Hyperglycemia and development of glomerular pathology: Diabetes compared with galactosemia

Kidney International, Vol. 36 (1989), pp. 41-45

Hyperglycemia and development of glomerular pathology: Diabetes compared with galactosemia RONALD L. ENGERMAN and TIMOTHY S. KERN Department of Ophthalmology, University of Wisconsin, Madison, Wisconsin, USA

Elevation of the blood concentration of an isomer of glucose, D-galactose, is known to be capable of reproducing many of the complications of diabetes in the retina, lens, peripheral nerve, and cornea [10, 11]. Retinal microaneurysms and other vascular lesions characteristic of diabetic retinopathy are elicited in dogs by experimental galactosemia [12]. Renal histology, in contrast,

Hyperglycemia and development of glomerular pathology: Diabetes compared with galactosemia. Dogs were randomly assigned to experimental galactosemia or diabetes, or to a normal untreated group, and diabetic animals were then randomly assigned to either poor or good glycemic control. At five years duration, kidneys from the animals were compared by quantitative stereology. Glomerulopathy appeared in the poor control diabetes group, and the thickness of glomerular capillary basement membrane, the glomerular tuft volume, and the fraction of glomerulus occupied by mesangium were each significantly greater than normal. The capillary filtering surface area per glomerulus was supra-

has seemed upon light microscopy to be little affected [13]. Experimental galactosemia of five years duration apparently failed to elicit the nephropathy seen in dogs diabetic equally long, notwithstanding elevations of tissue polyol as great or

normal also, but nonetheless was subnormal relative to glomerular volume. The development of glomerulopathy was significantly inhibited in dogs assigned to good glycemic control. In galactosemic animals, the

greater than seen in diabetic dogs. In the present report, effects

basement membrane thickness was greater than normal, but the gbmerular volume, fractional and absolute volumes of mesangium, and

of long-term galactosemia on the ultrastructure of the renal glomerulus have been quantitatively evaluated in dogs and compared with the effects of diabetes.

capillary filtering surface area remained normal. The polyol concentration in renal cortex seemed elevated by galactosemia no less than by diabetes, and was highest in galactosemia. The galactosemic animals are known to have developed a retinopathy morphologically compara-

ble to that of diabetic patients and diabetic dogs. Thus, sequelae of hyperglycemia sufficient to produce glomerular basement membrane

Methods

Healthy adult dogs (1-½ to 2-'/2 years old) were randomly

assigned to be made diabetic or fed a 30% galactose diet [12], or to remain as normal untreated (control) animals. The dogs were tes. random source animals of mixed breeds. Diabetes was induced by injection of alloxan monohydrate intravenously, and after a brief interval of hyperglycemia to insure that the animals were Hyperglycemia, the principal symptom of diabetes mellitus, comparably insulin-deficient, the diabetics were randomly asis of uncertain significance in the etiology of diabetic nephrop- signed to either poor glycemic control or good glycemic control athy. Elevation of the blood glucose concentration can alter as described previously [14]. The poor control group received renal function, and has been shown to increase glomerular unlimited food with insulin once daily at a dosage inadequate to thickening and retinopathy proved not necessarily sufficient to produce the mesangial expansion and gbomerular hypertrophy typical of diabe-

filtration rate (GFR) and renal plasma flow (RPF), and via prevent chronic hyperglycemia and glucosuria, whereas the nonenzymatic glycation of proteins, might increase the binding of plasma proteins to glomerular basement membrane and alter protein excretion [1—5]. Elevation of tissue polyoi concentration, another of the sequelae of hyperglycemia, possibly contributes to such functional changes, and diabetes-induced abnormalities of renal function in rats reportedly may be inhibited by pharmacologically inhibiting polyol production [6, 7]. To what extent the anatomic renal lesions of diabetes are a result of

good control group received a measured diet with insulin twice daily at dosages chosen so as to minimize hyperglycemia and glucosuria. Hemoglobin A1 (HbA1) was measured two to four times per year using minicolumns (Isolab, Inc., Akron, Ohio), after incubation to remove labile adducts. Blood glucose and plasma galactose concentrations have been reported previously [12].

Morphometry hyperglycemia and its sequelae is less clear. Thickening of At 60 months of study, animals were anesthetized with glomerular basement membrane has been reportedly produced in nondiabetic mice by injections of nonenzymatically glycated pentobarbital and, after tissue samples had been removed, were killed. Samples of renal cortex from mid-kidney were fixed for protein, but this finding has been disputed [8, 9].

light microscopy in 10% neutral formalin, and for electron microscopy in 4% phosphate buffered glutaraldehyde followed by post-fixation in 1% osmium tetroxide. Ultrathin sections of

Received for publication October 17, 1988 and in revised form February 16, 1989 Accepted for publication February 22, 1989

the centermost gbomerulus from three different epon blocks were placed on Butvar-coated copper slot grids (1 X 2 mm), stained with uranyl acetate and lead citrate, and viewed in a

© 1989 by the International Society of Nephrology 41

42

47

a-

:4-

r.

'I.

'

Wa

r

_

Engerman and Kern: Hyperglycemia and glomerular pathology

Fig. 1. Glomerular capillaries from dogs at 60 months of study: (a) normal, (b) galactosemia, (c) alloxan diabetes with poor glycemic control, (d) alloxan diabetes with good glycemic control. The basement membrane (r) of peripheral capillaries is thicker than normal in galactosemia and in poorly controlled diabetes. Large podocyte processes such as shown in Ic were common in poorly controlled diabetics, and possibly represent fusion of foot processes. The micrographs illustrate the magnification used for the morphometric analyses, but include only 14% of the area of each print analyzed. (x 17,270)

JEOL 100 CX electron microscope (JEOL, Tokyo, Japan). Glomeruli were randomly entered, and then 15 to 25 evenly

Table 1. Glycated hemoglobin and glycosuria of diabetic dogs and experimentally galactosemic dogs studied 5 years (mean SD)

spaced micrographs were taken and printed at a final magnification of about 18,000x [151. A calibration grid was photographed each day along with the glomeruli. Obliterated glomeruli were not studied (although none were encountered in these Normal epon blocks). Diabetes Glomerular volume was determined from sections (4 pm) of Good control Poor control formalin-fixed paraffin-embedded kidney. The tissue sections were stained with periodic acid-Schiff and hematoxylin, and Galactosemia

Hemoglobin

N

A1

%

9

5.9

0.3

6

6.1 10.8

0.7

6.8

0.2

7 5

1.3

Glycosuria % sugarg/kg/day free days

0

100

<1 10.8 2.5

II

68

4.6 0.4

0

0

viewed in a microscope with a camera lucida attachment. Outlines of glomerular tuft from 25 sequential glomeruli were traced onto paper, and the area of the tuft profile was determined using computer-assisted morphometry. Distances from a calibrated micrometer slide also were traced onto paper for use in calibrating the glomerular area measurements. The mean glomerular volume was calculated from mean glomerular area as described by Hirose and coworkers [16]. Fractional volume (volume density) of each glomerular component was determined from the electron photomicrographs

using a 77 point grid [15]. Total volume per glomerulus of each glomerular component was calculated from fractional volumes

and average glomerular volume for each animal. Capillary filtering surface area relative to glomerular volume (surface density of the filtration surface, Sly), and capillary length relative to glomerular volume (LIV) were determined by standard stereological methods [15, 17]. Capillary filtering surface area per glomerulus and capillary length per glomerulus were

43

Engerman and Kern: Hyperglycemia and glomerular pathology

Table 2. Glomerular m orphometry in diabetic dog s and experimentally gal actosemic dogs studied 5 years (mean

SD)

Diabetes Normal

Good control

Poor control

9

6

7

N Glomerular tuft vol

— x 106

2.0

0.6

2.5

0.9

5.1

1.0°

0.31

0.21

0.43

0.20

1.08

O.45°°

Galactosemia 5 1.9

0.24

0.4 0.13

Mesangium, volume pm3 X 106 Mesangium, volume % of glomerulus Mesangial matrix, vol % of glom Endothelium, vol

0.17

0.08

0.21

0.08

0.36 0.10

0.16

0.05

Epithelium, vol

0.36

0.13

0.42

0.19

0.74

0.24a,1

0.40

0.08

Capillary surface area mm2

0.33

0.09 0.02

0.38

0.12 0.02

0.60 0.l8

0.31

0.07

0.16

0.01

472

90b

14 6

18

3

25

93

Il

3

16 6

un3 x 106

.unh x 10

S/V m'

Capillary basement membrane width nm

0.17 361

76

0.16

416 34

10"

0.12 003ad 644

79a.c

11

4

8.0

3

a P < 0.01 compared with normal b P < 0.05 compared with normal P < 0.01 compared with diabetes good control d P < 0.05 compared with diabetes good control

calculated as the product of average glomerular volume and S/V or L/V, respectively. Thickness of glomerular peripheral capil-

absolute volume of mesangium, accordingly, also exceeded normal. The increase in the fractional volume of mesangium

lary basement membrane was determined by the orthogonal involved increases in the fractional volumes of both the mesanintercept method [181.

gial cells and matrix. The mesangium's fractional volume varied

Polyol accumulation In an ancillary experiment, dogs were made galactosemic (N = 3) or diabetic (poor glycemic control, N = 4) as described above or were kept as normal controls (N 4), and were killed

appreciably between animals, about 6 to 23% in normal dogs and 9 to 34% in poorly controlled diabetics. Other glomerular components in the poorly controlled diabetics (such as, capillary endothelium and visceral epithelium) also showed significant increases in absolute volume, and the capillary basement membrane was significantly thicker than normal. Capillary

after two to four months of study. Fresh kidney slices were frozen at —80°C for assay of polyol concentration in renal filtering surface density (Sly), was subnormal in the poor cortex. Polyol was assayed by gas chromatography in the case control group, but the reduction seems due to glomerular of galactosemia (mostly galactitol), or by enzymatic methods hypertrophy rather than to a loss of filtering surface area. The (to achieve the desired sensitivity for sorbitol) in tissue from filtering surface's absolute area in the poor control group was, diabetic and normal animals [19, 201. The level of renal sorbitol in fact, greater than normal. The ratio of capillary length to was similar in the two methods. glomerular volume (L/V) was not abnormal in any group (data Group comparisons were made using parametric methods

not shown). (ANOVA followed by a Tukey-Kramer test) or nonparametric Good glycemic control inhibited the development of the methods (Kruskal-Wallis followed by Mann-Whitney U tests of HbA1 concentration and tissue polyol concentration). Data are above glomerular abnormalities in diabetes. The good control group differed significantly from the poor control group, and not reported as mean SD. from normal, with respect to basement membrane thickness, Results glomerular volume, capillary filtering surface area per glomerDiabetic dogs assigned to the poor control group developed ulus, capillary length per glomerulus, S/V ratio, and volume chronic hyperglycemia, glycosuria, polyuria, and hyperphagia, occupied by endothelium and epithelium (Table 2). The fraction and their mean HbA, concentration for the five years of study of glomerulus occupied by mesangium in the good control group was significantly greater than normal (P < 0.01, Table 1, Fig. 1). was less than in poor controls, although not to a statistically In diabetic dogs assigned to good control, such metabolic significant degree. In contrast to the severe glomerulopathy seen in chronically abnormalities were absent or minimal and HbAJ was comparahyperglycemic diabetic dogs, the renal glomerulus of dogs ble to normal. Nondiabetic animals fed the galactose-rich diet showed levels of HbA, significantly greater than normal (P < galactosemic for five years seemed little affected. The glomer0.05) although less than the value for diabetics in poor control. ulus in galactosemia was not hypertrophic and was not signifiPoor control of diabetes resulted in glomerular hypertrophy cantly different from normal with respect to either the fractional and mesangial expansion (Table 2). The volume of the glomer- volume or absolute volume (per glomerulus) of mesangium, ulus greatly exceeded that of normal dogs and, moreover, the endothelium or epithelium. Neither the S/V ratio nor the total mesangium was disproportionately enlarged and occupied a capillary filtering surface area was abnormal. Nevertheless, the greater than normal fraction of the glomerular volume. The capillary basement membrane was significantly thicker than in

44

Engerman and Kern: Hyperglycemia and glomerular pathology

Table 3. Polyol concentration in renal cortex of dogs diabetic (poor glycemic control) or galactosemic 2 to 4 months (mean SD)

N Normal Diabetes Galactosemia a Measured

4 4 3

nmol/mg proteina 0.23 0.43 6.0

0,05

0.08 1.2

as sorbitol in normals and diabetics, and galactitol in

galactosemics

normal animals. The capillary basement membrane was not as thick as in the diabetic poor control group (P < 0.01).

In an ancillary study of polyol concentration in the renal cortex, diabetes of two to four months duration was shown to result in tissue polyol (sorbitol) levels significantly greater than in normal dogs (P <0.05; Table 3). Galactosemic (nondiabetic) dogs had tissue polyol (galactitol) concentrations several-fold greater than those of normal or diabetic dogs. Discussion

mental galactosemia, and significantly less than in diabetes. Nonetheless, the basement membrane of glomerular capillaries is found by us to be significantly thickened by experimental galactosemia. This basement membrane thickening occurs in the absence of any increase whatever in mesangial matrix, a basement membrane-like material known to become excessive in diabetes. Accumulation of mesangial matrix therefore seems unlikely to be produced by the same sequelae of hyperglycemia that cause accumulation of basement membrane in the capillary wall. Similarly, abnormalities of the glomerular capillary basement membrane and mesangial matrix in diabetic animals have been reported to be not equally reversible by islet transplantation [15, 18].

Dogs made experimentally galactosemic for several years develop a retinopathy that is remarkably similar to that seen in diabetic patients and diabetic dogs [12]. The retinopathy includes saccular capillary aneurysms, pericyte "ghosts", nonperfused acellular capillaries, and hemorrhage. Thus, galac-

tosemia sufficient to produce diabetic-like lesions in the

glomerular basement membrane and retina has proved to be not Diabetic dogs in poor glycemic control develop a nephrop- sufficient to elicit several renal abnormalities that are typical of athy closely resembling that characteristic of diabetes in pa- diabetes, notably mesangial expansion, glomerular hypertrophy tients. The renal abnormalities include nephromegaly, mesan- and obliteration, and nephromegaly. Consumption of the galactose-rich diet results in elevation of gial expansion, glomerular obliteration, occasional exudative

deposits and mesangial nodules, and excessive excretion of the blood levels of an isomer of glucose (galactose) in the protein [13, 21—24]. In addition, the glomerular volume and apparent absence of many other metabolic abnormalities of capillary filtering surface area become greater than normal in diabetes, such as insulinopenia and supranormal blood levels of diabetic dogs, as is the case also in diabetic patients prior to the appearance of clinical nephropathy. Later in the course of the

glucose, fatty acids, branched-chain amino acids and fibrinogen

[12]. Diabetic-like lesions produced by feeding the galactose nephropathy in patients the filtering surface area becomes appear to be secondary to the hyperglycemia (galactosemia) reduced, an anatomic change yet to be demonstrated in ani- itself and its sequelae. mals, and associated with serious impairment of renal function Two sequelae of hyperglycemia, excessive non-enzymatic glycation of protein and excessive production or accumulation Good glycemic control from the onset of diabetes tends to of polyol, have received much attention in recent years, and the inhibit the development of nephropathy in rats [26—28] and, as is magnitude of each in our galactosemic animals appears to be clear from the present studies, also in dogs. The nephropathy unlike that in our insulin-deficient animals. Tissue levels of observed in animals kept in poor glycemic control cannot be non-enzymatically glycated protein, represented in the present attributed therefore simply to nephrotoxic effects of alloxan or study by HbA1, tend to be less in galactosemic dogs than in aging. These findings lead one to hope that the development of insulin-deficient dogs. This may be due in part to a difference in diabetic nephropathy in patients also might be inhibited by severity of hyperglycemia, since blood hexose levels in insulinimprovement of glycemic control, at least if good control is deficient dogs tend to remain elevated at all hours of the day, established soon after the onset of diabetes. The evidence that whereas in galactose-fed dogs the blood hexose rises postpranthe nephropathy can be inhibited by good glycemic control dially but subsequently approaches normal during sleep or provides no information as to whether the inhibition is due to fasting. Tissue polyol concentration, in contrast, tends to be greater correction of the hyperglycemia itself, or is due instead to the coincident correction of other metabolic sequelae of deficient in galactosemic dogs than in insulin-deficient dogs, as shown for insulin activity (dysmetabolism of lipids, proteins, etc.). Hyper- renal cortex in the present study and for erythrocytes in a prior glycemia, although an appropriate target for clinical treatment, study [13]. Excessive production and accumulation of polyols is of uncertain significance in the pathogenesis of nephropathy. has been implicated in capillary basement membrane thickening In contrast to the marked nephropathy observed in diabetic in diabetes and experimental galactosemia, inasmuch as the dogs after five years of hyperglycemia, kidneys from dogs thickening in retinal vessels of diabetic and galactose-fed rats experimentally galactosemic equally long seem little affected. reportedly can be prevented by administration of an aldose [25].

Glomerular volume as well as the volume of the mesangium and

reductase inhibitor [29—31]. Comparison of the effects of exper-

cellular components of the tuft, and the capillary filtering imental galactosemia and diabetes described herein and in a surface area are not abnormal in galactosemic dogs, and clearly

are less affected than in diabetes. The results of the present ultrastructural analysis are consistent with our observations by light microscopy [13]. By light microscopy, the mesangial area (estimated semi-quantitatively), frequency of obliterated gbmeruli, and kidney weight were found to be normal in experi-

previous report [131 raise a possibility, however, that mesangial expansion, glomerular hypertrophy and obliteration, and seemingly other renal lesions typical of diabetes are not a result of

excessive tissue polyol concentrations alone, and thus may respond less well than perhaps other diabetic complications to treatment with aldose reductase inhibitors.

Engerman and Kern: Hyperglycemia and glomerular pathology

45

Acknowledgments This study was supported in part by Public Health Service research

13. KERN TS, ENGERMAN RL: Kidney morphology in experimental hyperglycemia. Diabetes 36:244—249, 1987

grant EYOO300 from the National Eye Institute, and by a research grant

microvascular disease in diabetes to metabolic control. Diabetes

from the Juvenile Diabetes Foundation International. We sincerely appreciate the technical assistance of M. Garment, M. Larson, and H. Wabers, and invaluable advice received from J. Basgen and Dr. M. Steffes concerning the morphometric techniques.

14. ENGERMAN RL, BLOODWORTH JMB JR. NELSON S: Relationship of 26:760—769, 1977

15. STEFFES MW, BROWN DM, BASGEN JM, MAUER SM: Ameliora-

tion of mesangial volume and surface alterations following islet transplantation in diabetic rats. Diabetes 29:509—515, 1980 16. HIR0SE K, OSTERBY R, NOZAWA M, GUNDERSEN HJG: Develop-

ment of glomerular lesions in experimental long-term diabetes in Ophthalmology, University of Wisconsin-Madison, Medical Sciences Center, 1300 University Avenue, Madison, Wisconsin 53706, USA.

the rat. Kidney mt 21:689—695, 1982 17. ELIAs H, HENNIG A, SCHWARTZ DE: Stereology: Applications to biomedical research. Physiol Rev 51:158—200, 1971 18. STEFFES MW, BROWN DM, BASGEN JM, MATAS AJ, MAUER SM:

References

Glomerular basement membrane thickness following islet transplantation in the diabetic rat. Lab Invest 41:116-1 18, 1979 19. KERN TS, ENGERMAN RL: Hexitol production by canine retinal

1. WISEMAN Mi, VIBERTI GC, KEEN H: Threshold effect of plasma glucose in the glomerular hyperfiltration of diabetes. Nephron 38:

microvessels. Invest Ophthalmol Vis Sci 26:382—384, 1985 20. MALONE JI, KNOX G, BENFORD 5, TEDEsCO 5: Red cell sorbitol: An indicator of diabetic control. Diabetes 29:861—864, 1980 21. BLOODWORTH JMB JR, ENGERMAN RL: Experimental diabetic

Reprint requests to Ronald L. Engerman, Ph.D., Department of

257—260, 1984 2. CHRISTIANSEN JS, CHRISTENSEN CK, HERMAN5EN K, PEDER5EN

EB, MOGENSEN CE: Enhancement of glomerular filtration rate and

renal plasma flow by oral glucose in well controlled insulindependent diabetics, Scand J Clin Lab Invest 46:265—272, 1986 3. BROWNLEE M, PONGOR S, CERAMI A: Covalent attachment of soluble proteins by nonenzymatically glycosylated collagen. J Exp Med 158:1739—1744, 1983

4. WILLIAMS SK, SIEGAL RK: Preferential transport of non-enzymatically glucosylated ferritin across the kidney glomerulus. Kidney ml 28:146—152, 1985 5. GHIAGERI GM, CANIANO G, DELFINO G, BIANCHINI F, QuEIR0L0

C: Glycosyl albumin and diabetic microalbuminuria: Demonstration of an altered renal handling. Kidney mt 25:565—570, 1984 6. BEYER-MEARS A, CRUZ E, EDELIST T, VARAGEANNIS E: Dimin-

ished proteinuria in diabetes by Sorbinil, an aldose reductase inhibitor. Pharmacol 32:52—60, 1986 7. GOLDFARB S, SIMMoNs DA, KERN EFO: Amelioration of glomer-

ular hyperfiltration in acute experimental diabetes mellitus by dietary myo-inositol supplementation and aldose reductase inhibition. Trans Assn Am Phys 99:67—72, 1986 8. MCVERRY BA, FISHER C, HoPe A, HUEHNS ER: Production of

pseudodiabetic renal glomerular changes in mice after repeated injections of glucosylated proteins. Lancet i:738—740, 1980 9. NATHKE HE, SIESS EA, WIELAND OH: Glucosylated plasma protein injection does not produce glomerular basement membrane thickening. Horm Metab Res 16:557—558, 1984 10. ENGERMAN RL, KERN TS: Experimental galactosemia produces diabetic-like retinopathy. Diabetes 3 l(Suppl 2):26A, 1982

glomerulosclerosis, in Secondary Diabetes: The Spectrum of the Diabetic Syndromes, edited by PODOLSKY 5, VISWANATHAN M, New York, Raven Press, 1980, pp. 521—540 22. STEFFES MW, MAUER SM: Diabetic glomerulopathy in man and experimental animal models. Int Rev Exp Pathol 26: 147—175, 1984 23. BROWN DM, ANDRES GA, HOSTETTER TH, MAUER SM, PRICE R,

VENKATACHALAM MA: Kidney complications. Diabetes 31(Suppl l):71.-8l, 1982 24. BLOODWORTH JMB JR: Experimental diabetic glomerulosclerosis. II. The dog. Arch Pathol 79:113—125, 1965

25. MAUER SM, STEFFES MW, GOETZ FC, SUTHERLAND DER,

BROWN DM: Diabetic nephropathy. A perspective. Diabetes 32(Suppl 2):52—55, 1983

26. RASCH R: Prevention of diabetic glomerulopathy in streptozotocin diabetic rats by insulin treatment. Glomerular basement membrane thickness. Diabetologia 16:319—324, 1979 27. RASCH R: Prevention of diabetic glomerulopathy in streptozotocin diabetic rats by insulin treatment. The mesangial region. Diabetologia 17:243—248, 1979

28. RASCH R: Prevention of diabetic glomerulopathy in streptozotocin

diabetic rats by insulin treatment. Kidney size and glomerular volume. Diabetologia 16:125—128, 1979

29. FRANK RN, KEIRN Ri, KENNEDY A, FRANK KW: Galactoseinduced retinal capillary basement membrane thickening: Prevention by sorbinil. Invest Ophihalmol Vis Sci 24:1519—1524, 1983 30. R0BI5ON WG JR. KADOR PF, KIN05HITA JH: Retinal capillaries: Basement membrane thickening by galactose prevented with aldose reductase inhibitor. Science 221:1177—I 179, 1983

11. Kn'osrnm JH: Aldose reductase in the diabetic eye. Am J Oph- 31. K0JIMA K, MATSUBARA H, HARADA T, MIzUN0 K, SUZUKI M, thalmol 102:685—692, 1986

12. ENGERMAN RL, KERN TS: Experimental galactosemia produces diabetic-like retinopathy. Diabetes 33:97—100, 1984

HOTTA N, KAKUTA H, SAKAMOTO N: Effects of aldose reductase inhibitor on retinal microangiopathy in streptozotocin diabetic rats. Jpn J Ophthalmol 29:99—109, 1985