Clinica Chimica Acta, 190 (1990) 249-262 Elsevier

249

CCA 04807

Urinary proteins and red blood cell membrane negative charges in diabetes mellitus Alfred Bernard ‘, Ali Ouled Amor ‘, Jocelyne Goemare-Vanneste ‘, Jean-Luc Antoine *, Robert Lauwerys I, Id& Colin 3, Bernard Vandeleene 3 and And& Lambert 3 Units of ’ Industrial Toxicology, 2 Occupational Dermatology and ’ Diabetoloa and Nutrition, University of Louvain, Brussels (Belgium) (Received

31 March

Key words: Proteinuria;

1989; revision received 29 May 1990; accepted Alcian blue; Glomerular polyanion; Diabetic nephropathy

5 June 1990)

Retinol-binding

protein;

Summary

The nature and origin of proteinuria in diabetes mellitus have been investigated by measuring the urinary excretion of seven specific proteins of low (µglobulin, retinol-binding protein) or high molecular weight (albumin, transferrin, hemopexin and IgG). Using the Alcian Blue binding test, we also measured negative charges on red blood cell (RBC) membrane which according to recent studies might mirror the glomerular polyanion charge. A group of 190 diabetics was examined, including 90 patients with type I diabetes, 23 type II diabetics treated with diet and/or hypoglycaemic agents and 77 longstanding type II diabetics requiring insulin therapy. With the exception of &-microglobulin all proteins measured were excreted in the urine of diabetics in significantly higher amounts than in controls. The assay of transferrin proved the most sensitive (58% positive) followed by albumin (49%), IgG (34%), hemopexin (28%) and retinol-binding protein (26%). Practically the same ranking was obtained when only type I diabetics were considered. RBC membrane negative charges were diminished in diabetics and negatively correlated with the urinary excretion of albumin (r = - 0.61, n = 190). RBC charges were also negatively correlated with other urinary proteins of high molecular mass (r between - 0.5 and - 0.2) but presented no relation with urinary &nicroglobulin or retinol-binding protein. The loss of RBC charges in diabetics most likely reflects the concomitant depletion of the glomerular polyanion responsible for the increased

Correspondence to: A. Bernard, Unit of Industrial Chapelle-aux-Champs, 1200 Brussels, Belgium.

0009-8981/90/$03.50

Toxicology,

0 1990 Elsevier Science Publishers

University

B.V. (Biomedical

of Louvain,

Division)

30.54.

Clos

250

glomerular leakage of high molecular mass plasma proteins. The preferential increase in transferrin excretion together with the progressive rise in the urinary excretion of IgG lead us to postulate that the loss of glomerular polyanion in diabetes is accompanied, from the early stage, by a progressive decrease in the size-selectivity of the glomerular filter. The urinary excretion of retinol-binding protein was weakly correlated with albuminuria (r = 0.26, n = 186). Eight % of diabetics showed an elevation of urinary retinol-binding protein without evidence of microalbuminuria, which clearly demonstrates that a proximal tubular impairment can occur independently of the glomerular alterations in the course of diabetic nephropathy. Introduction

The mechanisms responsible for the increased protein excretion at the early stage of diabetic nephropathy are not fully understood. Glomerular permeability to proteins is the result of various factors: renal hemodynamics, porosity and charge of the glomerular basement membrane (GBM) and the charge and size of filtered proteins. All these factors are altered in diabetes. At the early stage of the disease, the glomerular plasma flow and the intraglomerular pressure are increased and these changes are believed to be responsible for the increased protein excretion when glucose metabolism is not controlled [l-4]. For some authors, these hemodynamic alterations are major determinants of both the initiation and progression of diabetic nephropathy [3]. Several studies with neutral dextrans or specific proteins have demonstrated that the size-selectivity of the glomerular filter in diabetics is compromised by the development of a nonrestrictive shunt pathway which may emerge early in the course of the disease [4-81. Diabetic nephropathy in man or animal is also accompanied by a depletion of the glomerular polyanion (GPA) resulting from a loss of sialic acid and heparan sulfate [9-131. This phenomenon is presumably responsible for the reduction of charge-selectivity of the glomerular filter suggested by the preferential urinary excretion of albumin over IgG [4] or neutral dextran [5,14] in patients with microalbuminuria. But the opposite observation, i.e. the preferential excretion of IgG over albumin, has also been made in diabetics without microalbuminuria which indicates an increased anionic pore charge within the GBM [15]. The study of the fractional clearances of dextrans in diabetic rats also points to an increase in the negative charge density of the filtration barrier [16], which led to the hypothesis of a tubular origin for the diabetic microalbuminuria [17]. The situation is rendered even more confused by the fact that as a result of non enzymatic glycosylation albumin from diabetics presents charge and conformational changes which affect its renal handling at both the tubular and glomerular level [18-201. To gain further insight into the origin of diabetic proteinuria we have attempted to assess the charge- and size-selectivity of the glomerular filter by the measurement of specific urinary proteins and of red blood cell membrane negative charges, the latter reflecting the GPA charge [21,22].

251

Materials and methods The study was conducted on a group of 190 diabetics, including 90 patients with type I diabetes, 23 type II diabetics treated with diet and/or hypoglycaemic agents and 77 longstanding type II diabetics requiring insulin therapy. The mean age of these groups averaged 42.4 t_ 15.5 (SD), 59 + 11.9 and 60.2 f I2 years, respectively. The mean duration of diabetes was 15.2 f 12, 6.9 25.9 and 13.4 f 7.3 years respectively. For the total population, the mean age was 51.5 5 16 (range 14-81) yr and the mean duration of diabetes 14 f 11 (range l-47) yr. Thirty three patients had blood pressure readings higher than 160/95 mm Hg and twelve had an abnormal serum level of creatinine (> 0.115 mmol/l). Diabetic patients were compared with a control group of 57 individuals, aged 18 to 84 yr (mean t_ SD : 48.3 f 16.2). Patients (20 to 30 per week) were examined in the morning. Five ml of venous blood (in a tube containing 0.1 ml EDTA-Na, lo%, w/w) and a spot urine sample were collected. These samples were stored at +4*C until analysis. RBC membrane negative charges were measured by the binding of the cationic dye, Alcian Blue 8GX (AB). The analysis was carried in the 24 h of blood collection. The method i~tially described by Levin et al. [21] was modified as follows. RBCs were washed 4 times with phosphate buffered saline, 0.15 mol/l NaCl and 0.05 mol/l phosphate, pH 7.4. The AB solution (54% purity, from Sigma Chemical Co., St Louis, MO, USA) was prepared just before use by dissolving the dye (4 g/l) in a solution containing 25 mmol/l MgCl, and 0.15 mol/l NaCl. This solution was vigorously mixed for 10 min and the undissolved dye was removed by centrifugation (2000 x g x 10 min) and filtration on Whatman no. I filter, The final AB concentration was determined from the extinction coefficient 68.4 for the crude AB preparation). (E:? The results presented here were obtained with a saturating AB concentration of 1.5 g/l. RBCs were counted with a Technicon Autocounter (Technicon, Wemmel, Belgium). The concentrations of ~~-~croglobulin f&-m), retinol-boding protein (RBP), albumin, transferrin, hemopexin and IgG were determined by an automated latex agglutination method [23]. The antibodies against these proteins were obtained from Dako Immunoglobulin, Copenhagen, Denmark. The assays of B-m and RBP were calibrated with proteins purified as described previously [24] and that of albumin with the protein purchased from Sigma Chemical Co. A pool of normal sera was used to calibrate the assays of ~~sfe~n (3 g/l), hemopexin (0.75 g/l) and IgG (12.5 g/l). All these immunoassays had a detection limit of 5 pg/l and required a dilution of at least 20-fold of normal urine samples. Table I presents the main molecular features of the 7 proteins measured. Since the literature did not provide consistent data for the sizes of human transferrin and hemopexin, we fractionated a pool of 10 fresh normal sera on a Sephadex G-200 column. The peaks of transferrin and hemopexin were localized by measuring both proteins in the eluted fractions. The relative molecular masses (M,) and effective molecular radii given in Table I were estimated by reference to albumin and IgG (peaks also localized immunochemically).

252 TABLE I Protein characteristics Protein

&-microglobulin Retinol-binding protein Albumin Transferrin Hemopexin IgG

Ekctrophoretic mobiliti t% a2

Albumin

weight

Molecuiar radius @m>

11800 21400 66 300 81100 a 84600 a 156000

1.6 2.07 3.6 3.97 4.06 5.5

MOlSXXliW

’ Estimated in the present study by gel filtration of a serum pool on Sephadex G-200.

Creatinine was measured in urine and in serum by Jaffe’s method (125,261, respectively). Hem~ob~ A,, was determined by a c~o~to~ap~c technique [27]. The electrophoretic mobility of albumin and transfer& was examined on agarose gel with barbital buffer, pH 8.6, ionic strength 0.1 [28]. After electrophoresis the gel was cut into 0.5 cm pieces, which were placed in 0.5 ml of 0.15 mol/l NaCI. The eluted albumin and transfer& were measured using a latex assay [23]. Protein excretion was expressed in pg or mg per g of creatinine, the results obtained by this method being comparable with those based on the excretion rate (29-311. The results were evaluated by non parametric (Spearman’s correlation, r, and x2 test), or parametric tests after logarithmic transformation (Pearson’s correlation, Student’s t test and polynomial regression). Prevalences of abnormal values were calculated by using as upper limit of normal the geometric mean + 2 geometric SD of values found in the control group.

RG?Sl&S

Urinary excretion of proteins The mean urinary excretion of proteins, together with the prevalences of elevated values are presented in Table II for control subjects, type I diabetics and all diabetics. With the exception of &-m, all proteins were excreted in significantly greater amounts in diabetics (type I or all types) whether the comparison is made on the basis of mean values or of prevalences. The extent of proteimuia in diabetes was however variable. When all types of diabetes are taken together, the highest prevalence was observed with transferrin (58%), followed by albumin (49%), IgG (34%), hemopexin (28%) and RBP (26%). The ranking of prevalences was practicahy the same in type I diabetics except that the difference between transferrin and other proteins was still more pronounced (Table II).

creak)

50

54.8 (1.6-4000) 71.3 (14.5-727) 51 (20-151) 6.4 (2.4-58) 0.13 (0.03-1.8) 0.71 (0.26-2.4)

Xg (range)

and prevalences

90 84 60

3.5 0

74

87

86

n

Type 1

Diabetics

and controls

63.1 (1.5-6030) 155 (3.2-50800) ‘I 78 (10-7360) b 18.3 (2.5-3616) d 0.67 (0.06-138) d 1.62 (0.18-95) d

Xg (range)

values in diabetics

5.3

5.3

5.3

5.3

% Positive values

of abnormal

23.8 ’

47.6 d

26.7 ’

27.0 ’

21.8 a

3.5

% Positive values

121

177

190

153

186

182

n

56.9 (1.5-6030) 148 (3.2-50800) d 92 (10-41000) b 31.4 (2.5-5 200) d 1.14 (0.02-209) d 1.94 (0.08-193)

Xg (range)

Total population

33.9 d

57.6 d

48.9 d

27.5 d

25.8’

5.5

S Positive values

The means (geometric) were compared by the Student’s t test and the prevalences by the x2 test with Yates correction. The levels of significance were as follows: a P c 0.025; b P < 0.01; ’ P < 0.005 and d P < 0.001. Prevalences of elevated values were calculated by using as upper limit of normal the mean + 2 SD of log transformed values observed in controls: /12-m. 464 pg/g treat.; RBP, 248 pg/g treat.; hemopexin, 120 gg/g treat.; albumin, 24 mg/g treat; transferrin, 0.7 mg/g treat; IgG, 2.4 mg/g treat.

IgG (mg/g

57

57

57

@g/g treat.) Hemopexin

(gg/g treat.) Albumin (mg/g treat.) Transferrin (mg/g treat.)

51

RBP

n

Controls

of proteins

57

excretion

Urinary

h-m W/g treat.)

II

TABLE

254 TABLE III Correlation matrix between urinary proteins and RBC charges in diabetics (n between 121 and 190)

W Transferrin Albumin Hemopexin RBP A-m

RBC

/G-m

RBP

Hemopexin

Albumin

Transfertin

- 0.47 P

Urinary proteins and red blood cell membrane negative charges in diabetes mellitus.

The nature and origin of proteinuria in diabetes mellitus have been investigated by measuring the urinary excretion of seven specific proteins of low ...
1MB Sizes 0 Downloads 0 Views