Iffeet of a Low Protein Diet on he Relationship of noRenzymatic Glycation D Altered Renal Structure and Function in liabetes Mellitus

Kenneth R. Copeland, MSc,

Randall W. Yatscoff, PhD,

James A. Thliveris, PhD,*

S. Brian Penner, MDT

Adi Mehta, Mot

Departments of Clinical Chemistry,

*Anatomy and -/Medicine, Health Sciences Center, University of Manitoba, Winnipeg, Manitoba

ABSTRACT

Renal functional parameters including creatinine clearance, urinary al- bumin excretion, basement membrane thickening, and levels of nonen- zymatic glycation of glomerular basement membrane were studted in streptototocin-induced diabetic rats and age-matched controls sub- jected to low protein diet. In addition, these parameters were also assessed in diabetic and streptozotocin injected nondlabetic animals fed a 24% protein diet, which served as “positive controls.” While di- abetic animals from both diet groups had similar elevated glycated hemoglobin levels and increased levels of nonenzymatk glycation of glomerular basement membrane, these were significantly elevated as compared to insulin treated diabetic (euglycemic), age-matched con- trols on an 8% protein diet, and streptozotocin injected nondiabetic animals from both diet groups. However, urinary albumin excretion and creattnine clearance levels were significantly elevated only in the 24% protein diet fed diabetics over values seen in the various groups of animals on 8% and controls on 24% protein diet. In contrast, there were no statistical differences among diabetic, euglycemic and con- trol (8% and 24% protein) animals with respect to creatinine clear- ance, urinary albumin excretion, and glomerular basement membrane thickness. Taken together these data cast some doubt on the role of nonenzymatic glycation in the development of diabetic nephropathy. Moreover, hyperglycemia per se causes a compensatory increase in kidney size irrespective of protein intake; a low protein diet, however, inhibits the hyperftltration commonly seen in early diabetic nephropa- thy. The authors, thus, hypothesize that a low protein diet, by pre- venting compensatory increase in blood flow to survivlng nephrons, in some fashion protects these functional units from subsequent damage and possibly delays the onset of renal failure. (The Journal of Diabetic Complications 3;2:113-119, 1989)

INTRODUCTION

Reprint requests: Dr. Adi Mehta, Room

GG449, Health Sciences Center, Sec- tion of Endocrinology and Metabolism, 700 William Avenue, Winnipeg, Manitoba,

Canada, R3E 023.

Submitted for publication in April 1988; accepted in revised form in August 1988.

The most common and earliest structural manifestation of the abnor- malities induced by or associated with hyperglycemia in diabetes mel- litus is microangiopathy.’ This disorder, specific to small vessels, af- fects all organs of the body but is clinically manifest most commonly in the eye and kidney. Morphologic documentation of microangiopa- thy has focused on capillary basement membrane thickening (CBMT). At the light microscopic level, PAS positive thickening of the capillary and small vessel wall has been demonstrated in the retina* and kidney.3 Electron microscopy has documented glomerular4.5 as well as skeletal muscle6 capillary basement membrane thickening.

Recent studies have shown that CBMT can be correlated with the degree of glycemia in diabetes. An increase in blood glucose has been shown to result in increased levels of nonenzymatic glycation,’ a com- mon posttranslational modification that has been shown to occur for many proteins.8,g It is a result of direct chemical reactions between re- ducing sugars and primary amino groups of proteins, ultimately result- ing, via Amadori rearrangement, in the formation of a stable ketoamine

113

114

derivative.lOThe reaction is glucose dependent and is in- creased in hyperglycemic situations. Nonenzymatic gly- cation has been shown to alter the structure and function of many proteins, both in vitro and in vivo.@.ll.l2

The hyperglycemia found in diabetes has been shown to result in increased nonenzymatic glycation of pro- teins. This may precipitate changes in the composition and synthesis of glomerular basement membrane,13814 ul- timately leading to functional and morphologic changes in the kidney. The functional changes that occur early on in diabetic nephropathy include an increase in the glomerular filtration rate (GFR)ls and development of microalbuminuria,l6 most notably in poorly controlled diabetic subjects. In as much as hyperglycemia is associ- ated with increased GFR in diab8t8S,“,‘8 several Studies have suggested that glomerular hyperfiltration may be responsible for the initiation and progression of diabetic nephropathy.l*#20 Increased permeability of th8 glomeru- lar capillaries allows deposition of macromolecules in the mesangium, which eventually results in mesangial in- jury and glomerular sclerosis.21 A low protein diet, which has been shown to r8dUCe GFR,22 has been used in the prevention of glOmerUlar hyperfiltration in experimen- tal glomerulonephrosis.23 It therefore follows that dia- betic animals and humans subjected to a low protein diet should also show improvement in renal function.

In view of the abOV8, the present study was designed to examine the effect of a low protein diet on the rela- tionship of hyperglycemia, nonenzymatic glycation, and renal capillary structure and function, concurrently, in experimentally-induced diabetic rats.

METHODS

AnlIMk Thirty-five male Sprague-Dawley rats (300- 350 g) were injected intraperitoneally with a single dOS8 of streptozotocin (65 mg/kg) dissolved in cold citrate buffer (pli, 4.5). Hyperglycemia (ie., blood glu- cose greater than 20 mmol/L) and glyCOSUria W8r8 present 3 days after injection in 25 animals, while the Other 10 remained normoglyc8mic (str8ptozotocin injected nondiabetic group). The 25 diabetic animals were divided into the following: One group consisted of diabetic euglycemic animals (n=12) maintained at a

COPELAND ET AL.

blood glucose level of 4.5-6.5 mmol/L by daily injec- tion of protamine-zinc insulin (Connaught Laborato- ries, Willowdale, Ontario). The dosage ranged from 2-14 units/day depending on the animal and was adjusted according to biweekly blood glucose monitoring. An- other group (n=13) was treated in a similar manner, with the exception that the blood glucose levels were main- tained at 19-22 mmol/L (diabetic hyperglycemic group). Age matched controls COnSiSted of animals (n=8) in- jected only with the citrate buffer. Animals from the aforementioned fOUr groups were placed on an 8% pro- tein isocaloric diet (ICN Biochemicals Inc., Cleveland, Ohio). Another group of Str8ptOZOtOCin-induced diabet- ics (n=lO) and their Controls, StreptOZOtOCin-injected nondiabetics (n=lO) Served as positive controls, being placed on standard rat chow. In our previous study,24 the mOSt affected group Of animals were the untreated diabetics. We therefore felt that an appropriate compar- ison in the present investigation would be against such a group, deemed “positive COntrOlS,” i.e., diabetic ani- mals fed standard rat chow, which contains 24% protein. Moreover, the rationale for including the streptozotocin injected nondiabetics on a 24% protein diet as “drug con- trols” is that there were no differences between th8Se types Of animals and nOninjeCted animals in our earlier study. All animals were housed in individual cages for 6 months with food and water available as necessary. There was a 36% mortality of animals from all groups for the duration of the study; the total number of SUrViV- ing animals at the end of the study period was 40. The experimental protocol is Outlined in Figure 1.

Twenty-four hour urine collections and blood samples were obtained for analysis of creatinine clearance, gly- Cat8d hemoglobin, and urinary albumin at 6 months of study. Body weights of all the animals were recorded weekly for the duration of the study and kidney weights were measured when animals were killed.

Functimul &dies Blood glucose was monitored bi- weekly in whole blood obtained by tail bleed and quanti- fied using Dextrosticks read on a glucometer (Ames Divi- sion, Miles Laboratories Ltd., ReXdale, Ontario).

Glycated hemoglobin was measured by affinity chro- matography using a commercially available kit (GlyCO-

Experimental Design

Animals at Beginning of Study (n=63)

Age Matched Controls Diabetic Euglycemics Drug Controls Diabetic Hyperglycemics

Group I Low Protein Group 111 Low Protein (n=20) (n=23)

(n=3) (n=12)

A A

croup II Group VI Group IV Group V

Low Protein High Protein Low Protein High Protein

(n=lO) (n=lO) (n=13) (n=lO)

FIG. 1. Experimental protocol. At the start of the study experimental animals were allocated to six groups: age matched controls on 8% protein, drug controls on 8% and 24% protein, diabetic euglycemics on 8% protein, and diabetic hyperglycemics on 8% and

24% protein diets.

LOW PROTEIN DIET AND GLYCATION 115

Test, Pierce Chemical Co., Rockford, IL), as previously described.25 The between run coefficient of variation for this method was less than 10%.

Renal function was evaluated by creatinine clearance, which was expressed in relation to kidney weight at 6 months of study. Serum and urine creatinine were measured by an autoanalyzer (Beckman Astra, Beckman Inc., Brea, CA) using an alkaline picrate method.26

Urinary albumin was measured by radioimmunoassay as previously described,27.28 with the results being ex- pressed in pg/hr. The intra- and inter-assay coefficient of variations were 3.2 and 6.2% respectively.

Procurement of Tissues After 6 months of study, the an- imals were anesthetized with sodium pentobarbital and blood drawn by cardiac puncture for subsequent anal- ysis of creatinine, glycated hemoglobin, and blood glu- cose. Kidneys were rapidly excised, a slice of the left kidney obtained for electron microscopy, and the re- maining tissue placed in cold isotonic saline for sub- sequent isolation of the glomeruli.

Isolation of Glomerular Capillary Basement Membrane Glomer- uli were isolated from renal cortex by differential siev- ing through a series of nylon meshes according to the method of Cohen et a/,,29 with the material on the 88 pm and 105 pm screens collected as glomeruli. The base- ment membrane was isolated by osmotic lysis and se- quential detergent extraction using 3% triton-X 100, de- oxynuclease digestion, and sodium desoxycholate, as described by Carlson et a1.30 Although the amount of glomerular basement membrane isolated from each an- imal was not determined in this study, it was previously determined that approximately 2 mg of protein (as mea- sured by Lowry et a/.31) was isolated per kidney and was found to exhibit the characteristic electrophoretic pat- tern of pepsin digested type IV collagen as shown by previous investigators.13

Measurement of Menenzymatic Glycation The level of nonen- zymatic glycation was assayed for by borohydride re- duction and boronic acid affinity chromatography as previously described.32-34 Two milligrams of lyophyl- ized basement membrane was suspended in 200 ~1 of 0.10 N NaOH containing 625 &i of tritiated sodium borohydride [specific activity 500 mCl/mmol (Amersham Canada Ltd., Oakville, Ontario), and 12.5 mg of unla- beled sodium borohydride.] Reduction was carried out for 90 minutes on ice in a fumehood and was stopped by the addition of 4 ml 10% (w/v) trichloroacetic acid (TCA). The reduced basement membrane was washed a number of times with 10% TCA, followed by centrifu- gation and aspiration of the supernatant. The washed pellet was hydrolyzed with 1.5 ml of 6 M HCI for 16 hr at 110°C in sealed screw-top tubes. The pH was ad- justed to 8.5 by the addition of concentrated NaOH. The amino acid concentration of the hydrolysate was deter- mined by the ninhydrin procedure,35 using a standard curve produced with known concentrations of leucine; results were expressed in leucine equivalents. Four hun- dred microliters of the hydrolysate also was loaded onto

a m-aminophenylboronic acid column (Glycogel GSP columns, Isolab Inc., Akron, Ohio) and washed with 15 ml of 50 mfvl (Na),HPO,. The bound amino acids were removed by the addition of 2 ml 0.10 M HCI. One milliliter of the eluate was counted on a LKB 12-19 scin- tillation counter (LKB, Turko, Finland) using a quench corrected counting program. The results were expressed as CPM/pm01 leucine equivalents.

Morphologic Studies Small pieces of kidney cortex were fixed in 3% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4) for 2 hours at 4OC. Tissues were rinsed for 24 hours at 4OC in 0.1 M phosphate buffer (pH = 7.4) containing 0.2 M sucrose. The tissues were then postfixed for 2 hr at 4OC in 1% osmium tetroxide in 0.1 M phosphate buffer (pH = 7.4), dehydrated in as- cending concentrations of ethanol, and embedded in Epon 812. Thick sections were stained with toluidine blue and examined for routine orientation and assess- ment for the pcesence of glomerulosclerosis. Thin sec- tions were stained with uranyl acetate and lead citrate, viewed and photographed in a Philips EM 201 electron microscope. Quantification of kidney glomerular base- ment membrane was performed using the orthogonal intercept method of Jensen et a/.36 The actual measure- ments were carried out on micrographs (magnification, X20,000) with an electronic planimeter (Hewlett Packard digitizer, model 9874A) equipped with an electrosen- sitive cursor connected to a Hewlett Packard 9875A calculator/computer. The total number of micrographs assessed was 875. Glomerular capillary basement membrane thickness for each rat was measured on 25 randomly selected micrographs.

Statistics The experimental data were analyzed by the unpaired t-test to compare the means of the various groups. The level of significance was p = 0.05. The sta- tistical analysis was performed on a commercial package (Crunch Interactive Statistics Package (CRISP), Crunch Software, San Francisco, CA). The statistical analysis for the morphologic studies was by ANOVA and Tukey’s procedure.3’

RESULTS The functional and metabolic data for the animals at the completion of the study are presented in Table 1. The 8% and 24Oh protein diet hyperglycemic groups had a mean blood glucose level of 17.58 f 0.33 m and 18.38 f 0.62 mmol/L, respectively, which was significantly el- evated (p < 0.01) over the other groups: age-matched controls, streptozotocin nondiabetics, and diabetic eu- glycemics. There was no significant difference between the mean blood glucose levels of the former two or latter four groups. Similarly, both hyperglycemic groups had glycated hemoglobin values that were significantly ele- vated (p < 0.01) over the other four groups of animals.

The mean body weights of the 8% protein diet fed di- abetic hyperglycemic animals were 473 f 51 g, which was significantly decreased (p < 0.01) when compared with the diabetic euglycemics but not different from the noninjected and drug control animals. Moreover, the

116 COPELANO ET AL.

TABLE 1 Metabolic and Functional Parameters for Animals on 8% and 24% Protein Diets at Six Months of Study*

Corrected Creatlnine Clearance

Weight Blood Glucoee Urinary Albumin

o/o Glycated Kidney Weight Excretion Group (a) (mmol/l) Hemoglobin (a)

(ml/mln/gm (#g/hr) kldney weight)

Group I 518 f 92 7.10 f 1.83 4.43 f 1.04 1.31 f 0.14 2.40 f 1.04 0.27 * 0:lS (n = 8)

Group II 525 f 66 8.80 f 0.73 4.32 f 1.27 1.40 f 0.13 3.33 f 2.71 0.25 * 0.16 (n = 6)

Group Ill 660 f 84t$ 8.78 f 0.80 4.66 f 1.05 1.48 f 0.10 12.29 f 8.16 0.51 f 0.30 (n = 6)

Group IV 473 f 115 17.58 f 0.333 17.40 f 4.418 2.27 f 0.345 138 f 180 0.53 * 0.29 (n = 7)

Group V 592 f 5611 16.38 f 0.625 15.93 f 4.665 2.24 f 0.365 1932 f 17OO!jll 1.06 f 0.49#7 (n = 7)

Group VI 628 f 39t 7.94 f 1.06 5.37 f 1.07 1.89 f 0.08”tt 236 f 242 0.48 f 0.14 (n = 8)

Group I, 8% citrate buffer injected controls; Group II, 8% drug injected non-diabetics; Group Ill, 8% diabetic euglycemics; Group Iv, 8% diabetic hyperglycemics; Group V, 24% diabetic hyperglycemics; Group VI, 24% drug injected non-diabetics.

‘Values expressed as mean f SD; Wignifiwntly different at p < 0.01 from other groups; tsignificantly different at p < 0.05 from groups I, II; $Significantlydifferentatp<O.Ol fromgroup IV; II Significantlydifferentatp<O.O5fromgroups IV;#Significantlydifferent atp<O.Ol fromgroups I, II;1Significantlydifferentatp<O.O5from groupslll, IV: l * Significantlydifferentatp<O.Ol fromgroups), II, III; and ttsignificantly different at p < 0.05 from groups IV, V.

diabetic euglycemic animals had greater weights (p < 0.05) than either of the control groups on the same diet. Moreover, kidney weights of the diabetic hyperglycemic groups were significantly increased (p < 0.01) over that seen in the diabetic euglycemic, age-matched con- trols, and streptozotocin nondiabetic animals. In con- trast, there were no statistical differences among the latter three groups. Body weights for both groups of animals fed a 24% protein diet were higher than those seen in the hyperglycemic, but not in the other groups of animals fed an 8% protein diet. Kidney weights for the 24O/6 protein hyperglycemic groups were higher (p < 0.01) than values seen for controls and euglycemics but not different from hyperglycemics fed the 8% protein diet. On the other hand, kidney weights for drug controls fed the 24% protein diet were higher (p < 0.01) than con- trols and euglycemics fed an 8% protein diet but lower (p < 0.05) than both hyperglycemic groups.

The mean glomerular filtration rate (GRF), as mea- sured by creatinine clearance, was not statistically differ- ent among the four groups of animals on the 8% protein diet. Assessment of urinary albumin profiles showed a marked variability in the range of values within each group, most notably in the untreated diabetic animals. However, even though there was a trend toward elevated levels in three of the seven diabetic animals, thevalues for this group as a whole did not approach statistical significance overthe other groupsof animals. In contrast, the hyperglycemics, but not drug control animals, fed the 24% protein diet had significant elevation of creatinine clearance and urinary albumin excretion.

The level of nonenzymatic glycation of glomerular basement membrane isolated from the animals in the present study is shown in Table 2. Both groups of hy- perglycemic diabetic animals had significantly elevated values (p < 0.01) in comparison to the other four groups. There was no statistical difference in nonenzymatic gly-

cation among control and diabetic euglycemic groups of animals.

The morphometric data representing glomerular base- ment membrane thickness is shown in Table 3. As can be seen, there were no statistical differences in this pa- rameter among the various groups of animals. Moreover, routine histologic assessment of the glomeruli, i.e., pres- ence of glomerulosclerosis, revealed no apparent struc- tural abnormalities of these units among the six groups of animals.

TABLE 2 Nonenzymatic Glycatlon of Glomerular Capillary Basement Membrane in Animals on 8% and 24% Protein Diets at Six Months of Study*

Group

Group I (n =6)

Group II (n = 6)

Grouo III (n A 6)

Group IV

Level of nonenzymatic glycation of glomerular basement membrane

(CPMlrmol leucine equivalents)

668 f 140

548 k 167

556 ZIZ 159

1265 k 273t (n = 7)

Group V (n = 7)

Group VI (n = 6)

1151 f 272t

354 f 204

Group I, 8% citrate buffer injected controls; Group II, 8% streptozotocin injected non-diabetics; Group III, 8% diabetic euglycemics; Group IV, 8% diabetic hyperglycemics; Group V. 24% diabetic hyperglycemics; Group VI, 24% drug injected non-diabetics.

‘Values expressed as mean + SD. tsignificantly different at p < 0.01 from other groups.

LOW PROTEIN DIET AND GLYCATION 117

TABLE 3 Morphometric Analysis of Glomerular Capillary Basement Membrane Thickness for Animds on 8% and 24% Protein Diets at Six Months of Study*

Group

Group I (n = 6)

Group II (n = 6)

Group 111 (n = 6)

Group IV (n = 7)

Group V (n = 7)

Group VI (n =8)

Besement Yembrene Thickness (microns)

0.164 f 0.004

0.162 f 0.002

0.180 i 0.002

0.194 f 0.002

0.207 f 0.007

0.191 f 0.009

Group I, 8% citrate buffer injected controls; Group II, 8% drug injected non-diabetics; Group III, 8% diabetic eugiycemics; Group IV, 8% diabetic hyperglycemics; Group V, 24% hypergly- cemics; Group VI, 24% drug injected non-diabetics.

‘Values expressed as mean f SEM.

DISCUSSION It has been proposed that the beneficial effects of a low protein diet on renal disease may be attributed to a reduction in protein intake, which in turn lim- its the hemodynamic changes occurring in surviving nephrons,22J3,38J9 Such studies were designed to in- crease GFR by renal ablation or experimentally induced glomerulonephritis. It was noted that these procedures resulted in altered hemodynamic patterns, namely hy- perfiltration and hyperperfusion, whereby the remaining glomeruli compensated for the loss of adjacent func- tional renal mass by increasing their blood flow and fil- tration rate. Animals placed on low protein diets dem- onstrated a reversal of the above, concurrent with an attenuation of glomerular structural abnormalities.

An increase in GFR is a common occurrence early in diabetes mellitus.15 The present study was therefore de- signed to assess whether a low protein diet would af- fect not only GFR early on in diabetes but, in addition, other features associated with this disease, such as renal hypertrophy,ls proteinuria,l6 and glomerular structural alterations.24 The diabetic hyperglycemic group fed an 8% protein diet exhibited increased kidney weight and nonenzymatic glycation of glomerular basement mem- brane as compared to the insulin-treated diabetic and control animals fed a similar diet. On the other hand, there was no statistical differences in GFR, measured by creatinine clearance and urinary protein excretion, among the various groups on the 8% protein diet. This is in contrast to the results seen in the diabetic animals fed a 24% protein diet as well as in the observations from our earlier study in which rats were fed standard rat chow. In our previous study, just as in our present group of hyperglycemic rats on a 24% protein diet, there was an elevated creatinine clearance, consistent with the early stages of diabetic nephropathy24 and enhanced urinary

protein excretion over values seen in insulin treated and

control animals. Taken together, these observations Sug- gest that protein restriction may be beneficial in ame- lioration of renal dysfunction in diabetes despite persis- tent hyperglycemia. These findings are supported by the studies of Hostetter et a/22 and Neugarten et 8Lz3 In the study by Hostetter et a/,22 it was noted that restriction of dietary protein prior to renal ablation blunted the subse- quent increase in glomerular capillary pressure, as well as structural lesions, seen in renal ablated rats main- tained on standard chow containing a high protein con- tent. Neugarten et a/,23 studying glomerulonephritis in- duced in rats by injection of goat gamma globulin, noted that animals on a low protein diet showed normal serum creatinine, an absence of proteinuria, and amelioration of glomerular structural changes. This was in marked contrast to high protein fed nephritic rats that mani- fested proteinuria, decreased creatinine clearance, an elevation in serum creatinine and glomerular structural alterations. From the aforementioned studies, it would appear that increased protein excretion may be a func- tion of the higher ingested protein load or, more likely, that a higher ingested load constituted a greater stress to the kidney, overwhelming its compensatory reserve. Our studies showed that while hyperglycemia per se continues to be nephrotoxic, as indicated by an equal increase in renal mass between the two hyperglycemic groups, a similar protection from hyperperfusion (as re- flected by no change in creatinine clearance among the various groups of animals on 8% protein) could be in- duced by a low protein diet. Thus, such a diet prevents the hyperglycemia-induced increase in blood flow to nephrons and may thereby protect these functional units from subsequent damage and possibly delay the onset of renal failure.

No differences in glomerular basement membrane thickness was found among the six groups of animals tested despite the higher level of nonenzymatic glycation in the untreated diabetics on either diet. Similar morpho- logic results were noted in our previous study, coupled with the early impairment of renal function in the diabetic hyperglycemic group, as measured by increased crea- tinine clearance and urinary albumin excretion.24 These findings suggest that basement membrane thickness per se may not be important in the development of incipient diabetic nephropathy. This is consistent with the work of Mauer et a1.,4 which suggested that expansion of the mesangium, rather than glomerular basement membrane thickening, leads to glomerular functional deterioration in diabetes by restricting the glomerular capillary vascu- lature and its filtration surface.

The degree of nonenzymatic glycation of basement membrane was statistically higher in both groups of diabetic animals maintained hyperglycemic irrespective of protein load as compared to controls and the eu- glycemic diabetics. This is not surprising because nonen- zymatic glycation of protein is directly related solely to the degree of hyperglycemia. In addition, the increase in nonenzymatic glycation found was at approximately the same level in the present investigation as that re- ported in our previous study on diabetic animals fed a high protein diet.24 It has been suggested that increased

COPELAND ET AL.

levels of nonenzymatic glycation of protein in diabetics could lead to structural and functional changes, includ- ing changes in net charge.B,‘* The removal of the posi- tive charge on hydroxylysine by nonenzymatic glycation may prevent interchain cross-linking” of type IV colla- gen, the major component of glomerular basement mem- brane. This could interfere with the packing of collagen fibrils, resulting in increased pore size and subsequent increased membrane permeability. The latter could re- sult in increased passage of albumin through the filtra- tion barrier, resulting in microalbuminuria. Such a hy- pothesis would seem untenable from our present and previous studies, as hyperglycemic groups fed different protein diets, although having similar degrees of nonen- zymatic glycation, exhibited significantly different levels of microalbuminuria.

In summary, therefore, keeping in mind that renal function did not appear compromised in low-protein fed diabetics in the present study but did show changes in high-protein fed diabetics in the current and previous studies,*’ our present investigation leads us to suggest that nonenzymatic glycation, like basement membrane thickening, may not play an important role in the on- set and/or progression of nephropathy. What does ap- pear relevant from the results of the present study is that restricting dietary protein, coupled with good glycemic control, may serve to prevent glomerulopathy in high risk groups of individuals such as Type I diabetics.

ACKNOWLEDGMENTS

This work was supported by grants received from the Canadian Diabetes Association and the Manitoba Health Research Council by Dr. R.W. Yatscoff. We wish to thank Ms. V. Saunders and Ms. D. Love for their excellent technical assistance.

REFERENCES

1. Eleatar S, Bergman M, Felig P: Endocrine pancreas: diabetes mellitus, in Felig P, Baxter JD, Broadus AE, Frohman LA (eds), Endocrinology and Metabolism. New York, McGraw-Hill, 1987, pp. 1043-1178.

2. Friedenwald JS: A new approach to some problems of reti- nal vascular disease. Am J Opthamo/32:48?-498, 1949.

3. McManus JFA: The periodic acid routine applied to the kidney. Am J P&ho/ i4:843-654, 1948.

4. Mauer SM. Steffes MW. Ellis EN. Sutherland DER. Brown DM, Goetz FC: Stru&ral-functibnal relationships in dia- betic nephropathy. J C/in Invest 74:1143-l 155, 1984.

5. Farquhar MG, Hooper J, Moon HD: Diabetic glomeruloscle- rosis: electron and light microscopic studies. Am J Pathol 351721-754. 1959.

6. Siperstein MD, Unger RH. Madison LL: Studies of muscle capillary basement membrane in normal subjects, diabetic and prediabetic patients. J C/in invest 47:1973-1999, 1968.

7. Day JF, Thorpe SR, Baynes JW: Non-enzymatic glycosy- lation of rat serum proteins in vitro and in vivo. Fed Proc 38:418, 1979.

8. Kennedy L, Baynes JW: Non-enzymatic glycosylation and the chronic complications of diabetes: an overview. Dia- betologia 26:93-98, 1984.

9. Kirschenbaum DM: Glycosylation of proteins: its compli- cations in diabetic control and complications. Pediatr C/in North Am 31:611-621, 1984.

10. Bunn HF, Haney DN, Kamin S, Gabbay KH, Gallop PM: The biosynthesis of human hemoglobin A,,. slow glycosylation of hemoglobin in vivo. J C/in invest 5771652-1659, 1976.

11. Shaklai N. Garlick RL. Bunn I-IF: Nonenzvmatic alvcosvla- tion of human serum.albumin alters its confor&Gon Hnd function. J Viol Chem 259:3812-3817, 1984.

12. Means YE, Chang MK: Nonenzymatic glycosylation of pro- teins: structure and function changes. Diabetes 31:1-4, 1982.

13. Trueb 8, Fluckiger R, Winterhalter KH: Nonenzymatic clyco- sylation of basement membrane collagen in diabetes melli- tus. Collagen Relat Res 4:239-251, 1984.

14. Cohen MP, Urdanivia E, Surma M, Ciborowski CJ: Nonenzy- matic glycosylation of basement membranes, in vitro stud- ies. Diabetes 31:367-371, 1981.

15. Mogensen CE: Renal functional changes in diabetics. Dia- betes 25(2):872-879, 1976.

16. Viberti GC, Pickup JC, Jarrett RJ, Keen H: Effect of control of blood glucose on urinary albumin excretion of albumin and BBmicroglobulin in insulin-dependent diabetes. N Engl J Med 300:638-641, 1979.

17. Mogensen CE: Glomerular filtration rate and renal plasma flow in normal and diabetic man during elevation of blood suaar levels. Stand J C/in Lab invest 28:177-182. 1971.

18. Blintz RC, Tucker BJ, Bushwa L, Peterson OW: Mechanism of diuresis following acute modest hyperglycemia in the rat. Am J Physio/244:F185-F194, 1982.

19. Hostetter TH, Troy JL, Brenner BM: Glomerular hemo- dynamics in experimental diabetes mellitus. Kidney Int 19:410-415, 1981.

20. Hostetter TH, Rennke HG, Brenner BM: The case for in- trarenal hypertension in the initiation and progression of diabetic and other glomerulopathies. Am J Med 72:375-380, 1982.

21. Olson JL, Hostetter TH, Rennke HG, Brenner BM, Venkat- achalam MA: Altered glomerular permselectivity and pro- gressive sclerosis following extreme ablation of renal mass. Kidney Int 22:112-126, 1982.

22. Hostetter TH, Olson JL, Rennke HG, Venkatachalam MA, Brenner BM: Hyperfiltration in remnant nephrons: a po- tentially adverse response to renal ablation. Am J Physiol 241 :F85-F93, 1981.

23. Neugarten J, Feiner HD, Schacht RG, Baldwin DS: Amelio- ration of experimental glomerulonephritis by dietary protein restriction. Kidney Int 24:595-601 I 1983.

24. Copeland KR. Yatscoff RW, Thliveris JA. Mehta A, Penner B: Non-enzymatic glycation and actual renal structure and function in the diabetic rat. Kidnev Int 32:664-670, 1987.

25. Yatscoff RW, Tevaarwerk GJ, Ciarson CL, Warlock LM: Evaluation of an affinity chromotographic procedure for the determination of glycosylated hemoglobin (HbA,). C/in Biochem 16:291-295, 1983.

26. Fabing DL, Ertingshausen G: Automated reaction-rate method for the determination of serum creatinine with the Centrifichem. C/in Chem 17:696-699, 1971.

27. Miles DW. Moaensen CE. Gundersen HJG: Radioim- munoassay for irinary albumin using a single antibody. Stand J C/in Lab invest 26:5-l 1, 1970.

28. Woo J, Floyd D, Cannon DC, Kahan B: Radioimmunoassay for urinarv albumin. C/in Chem 24:1464-1467. 1978.

29. dohen lip, Carlson EC: Preparation and analysis of glomerular basement membrane, in Larner J, Pohl SL (eds), Methods in Diabetes Research, New York, John Wiley and Sons, Inc., 1984, pp. 357-375.

30. Carlson EC, Brendel K, Hjelle JT, Mezan E: Ultrastructural and biochemical analyses of isolated basement membrane from kidney glomeruli and tubules and brain and retinal microvessels. J Ultrastructural Res 62:26-53, 1978.

31. Lowry OH, Rosebrough NJ, Farr AL, Randall RS: Protein measurement with the folinphenal reagent. J 5iol Chem 123:265-275, 1951.

32. Shaklai N. Garlick RL. Bunn HF: Nonenzymatic glyCOsYla- tion of human serum.albumin alters its conformaiion &d function. J Biol Chem 259:3812-3817, 1984.

33. Garlick RL, Mazer JJ: The principle site of nonenzymatic

LOW PROTEIN DIET AND GLYCATION

glycosylation of human serum albumin in vivo. J Biol Chem 258:8142-8146, 1983.

34. Yue OK, McLennan S, Handlesman DJ, Delbridge L, Reeve 38. T, Turtle JR: The effect of salicylates on nonenzymatic gly- cosylation and thermal stability of collagen in diabetic rats. Diabetes 331745-751, 1984. 39.

35. Moore SJ: Amino acid analysis: Aqueous dimethyl sulfoxide solvent for the ninhydrin reaction. J Biol Chem 243:8281- 6283, 1968. 40.

of Statistics in Biological Research. San Francisco, W.H. Freeman and Company, 1984. El Nahas AM, Masters-Thomas A, Brady SA, et al: Selective effect of low protein diets in chronic renal diseases. kit Med J 289:1337-1341, 1984. Meyer TW, Anderson S, Rennke HG, Brenner BM: Reversing glomerular hypertension stabilizes established glomerular injury. Kidney Int 31:752-759, 1987. Mogensen CE, Christensen CK: Predicting diabetic nephro-

36. Jensen EB, Gundersen HJG, Osterby R: Determination pathy in insulin-dependent patients. N Engl J Med 311:89-

of membrane thickness distribution from orthogenal inter- 93, 1984.

cepts. J Microscopy 115:19-33, 1978. 41. Ireland JT, Vibreti GC, Watkins PJ: The kidney and urinary

37. Sokal RR, Rohlof FF: Biometry. The Principles and Practice tract, in Keen H, Jarrett J (eds), Complications of Diabetes, London, Edward Arnold, 1982, pp. 137-178.