g/l), glycated apolipoprotein Bconcentration (Apo B-G mg/l), and rate of apolipoprotein B glycation (Apo B-G %)innon-diabetic control subjects and diabetic patients at day of admission (D0).

Mean Ý standard deviation. a = p < 0.02 vscontrols. b = p < 0.05 vs controls.

TABLE II.    Mean daily plasmaglucose concentration (GLY), serum fructosamine concentration (FRU) and percentage ofglycated haemoglobin (HbA1c) in control subjects and diabetic patients with a favourable(D. fav.) or unfavourable course (D. unfav.) at day of admission (D0) and after5 days of treatment (D5).

Mean Ý standard deviation. a = p < 0.05 vscontrols. b = p < 0.05 : D5 vs D0.c = p < 0.05 : D. fav. vs D. unfav.

Table III.    Serum triglycerides(TRIGLY.), total cholesterol (CHOL. T), and total apolipoprotein B concentrations (Apo BT) in control subjects and diabetic patients with a favourable (D. fav.) or unfavourable(D. unfav.) course at day of admission (D0) and after 5 days of treatment(D5).

Mean Ý standard deviation. a = p < 0.05 vscontrols. b = p < 0.05 : D5 vs D0.c = p < 0.05 : D. fav. vs D. unfav.

TABLE IV.    Serum glycatedapolipoprotein B concentration (Apo B-G mg/l)and percentage of apolipoprotein B glycation(Apo B-G %) in control subjects and diabetic patients with a favourable (D. fav.) orunfavourable course (D. unfav.) at day of admission (D0) and after 5 daysof treatment (D5).

Mean Ý standard deviation. a = p < 0.05 vscontrols. b = p < 0.05 : D5 vs D0.c = p < 0.05 : D. fav. vs D. unfav.

REFERENCE(S)

  1    Kennedy L, Baynes JW. Non-enzymatic glycosylationand the chronic complications of diabetes : an overview. Diabetologia, 1984, 26,93-98.

  2    Schleicher E, Deufel T, Wieland OH. Non-enzymaticglycosylation of human serum lipoproteins. FEBS Lett, 1981, 129, 1-4.

  3    Lyons TJ, Baynes JW, Patrick JS, Colwell JA,Lopes-Virella MF. Glycosylation of low density lipoprotein in patients with type1 diabetes : correlations with other parameters of glycaemic control. Diabetologia,1986, 29, 685-689.

  4    Kobayashi K, Ogasahara N, Sakoguchi T, Kimura M,Taniuchi K, Matsuoka A. Agarose gel electrophoresis with nitroblue tetrazolium colorationfor determination of glycated lipoproteins in serum from diabetics. Clin Chem,1990, 36, 65-69.

  5    Caines PS, Cheung RMC, Draisey TF, Yee GI.Colorimetric determination of non-enzymatically glycated apolipoprotein B. Clin Chem,1989, 35, 1074-1077.

  6    Jack CM, Sheridan B, Kennedy L, Stout RW.Non-enzymatic glycosylation of low-density lipoprotein. Results of an affinitychromatography method. Diabetologia, 1988, 31, 126-128.

  7    Kobayashi K, Ogasahara N, Sakoguchi T, Taishi K,Makimura H, Matsuoka A. Correlation between glycated lipoproteins and fructosamine levelin serum. Chem Pharm Bull, 1991, 39, 2149-2151.

  8    Tames FJ, Mackness MI, Arrol S, Laing I,Durrington PN. Non-enzymatic glycation of lipoprotein B in the sera of diabetic andnon-diabetic subjects. Atherosclerosis, 1992, 93, 237-244.

  9    Curtiss LK, Witztum JL. Plasma apolipoproteins AI,AII, B, CI et E are glycosylated in hyperglycemic diabetic subjects. Diabetes,1985, 34, 452-461.

10    Shea EA, Cohen MP. Immunologic detection and measurement ofglycated apolipoprotein B with site-specific monoclonal antibodies. J ImmunolMeth, 1993, 162, 85-95.

11    Chan L. Apolipoprotein B messenger RNA editing and update. Biochimie,1995, 77, 75-78.

12    Trinder P. Enzymatic assay for plasma glucose. Ann ClinBiochem, 1964, 6, 24-28.

13    Baker JR, Metcalf PA, Johnson RN, Newman D, Rietz P. Use ofprotein-based standards in automated colorimetric determinations of fructosamine in serum.Clin Chem, 1985, 31, 1550-1554.

14    Nagele V, Haegele EO et al. Plasma triglyceride assayusing an enzymatic method. J Clin Chem Biochem, 1984, 22, 165-174.

15    Allain CC, Poon LS, Chan CS. Determination of plasmacholesterol with an enzymatic method. Clin Chem, 1974, 20, 470-475.

16    Brown MS, Goldstein JL. Scavenging for receptors. Nature,1990, 343, 508-509.

17    Ginsberg HN. Lipoprotein physiology in nondiabetic anddiabetic states. Relationship to atherosclerosis. Diabetes Care, 1991, 14,839-855.

18    Gugliucci A, Dumont S, Siffert JC, Stahl AJC. In vitroglycated low-density lipoprotein interaction with human monocyte-derived macrophages. ResImmunol, 1992, 143, 17-23.

19    Cohen MP, Lautenslager G, Shea E. Glycated LDL concentrationsin non-diabetic and diabetic subjects measured with monoclonal antibodies reactive withglycated apolipoprotein B epitopes. Eur J Clin Chem Clin Biochem, 1993, 31,707-713.

20    Panteghini M, Cimino A, Pagani F, Girelli A. Non-enzymicglycation of apolipoprotein B in patients with insulin- and noninsulin-dependent diabetesmellitus. Clin Biochem, 1995, 28, 587-592.

21    Ziegler O, Méjean L, Igan B, Fruchart JC, Drouin P, FiévetC. Accessibility of human apolipoprotein B-100 epitopes in insulin-dependentdiabetes : relation with the surface lipid environment of atherogenic particles. Diabetes& Metabolism, 1996, 22, 179-184.

22    Menini T, Gugliucci A, Sodahlon YK, Stahl AJC, Blicklé JF,Brogard JM. Glycated immunoglobulin M in diabetic patients. Ann Biol Clin, 1993, 51,887-891.

23    Gugliucci A, Menini T, Stahl AJC. Glycation of fibrinogen indiabetic patients : a practical colorimetric assay. Glycosylation & Disease,1994, 1, 177-183.