Aminoguanidine ameliorates changes in the IGF system in experimental diabetic nephropathy

Introduction Background. Formation of advanced glycation endproducts (AGEs) has been implicated in the developNephropathy occurs in up to 30% of diabetic patients ment of diabetic complications. As well as causing and is a significant cause of morbidity and mortality. changes in structural proteins, AGEs may also alter Hyperfiltration and microalbuminuria are hallmarks gene expression of growth factors in vitro. The insulinof early nephropathy, with the subsequent development like growth factor (IGF ) system, including IGF-I and of macroalbuminuria, hypertension, and progressive modulatory IGF binding proteins (IGFBPs), is dysrenal impairment [1]. Nephropathy is characterized regulated during the development of diabetic by renal enlargement and a number of ultrastructural nephropathy. abnormalities, including basement membrane thickenMethods. Quantitative in situ hybridization histocheming and mesangial expansion [1]. istry and immunohistochemistry were used to deterAlthough hyperglycaemia is a prerequisite for the mine the effects of aminoguanidine, an inhibitor of development of diabetic nephropathy, the mechanisms AGE formation, on gene expression of IGF-I and underlying its development are not completely underIGFBPs in kidneys of long-term (8 months duration) stood. Increased formation of advanced glycation endstreptozotocin-diabetic rats. products (AGEs), which result from the non-enzymatic Results. Diabetes was associated with increased renal reaction of glucose with biologically relevant molecules expression of IGFBP-1 mRNA (diabetes 824±236 vs such as structural proteins, has been implicated in the control 264±76 arbitrary units, P<0.01) and development of diabetic complications including decreased expression of mRNAs for IGF-I (diabetes nephropathy [2]. Indeed, treatment of diabetic rats 39±7 vs control 185±23 arbitrary units, P<0.001) with aminoguanidine, an inhibitor of AGE formation, and IGFBP-4 (diabetes 139±25 vs control 383±54 significantly reduces albuminuria and mesangial expanarbitrary units, P<0.001). Aminoguanidine treatment sion in streptozotocin-(STZ)-diabetic rats [3]. inhibited the effects of diabetes on renal expression AGE formation may directly result in the developof mRNA for IGF-I, IGFBP-1 and IGFBP-4. The ment of structural changes in tissues by chemical changes in IGF-I and IGFBP-1 mRNA levels were means, including formation of cross-links between reflected in altered peptide levels. In diabetic kidneys, structural proteins that may alter properties such as IGFBP-5 mRNA levels were slightly decreased to 75% susceptibility to breakdown. However, there is also in of control levels (P<0.01); aminoguanidine had no vitro evidence that AGEs may also stimulate the syneffect on IGFBP-5 mRNA levels. thesis of growth factors, including TGF-b and IGF-I Conclusions. These results suggest that amelioration of [4], which have been implicated in the development of changes in the renal IGF system by aminoguanidine diabetic complications [5–7]. may contribute to the renoprotective effects of the Considerable evidence links dysregulation of the latter, which have been previously shown to inhibit IGF system and the development of experimental structural and functional aspects of diabetic nephrodiabetic nephropathy. Kidney IGF-I levels are elevated pathy in the rat. during the first few days of STZ diabetes, immediately preceding the rapid phase of renal growth [8,9].

(5 min), following which sections were incubated with probes decreased albuminuria and structural changes comin humidified (50% formamide) chambers for 14-16 h at pared with 'IGF-I-replete' control diabetic rats [12,13].diabetic rats, renal IGF-I levels normalize or decrease 50% formamide at 55°C for 1.5 h.Sections were then washed below control levels (reviewed in [6 ]), indicating either and treated with 150 mg/ml RNAase A for 1 h at 37°C that the predominant role of the IGF system is in the followed by a wash in 2×SSC at 55°C for 45 min.After development of the early changes of diabetes or that final dehydration through graded ethanol, slides were air- more sustained changes in the other components of dried and autoradiographed with Kodak XAR ( Eastman the IGF system are responsible for continuing effects.Kodak, New York) film for 1-3 days at room temperature.
may potentiate or inhibit IGF actions [14].Renal expression of IGFBPs is altered in STZ-diabetes Quantitation of in-situ hybridization histochemistry [15][16][17][18][19][20].Major changes include increased IGFBP-1 [22] levels and decreased IGFBP-4 levels.The aim of this study was to investigate whether aminoguanidine, Autoradiographs were quantitated using the MCID (Imaging which inhibits the development of diabetic nephro-Research, Ontario, Canada) computerized image analysis pathy, also normalizes changes in expression of IGF-I system.Regions of interest were traced out, and optical and IGFBPs 1-6.
densities within those regions were quantified and calibrated by reference to 14C-radioactive standards (Amersham, UK ) which were simultaneously exposed with slides to X-ray film.

Subjects and methods
Results were corrected for non-specific labelling by subtracting the signal obtained using sense probes from that Animals obtained using antisense probes.
Male Sprague-Dawley rats weighing 200-250 g were made Immunohistochemistry diabetic by intravenous injection of streptozotocin (55 mg/kg in citrate buffer, pH 4.5).Control rats were injected with Standard techniques were employed using rabbit anti-mouse citrate buffer alone.Control and diabetic rats were random-IGFBP-1 and rabbit anti-human IGF-I polyclonal antisera ized to receive either aminoguanidine (AG, 1 g/l in drinking (GroPep, Adelaide, Australia), both of which cross-react water for the entire 8 months of the study) or no treatment.with rat antigens, at dilutions of 15200 and 15100 respect-Diabetic rats did not receive insulin therapy.Urine of diabetic ively.The Elite Vectastain ABC kit ( Vector Laboratories, rats was regularly tested for ketones ( Ketostix, Bayer, Burlingame, CA) and the peroxidase method of labelling Mulgrave, Australia) and these were not detected.After 8 were used according to the manufacturer's instructions.months, blood was taken from the tail vein for glucose Negative control slides were incubated with normal goat measurement and animals were killed by intravenous serum instead of primary antibody.Nembutal injection.Kidneys were removed, weighed, and fixed in 10% neutral buffered formalin.This protocol was approved by the Austin and Repatriation Medical Centre Statistics Animal Welfare Committee.
Quantitative in-situ hybridization histochemistry data were log-transformed to stabilize variance prior to analysis.

In-situ hybridization histochemistry
Results were analysed by analysis of variance followed by Fisher's protected least significant difference test for specific In-situ hybridization histochemistry was performed as previcomparisons between groups.Results are expressed as ously described [21].Briefly, 35S-labelled complementary mean±SE.Two-sided P values are shown.(antisense) or non-complementary (sense) RNA probes were synthesized for rat IGFBP-1-6 and IGF-I.The cDNAs for IGFBPs 1-6 were kindly provided by Dr S. Shimasaki Results ( Whittier Institute, La Jolla, CA), and for IGF-I by Dr C. T. Roberts Jr (NIH, Bethesda, MD).Probes were adjusted As expected, diabetic rats had higher plasma glucose to an average length of 150 bases by alkaline hydrolysis.The average specific activity of the RNA probes generated was levels and urinary volumes than control rats, but 3×109 c.p.m./mg RNA.aminoguanidine had no effect on severity of diabetes Kidney sections (4 mm) were digested with Pronase E as measured by either parameter ( Table 1).Control (Sigma, 125 mg/ml ), washed in 0.1 M sodium phosphate and diabetic rats treated with aminoguanidine weighed buffer, pH 7.2, ultrapure water, dehydrated in 70% ethanol, less than vehicle-treated rats (P=0.003,Table 1).and air-dried.The 35S-labelled RNA probes (5×105 However, diabetes resulted in a similar degree of weight c.p.m./25 ml hybridization buffer) were added to a hybridizaloss in vehicle-and aminoguanidine-treated rats tion buffer consisting of 300 mM NaCl, 10 mM Tris-HC1, (diabetes×treatment interaction, P=0.23).Absolute and relative kidney weights were significantly higher 1×Denhardt's solution, 50 mg/ml yeast RNA, 50% deionized in diabetic rats (P<0.0001,Table 1) and treatment formamide, and 10% (w/v) dextran sulphate.The hybridization buffer containing labelled probe was preheated to 80°C with aminoguanidine had no effect on either parameter.Aminoguanidine ameliorates diabetes-related changes in (results not shown), were decreased by 79% compared with untreated control rats (P<0.001, Figure 4).In renal IGF-I, IGFBP-1, and IGFBP-4 expression aminoguanidine-treated diabetic rats, levels of IGF-I As previously described [15][16][17][18][19][20], the major effects of mRNA were decreased by only 39% compared with diabetes on gene expression of IGF system components aminoguanidine-treated controls, a difference that did were increased IGFBP-1 mRNA levels and decreased not reach statistical significance (P<0.10, Figure 4).IGFBP-4 and IGF-I mRNA levels.
Levels of IGF-I mRNA were significantly higher in Overall, IGFBP-1 mRNA levels were increased by aminoguanidine-treated than untreated diabetic rats 212% in untreated diabetic rats (P<0.01, Figure 1a (P<0.01, Figure 4).and b), whereas this increase was abrogated by amino-Immunohistochemical analysis demonstrated IGF-I guanidine treatment (control+AG vs diabetes+AG, peptide in distal tubules and thick ascending limbs of P>0.20).The major sites of IGFBP-1 mRNA expresthe cortex and outer medulla (Figure 2e).Levels of sion were the inner stripe of the outer medulla, IGF-I peptide were decreased in diabetic rats reflecting expression in distal tubules and thick ( Figure 2f ), and aminoguanidine treatment restored ascending loops of Henle, and the cortex, representing peptide levels towards normal (Figure 2h).Changes in cortical collecting duct expression (results not shown).IGF-I mRNA and peptide levels were therefore IGFBP-1 mRNA levels were significantly increased by coordinate.206 and 219% in the inner stripe of the outer medulla and cortex of untreated diabetic rats respectively, whereas levels in aminoguanidine-treated diabetic rats Aminoguanidine has no effect on other IGFBPs were not significantly higher than in their non-diabetic IGFBP-5 mRNA was predominantly expressed in the controls ( Figure 1a,c and d).
inner medulla and inner stripe of the outer medulla Immunohistochemical analysis showed that with lower levels of expression in the cortex (results IGFBP-1 protein was most prominent in collecting not shown).Expression of IGFBP-5 mRNA was 25% ducts ( Figure 2a).IGFBP-1 was also found in distal lower in diabetic than control rats (diabetes 4069±259 tubules and thick ascending limbs of the cortex vs control 5459±578 arbitrary units, P<0.01) and and outer medulla.IGFBP-1 levels were markedly aminoguanidine treatment had no effect on levels in increased in all of these sites in diabetic rats diabetic or control animals.Diabetes and aminoguani-(Figure 2b).Aminoguanidine attenuated the diabetesdine treatment had no effect on expression of mRNA related increase in IGFBP-1 levels (Figure 2c).for IGFBPs-2, -3 and -6 (results not shown).Changes in IGFBP-1 mRNA and protein levels were therefore coordinate.
Steady-state cortical IGFBP-4 mRNA levels were

Discussion
decreased by 64% in untreated diabetic rats compared with untreated controls (P<0.001, Figure 3).Although levels of IGFBP-4 mRNA were decreased by 36% in Dysregulation of the IGF system has been implicated in the development of experimental diabetic nephro-aminoguanidine-treated diabetic rats compared with aminoguanidine-treated controls, this difference was pathy [6 ].Renal IGF-I levels are transiently elevated in the first 24-72 h after induction of STZ-diabetes not statistically significant (P<0.10, Figure 3).Levels of IGFBP-4 mRNA were significantly higher in amino-and a role for increased IGF-I activity has been postulated in the development of early diabetes-related guanidine-treated than untreated diabetic rats (P<0.001, Figure 3).Cortical IGFBP-4 mRNA pre-kidney growth [8,9].Long-term studies of diabetic growth hormone/IGF-I-deficient diabetic rats [13] and dominantly reflects proximal tubular expression as confirmed by light microscopic examination of emul-octreotide-treated rats [11], both of which have decreased urinary albumin excretion and glomerular sion-dipped slides (results not shown).
In vehicle-treated diabetic rats, levels of IGF-I volume, also suggest a role for the IGF system in the longer-term development of structural and functional mRNA in the inner stripe of the outer medulla, representing expression in thick ascending loops of Henle changes of diabetic nephropathy.
In the present study, long-term STZ-diabetes resulted in changes in renal gene expression of a number of components of the IGF system.Firstly, levels of IGF-I mRNA were decreased.Although some [23,24], but not other [25], authors have found transient increases in renal IGF-I mRNA levels soon after the induction of STZ-diabetes, the observations of the present study are consistent with studies demonstrating decreased IGF-I mRNA levels in kidneys from longterm diabetic rats [15,19,26 ].Second, expression of IGFBPs, which are important regulators of IGF actions [14], is altered early in STZ-diabetic kidneys with many changes persisting for up to 6 months, the longest duration previously studied [19,20].Cortical IGFBP-1 mRNA levels were increased in STZ-diabetic kidneys in the present study, consistent with previous observations [16,19,20,27].IGFBP-1 mRNA levels in the inner stripe of the outer medulla were also increased in diabetic rats in the present study as was observed in our previous study of short-term diabetic rats [20].These findings contrast with the markedly decreased levels of medullary IGFBP-1 mRNA observed by Landau et al. in both short-and long-term diabetic rats [19].These inconsistencies may relate to differences in rat strains or severity of diabetes.
The changes in IGF-I and IGFBP-1 mRNA levels were reflected at the protein level in the present study.Localization of proteins was more widespread than that of the mRNA, which is consistent with previous studies [28,29].IGF-I and IGFBP-1 found at sites other than those of mRNA expression may reflect paracrine transport of proteins synthesized in the kidney or proteins sourced from the circulation.
Cortical IGFBP-4 levels were decreased in diabetic rat kidney, which is also consistent with previous studies [19,20,27].Unfortunately, attempts to perform IGFBP-4 immunohistochemistry with two different antisera were unsuccessful in the present study.A small decrease in IGFBP-5 mRNA was also observed in diabetic kidneys that is similar in magnitude to the decrease in cortical IGFBP-5 mRNA levels previously reported [19].In that study, medullary IGFBP-5 levels were unchanged after 180 days of diabetes.Whereas aminoguanidine attenuated diabetes-related changes in expression of IGF-I, IGFBP-1, and IGFBP-4, expression of IGFBP-5 mRNA was not influenced.This suggests that components of the renal IGF system may respond differently to diabetes-induced AGE accumulation.
The diabetic rats in the present study were not treated with insulin and had body weights ~30% lower than control rats.Although it is possible that haemo-  Aminoguanidine is an inhibitor of AGE formation.IGF actions under different conditions [14], their role in the development of diabetic kidney disease has not Aminoguanidine treatment substantially retards the development of albuminuria [3,[30][31][32][33] and prevents been completely defined.However, IGFBP-4 is thought to be a purely inhibitory IGFBP [14], so that relatively mesangial expansion [3,31,33] in diabetic rats.Some [33,34] but not all [3,31] studies have shown that increased levels of this IGFBP are likely to inhibit IGF actions in the kidney.Aminoguanidine abrogated the aminoguanidine also inhibits glomerular basement membrane thickening induced by diabetes.In the pre-diabetes-induced increase in IGFBP-1 mRNA and protein levels.Cell-associated IGFBP-1, which may be sent study, aminoguanidine treatment significantly blunted the diabetes-related reduction in IGFBP-4 present on proximal tubular cells of diabetic rat kidneys [16 ], has been implicated in potentiation of IGF mRNA levels.Since IGFBPs may inhibit or potentiate actions [35].Inhibition of the elevation of levels in this IGFBP by aminoguanidine may therefore also act been increasing interest in the role of the tubulointerstitium in diabetic nephropathy, with declining renal to inhibit renal IGF actions.
The present study did not address the mechanism function more closely correlated to tubulointerstitial rather than glomerular changes [39,40].Renal tubules by which aminoguanidine affected the IGF system.It is unlikely that AGEs play a role in the changes in the are a major site of AGE accumulation and many of the receptors involved in mediating AGE-induced IGF system that occur soon after the induction of STZ-diabetes, since they would not have had time to tissue injury or in the clearance of AGEs are located in tubules [41,42].Renoprotective therapies have been accumulate.AGE formation may lead to the persistence of abnormal expression of IGF system compon-shown to ameliorate tubular as well as glomerular injury in experimental and human diabetic nephro-ents by signalling via specific AGE receptors [36 ].Aminoguanidine treatment would inhibit formation of pathy [43,44].
AGEs have been shown to increase IGF-I expression AGE ligands and thereby result in 'normalization' of IGF system expression following the initial perturba-in monocytes [36 ] and glomerular mesangial cells [4] via a receptor-specific mechanism.It may appear that tion due to induction of diabetes.An alternative, indirect mechanism for the effect of aminoguanidine is these findings contradict the present observations that an inhibitor of AGE formation ameliorates the that AGEs regulate synthesis of extracellular matrix components [37], which in turn may regulate expres-decrease in renal IGF-I levels in diabetic rats.However, these in-vitro experiments do not take the in-vivo milieu, sion of IGFBPs [38].
Many of the changes in the renal IGF system were whereby renal IGF-I mRNA expression is decreased in long-term diabetic rats, into account.most readily detected in tubules.Indeed, there has

60°C.
Sections were washed in 2×standard saline citrate Following the initial rise in kidney IGF-I levels in (2×SSC; 0.3 M NaCl, 0.33 M Na 3

Fig. 1 .
(a) IGFBP-1 mRNA expression in rat kidney.(A) Control, dynamic and metabolic changes may have contributed to the changes in renal IGF expression in the diabetic diabetes+aminoguanidine.(b) Quantitation of cortical IGFBP-1 rats, it should be noted that the rats were not ketotic mRNA expression in control (open bars) or diabetic (shaded bars) or severely catabolic as ketonuria was absent and renal rats treated with water or aminoguanidine (AG).(c) Quantitation of cortical IGFBP-1 mRNA expression control (open bars) or hypertrophy was observed.Further, since aminoguanidiabetic (shaded bars) rats treated with water or aminoguanidine dine ameliorated many of the diabetes-associated (AG).(d ) Quantitation of IGFBP-1 mRNA expression in the inner changes in the renal IGF system without affecting stripe of the outer medulla of control (open bars) or diabetic (shaded glycaemic control or changes in body weight, it is bars) rats treated with water or aminoguanidine (AG).Results, likely that these abnormalities are not merely related expressed as arbitrary units of optical density (OD), are shown as mean±SE (n=4-6).*P<0.02**P<0.01,control vs diabetes.tohaemodynamic and metabolic changes.
17. Ooi GT, Tseng LY-H, Tran MQ, Rechler MM.Insulin rapidly In conclusion, long-term treatment of diabetic rats decreases IGFBP-1 gene transcription in streptozotocin-diabetic with aminoguanidine, which has been previously rats.Mol Endocrinol 1992; 6: 2219-2228 shown to inhibit the development of structural and 18. Flyvbjerg A, Kessler U, Dorka B, Funk B, Orskov H, Kiess W. functional aspects of experimental diabetic nephro-Transient increase in renal insulin-like growth factor binding proteins during initial kidney hypertrophy in experimental dia-pathy [3,30-33], ameliorated diabetes-induced changes betes in rats.Diabetologia 1992; 35: 589-593 in expression of IGF-I, IGFBP-1, and IGFBP-4.Since 19.Landau D, Chin E, Bondy C et al.Expression of insulin-like dysregulation of the IGF system has been implicated growth factor binding proteins in the rat kidney: effects of longin the development of diabetic nephropathy, these term diabetes.Endocrinology 1995; 136: 1835-1842 findings suggest a possible mechanism that may con-20.Price GJ, Berka JL, Werther GA, Bach LA.Cell-specific regula- tion of mRNAs for IGF-I and IGF-binding proteins 4 and 5 in tribute to the protective effects of aminoguanidine.streptozotocin-diabetic rat kidney.J Mol Endocrinol 1997; 18: 5-14

Table 1 .
Characteristics of rat groups