Glycosaminoglycans

Because diabetic nephropathy occurs in 20 to 40% of patients with diabetes mellitus, it has become one of the most important causes of end-stage renal disease. The latest incidence figures vary from 16% of patients starting with dialysis in The Netherlands (1) to 42% in the United States (2). Diabetic nephropathy has a large impact in terms of associated morbidity and mortality for the individual patient and in terms of costs for health care (3). The duration and efficacy of treating hyperglycemia (4), as well as BP regulation (5, 6) and genetic factors (7, 8, 9, 10), are probably all important in the pathogenesis of diabetic nephropathy in insulin-dependent diabetes mellitus (IDDM) and non-insulin-dependent diabetes mellitus (NIDDM). Several retrospective and prospective studies have demonstrated that micro- and macroalbuminuria predict cardiovascular morbidity and mortality in diabetes mellitus (reviewed in reference (11). Although Rossing et al. recently reported that the prognosis for patients with overt diabetic nephropathy has improved, 15% of IDDM patients with normoalbuminuria, 25% with microalbuminuria, and 44% with overt nephropathy died during their 10-yr follow-up study (12). How albuminuria and proteinuria are linked with macroangiopathy is poorly understood, and many hypotheses have been formulated and tested. The Steno hypothesis, first advanced by Deckert et al. in 1988, held that a genetic defect in the regulation of heparan sulfate (HS) production by endothelial, myomedial, and mesangial cells determines the susceptibility of diabetic patients to develop proteinuria and angiopathy with its associated cardiovascular risk (13). The sulfation pattern of the glycosaminoglycan (GAG) side chains of HS proteoglycans (HSPG) plays a pivotal role in this hypothesis. The central idea is that diabetic patients susceptible to nephropathy and macroangiopathy have a genetic trait leading to a lower activity of the enzymes responsible for GAG sulfation under high glucose conditions. The resulting undersulfated GAG chains would then play a crucial role in the pathogenesis of proteinuria (due to the loss of anionic charges in the glomerular basement membrane [GBM]) and its morphologic substrate, diabetic nephropathy, as well as in the pathogenesis of diabetic micro- and macroangiopathy. Although genes have been cloned for N-deacetylase/N-sulfotransferase, 3-O-sulfotransferase, and 6-O-sulfotransferase enzymes (reviewed in reference (14), it is still unclear whether different allotypes of these enzymes with a different susceptibility to high glucose concentrations really exist. Although the Steno hypothesis has not been formally proven, it has stimulated in vitro studies and therapeutic trials in experimental animal models and patients. Pathogenesis of Diabetic Nephropathy and the Possible Role of Abnormal GAG Metabolism Possible GAG metabolism abnormalities in diabetic nephropathy were originally investigated for the following reasons: (1) albuminuria and proteinuria appear in diabetic nephropathy; (2) this phenomenon suggests abnormal GBM permeability; and (3) GAG, in particular HS, were thought to be important determinants of GBM permeability. Several groups reported a decreased 35S sulfate incorporation in the GBM of diabetic glomeruli (15, 16). In mice and rats with diabetes, reduced synthesis of glomerular proteoglycans and basement membrane HSPG was found (17, 18, 19). However, the findings of the numerous sulfate incorporation experiments are not without controversy, and a marked increase in radiolabeled sulfate incorporation in proteoglycans in diabetic tissues has also been reported (reviewed in reference (20). Studies using biochemical techniques to measure GAG content of kidneys obtained at autopsy demonstrated that GBM of patients with diabetic nephropathy contained less GAG than kidneys of nondiabetic control subjects (21, 22). Similar changes in HS content in the intima of the aortas of patients with diabetes mellitus have been observed (23), suggesting that the abnormalities in HS metabolism are not necessarily restricted to the kidney. This would explain the association between cardiovascular mortality and diabetic nephropathy. Biochemical techniques, as well as immunohistology and histochemical staining procedures combined with electron microscopy morphometric studies have been used to examine glomerular HSPG and GAG content in diabetic nephropathy. With the last technique, Vernier et al. (24) described a reverse correlation between GBM HSPG expression and mesangial expansion in diabetic nephropathy. In the past decade, new monoclonal antibodies against GAG chains and core proteins of various HSPG have become available (25). JM-403, a very interesting monoclonal antibody that reacts with HS-GAG chains, mainly stains the GBM in normal kidneys, but largely fails to stain tubular basement membranes. The epitope recognized by JM-403 contains one or more N-unsubstituted glucosamine and D-glucuronic acid units, and is located in a region of the HS chain composed of mixed N-sulfated and N-acetylated disaccharide units (26). Using this antibody, we could show that a decreased GBM staining intensity (Figure 1) correlated with proteinuria, expressed as a function of creatinine clearance in patients with diabetic nephropathy (27). We also could show that staining of skin basement membrane was significantly reduced in patients with diabetic nephropathy compared to patients with long-standing diabetes without nephropathy (28). It should be noted that the exact meaning of a decreased staining by this antibody under pathologic conditions in relation to total GAG content or sulfation is not yet known. The first major basal membrane HSPG to be identified was perlecan (20); however, immunohistologic staining with antibodies against perlecan HSPG core proteins did not reveal any significant changes in diabetic nephropathy (25, 27). In 1999, three HSPG core proteins have been identified in the GBM: perlecan, agrin (29), and collagen XVIII (30). Interestingly, domain-specific monoclonal antibodies revealed a differential expression of agrin in glomerular and tubular basement membranes, suggesting alternative splicing and/or posttranslational processing of the molecule within the kidney (31). Agrin is present only in the kidney and the nervous system, and studies on renal agrin metabolism in diabetes are currently in progress.Figure 1.: (A) Control section stained with a monoclonal antibody directed against heparan sulfate (HS) (JM-403). A linear staining pattern of high intensity is observed in the glomerular basement membrane. Magnification, X250. (B) Kidney section of a patient with diabetic nephropathy showing mesangial expansion without nodular lesions stained with the same anti-HS antibody. Compared with Panel A, the intensity of staining of HS chains is clearly diminished. From reference 27. Reproduced with permission from Springer-Verlag, Heidelberg, Germany.The effects of the diabetic milieu on glomerular HSPG synthesis have been studied in vitro by different groups, using glomerular cell cultures (reviewed in reference (32). When human glomerular visceral epithelial and mesangial cells were cultured under normal (5 mM) and high (25 mM) glucose conditions, and then stained with the same monoclonal antibody (JM-403) that had been used on the tissue sections of diabetic kidneys, a decreased extracellular matrix staining after culture in 25 mM glucose was observed (33). The rise of metabolic labeling also disclosed an altered proteoglycan production under high glucose conditions, with predominantly a decrease in HS, compared with dermatan or chondroitin sulfate proteoglycan. N-sulfation analysis of HSPG produced under high-glucose conditions revealed less di- and tetrasaccharides, compared with larger oligosaccharides, indicating an altered sulfation pattern. Renal function deterioration and proteinuria in diabetic patients are correlated with mesangial expansion as the main morphologic parameter (34). Because angiotensin-converting enzyme (ACE) inhibitors are known to slow the progression of diabetic nephropathy (5) and because we were interested in the role of HSPG in diabetic nephropathy, we studied the effect of angiotensin II (AngII) on mesangial HSPG production. Because AngII is also known to stimulate extracellular matrix production through transforming growth factor-β (TGF-β) production in rat mesangial cells (35), we also studied the role of TGF-β in this regard (36). Metabolic labeling studies revealed that AngII induced a decrease in HSPG synthesis, with a decrease in N-sulfation of the GAG side chains. Enzymelinked immunosorbent assay measurements using JM-403 confirmed that AngII decreased HS production. AngII increased TGF-β production in a dose-dependent manner. Specific mRNA for perlecan HSPG decreased, while mRNA for TGF-β increased after incubation with AngII. Blockade of the subtype 1 AngII receptor (ATR1) reversed the effects of AngII on both HSPG and TGF-β production. Coincubation of the mesangial cells with neutralizing antibodies against TGF-β did not prevent the AngII-induced reduction of HS. These results indicate that the decrease in HS synthesis induced by AngII is not mediated by an increase in TGF-β, but, on the contrary, the increase in TGF-β partially counteracts the inhibition of HS production by AngII. Because defects in HS-GAG synthesis are so striking in in vitro models of diabetic nephropathy (induced either by high glucose concentrations or AngII), and because decreased renal and extrarenal JM-403 staining seems to be a rather specific marker for nephropathy in patients with diabetes, treatment of diabetic nephropathy with HS-GAG-like substances can be viewed as an experiment to test the Steno hypothesis. Heparin and GAG Renoprotection in Experimental Diabetic Nephropathy Indeed, numerous reports showed that heparin and more generally GAG prevent and cure experimental diabetic nephropathy (37, 38, 39, 40, 41). However, it soon became clear that the activity of these drugs could not be explained, according to the Steno hypothesis, only in terms of recovery of the diabetes-induced abnormalities in HSPG metabolism, and restoration of anionic-HS charges in glomerular and other basement membranes. Several hypotheses to explain the renoprotective effect of heparin and GAG in experimental diabetes were formulated: downregulation of proteases, modulation of extracellular mesangial matrix synthesis, and restoration of GBM anionic charges (reviewed in references (42) and (43). Most in vitro and in vivo data support the idea that the degradation of extracellular matrix is impaired in diabetic nephropathy (44) and that heparin and GAG may correct the balance between collagen synthesis and degradation by acting on synthesis rather than on degradation (45). Heparin and GAG improve glomerular permselectivity to proteins, as deduced by evaluating the fractional clearances of neutral and anionic dextrans (37). This has been indirectly confirmed by a number of investigators reporting a reduction in the albumin excretion rate after heparin treatment in diabetic animals. This effect is probably related to the activity of heparin on extracellular matrix and GBM protein synthesis, rather than adhering to the GBM and thereby correcting the charge deficiency. The activity of heparin on the sulfation and synthesis of proteoglycans (37, 46) and on the collagen IV/perlecan mRNA ratio might be relevant to maintaining the authentic architecture of the membrane with normal permeability characteristics (37). However, although a recent study confirmed some of the glomerular morphologic effects, it reported increased albuminuria after heparin therapy (47). We suggest that this contrasting result might be due to the high heparin dosage and the type of commercial heparin used in this study (48). Indeed, at such high dosages, heparin appears in the glomerular ultrafiltrate; it then may interfere with the charge-dependent proximal tubule reabsorption of albumin due to its high degree of sulfation, thereby increasing albuminuria. Furthermore, it should be emphasized that when referring to heparin, it is generally overlooked that heparin is a heterogeneous group of polysaccharides, with varying degrees of sulfation, molecular weights, and biologic activities. Commercial heparins are similar regarding their anticoagulant activity, but not with respect to many other activities. For example, let us consider anti-proliferative activity, which is probably heparin's best-investigated non-anticoagulation-related activity. Although heparins from different suppliers have the same anticoagulant activity, they show a broad range of anti-mitogenicity, and some are even mitogenic (49). Furthermore, depending on its dosage, the same heparin preparation may be both anti-proliferative and proliferative (49). Similarly, in studies on the effect of heparin and GAG in experimental diabetic nephropathy, despite differences in drug formulation and dosage, some effects were constant, especially the effects on the ultrastructure of the glomerulus, whereas others, such as those on albuminuria, were contradictory. Most likely these differences were due to the quality and the quantity of heparins used in the different protocols. It is also possible that the renoprotective effect of heparin and GAG in diabetic nephropathy depends on the modulation of the overactivated TGF-β cascade. Although TGF-β was shown to increase HSPG synthesis (36), a potentially favorable effect according to the Steno hypothesis, as a whole it has a more prominent and pivotal role in the pathogenesis of glomerulosclerosis. Thus, TGF-β inhibition might have a useful impact in the evolution of diabetic nephropathy. Indeed, GAG therapy in long-term diabetic rats prevents characteristic manifestations of diabetic nephropathy that are possibly related to TGF-β activity, mesangial matrix expansion, and deposition of periodic acid-Schiff-positive material, collagen III, and α1(IV) collagen (39). Heparin/GAG can dissociate TGF-β from α2-macroglobulin in serum, a protein possibly in the clearance of TGF-β should increase the of TGF-β, mesangial matrix production and diabetic nephropathy rather than it is found in a number of However, in can also have at in the TGF-β cascade. Heparin/GAG may interfere with or increased We have data indicating that GAG not interfere with receptor or with or to and by to TGF-β receptor and/or or by with with type II receptor the contrary, GAG on increased of In cultured mesangial we found that high and similar to treatment with Because the effects of high glucose concentrations on mesangial matrix production have been to protein it is that renoprotective GAG heparins and dermatan also the of mRNA obtained with These data that GAG TGF-β function by of mRNA and against inhibition of TGF-β by tissue and to TGF-β The hypothesis that can expression was in an in vivo study by evaluating mRNA expression with in Indeed, treatment the increase in expression in glomerular and tubular these findings support the that are in the inhibition of most as experiments by stimulated activity without basal activity. Interestingly, in mesangial the of renoprotective GAG has effect on basal mRNA This the that treatment has effect on mRNA in nondiabetic animals. the data indicate that GAG treatment prevents of mRNA in mesangial cells induced by different but not basal In the of all activity may be because of the risk of and cell The that GAG its basal expression seems to against such Indeed, in vivo long-term to studies showed that GAG treatment prevents diabetes-induced of with of or (37, 38, 41). Heparin and GAG Renoprotection in Diabetic Nephropathy experimental findings and the that increased expression in with diabetic nephropathy suggest that the results obtained in animal studies may also be relevant to human disease. reports have described favorable results of GAG treatment on proteinuria in diabetic nephropathy with a molecular heparin reduced albuminuria in both micro- and IDDM patients a of GAG mainly of HS, also could lower proteinuria in a study in IDDM patients with albumin excretion A of was for the effect to become with in BP and creatinine a formulation composed of the GAG heparin and dermatan that were in diabetic nephropathy in the experimental was reported to albuminuria in IDDM and patients and this effect after its Although treatment in IDDM seems to be in microalbuminuria, this effect is observed in only to of patients. It is known that in patients is not by the diabetic nephropathy because nondiabetic changes were reported in of the patients Because the effects of heparin the various experimental diabetic nephropathy is more to heparin than we suggest that in some patients and other not albuminuria due to a dosage that is to the renal The could also be related to differences in the type of the In this it is interesting to that which results in patients has an molecular of for and has an to activity ratio of Thus, in patients in the study by et al. less molecular heparin than in the study by et al. Furthermore, differences in molecular and and activity may be of different renoprotective of the heparin formulation Although the studies were to whether GAG treatment in diabetic patients is of diabetic nephropathy, of one of its of the studies with in diabetic patients with of the of GAG A major of in the treatment of diabetic patients with GAG and heparin has been the risk of in subjects with due to the of new Although the used were only to we recently demonstrated that therapy with the GAG was more we observed a reduction in This was because it is known that a reduction in such abnormalities is not likely to within a How can these findings be growth plays a pivotal role in diabetic both the proliferative and Interestingly, under conditions, some are and by extracellular in the sulfation of GAG in diabetes could to an altered of and GAG can reverse this phenomenon (37). Thus, the effect of on might well be related to the modulation of in the may also a in long-term treatment with The of with heparin and GAG is Indeed, it depends on both the of the molecule and the of The is very anticoagulant activity is largely on the degree and pattern of GAG sulfation, so that of GAG might be for the other anticoagulant activity also depends on the type of molecular heparin is whereas chondroitin and dermatan sulfate are not a of in long-term experiments we did not any in either or dermatan (37, The of the is that to therapeutic heparin be in this its concentrations are high to the cascade. When heparin is through other its are and anticoagulant be to the of these drugs as well as increase patient it would be relevant to study the and of heparin and GAG through different A reports have demonstrated that heparin and dermatan sulfate have of heparin to et al. found as large as with activity in the Furthermore, heparin was shown to be at the in to prevent in but with significant anticoagulant effect due to its by the sulfate is also in at the with a to These the of of GAG in terms of the anticoagulant and activity, but studies in diabetic patients with show that after GAG their effects not only on but, also on albuminuria. is reported as a possible effect of long-term heparin but studies have this Interestingly, however, and to the it was reported that a of heparin therapy in rats with renal did not metabolism The risk of antibody is very with GAG therapy because molecular GAG are the of GAG a lower risk of antibody compared with antibodies or is a of heparin therapy that is associated with due to the of between heparin, and Although of patients heparin may develop this its is significantly lower in patients with molecular heparin and GAG probably because of the lower for these have been used to heparin to and The of other GAG, such as chondroitin or dermatan an interesting because is less likely to however, is with these in patients. and Heparin/GAG inhibitors decrease progression of diabetic nephropathy and proteinuria by that only in can be to control of and and might be related to an increased activity of the angiotensin synthesis of and the effects of the Indeed, AngII may as a growth and a It and/or of glomerular and tubular the synthesis of collagen and the synthesis of perlecan, and extracellular matrix Although and experimental findings support the of an between diabetes and this occurs is A number of that may be rather than as in the of high glucose However, the that AngII glucose and of the glucose in a number of different cells the that a increased activity of the at the kidney for due to a particular genetic might to a glucose in diabetic patients than in diabetic subjects with normal renal activity The of such a phenomenon is because a number of abnormalities to to diabetic nephropathy extracellular matrix protein synthesis, TGF-β are related to the rather than to the extracellular glucose Furthermore, between the effects of and high glucose might also exist. effects similar or in cultured renal in renal cells is of diabetes and high glucose concentrations and is important in the high expression of collagen and TGF-β in renal cells AngII in mesangial and tubular cells through as well It is possible that hyperglycemia and AngII might effects on and Indeed, the effects of these conditions on the of proximal tubular cells was demonstrated In vitro studies on cultured mesangial and proximal tubular cells have shown that AngII and mitogenic of those induced by growth factors TGF-β and growth AngII is a of TGF-β, and some of its reported effects are mediated by expression of this growth TGF-β by AngII is also a phenomenon similar to by The of available data has to the that TGF-β might the between glomerular a and and This can be to diabetic nephropathy; TGF-β could be the between and with a pivotal role between AngII and high glucose However, it is whether high glucose and AngII the same the same or different in TGF-β although it was demonstrated in glomerular epithelial cells that at some effects are suggesting different is between the biologic effects of AngII and TGF-β on mesangial cells in as clearly shown by differential effects on perlecan production (36). The of between high and TGF-β, is but its to is not a of it may the of for the treatment of diabetic nephropathy. clearly shown by the to this on the different by high and TGF-β in diabetes mellitus most likely drugs that on different the for for diabetic nephropathy to to metabolic control and modulation seems to be In this it seems interesting that on different than and as a the of inhibitors and GAG could the TGF-β in the thereby increasing the of glomerulosclerosis. on the regarding the therapeutic of heparin in in the of Kidney after more than of and The at that was that were of in the treatment of renal disease. Indeed, after the first experiments with heparin, the in the was on for the treatment of a very and glomerular it to heparin with The Steno hypothesis a crucial role for in the pathogenesis of diabetic nephropathy. We may that are some specific in the GAG chains of heparin which are useful for and not necessarily have a with the anticoagulant activity of We that experimental and human data on the effects of GAG treatment on and proteinuria in diabetic nephropathy, the past have disclosed a new therapeutic to improve the treatment of a patient with a The of less with the favorable that these may have on and other could this of treatment even more studies be to and the and most of all whether in not only proteinuria, but also renal function and morphologic abnormalities can be and of the of and of for the in the of the