Proteinuria-Lowering Effect of Heparin Therapy in Diabetic Nephropathy without Affecting the -Angiotensin- Aldosterone System

Urs Benck,* Sarah Haeckel,* John H. Clorius,† and Fokko J. van der Woude* *V. Department of Medicine, University Hospital Mannheim, Mannheim, and †Department of Nuclear Medicine, German Cancer Research Center Heidelberg, Heidelberg, Germany

Angiotensin-converting enzyme inhibitors and angiotensin II (AngII) type 1 receptor blockers lower proteinuria and preserve renal function in diabetic nephropathy (DN). The antiproteinuric effects are greater than their blood pressure reduction, involving the sieving properties of the glomerular filter. In DN, glomerular staining for heparan sulfate proteoglycans is decreased. AngII inhibits heparan sulfate synthesis. Also, heparins modulate AngII signaling in glomerular cells, inhibiting aldosterone synthesis and lowering proteinuria in DN. Is the antiproteinuric effect of heparins due to its interference with the renin-angiotensin-aldosterone system? Ten volunteers each with DN and glomerulonephritis and control subjects were examined before and after low-dosage enoxaparin. Renal hemodynamics were determined with 99mTc-DTPA and 131I- hippurate clearance. Glomerular filtration rate (GFR), effective renal plasma flow, mean arterial pressure, and rate were measured at baseline and during AngII infusion before and after enoxaparin while on normal salt and salt restriction. Enoxaparin did not lower aldosterone levels. GFR remained stable in all groups. AngII caused a significant decrease in effective renal plasma flow, whereas mean arterial pressure and heart rate increased significantly. Enoxaparin did not influence the AngII-induced changes of renal hemodynamics during normal salt intake or salt restriction. All groups showed identical responses to AngII before and after enoxaparin. In patients with diabetes, enoxaparin caused a significant decrease in proteinuria. It is concluded that the antiproteinuric effect of heparins in DN cannot be explained via interaction with the renin-angiotensin-aldosterone system. The absence of hemodynamic changes combined with reduced proteinuria point to intrinsic alterations in the glomerular filter. The effects were seen only in DN, not in glomerulonephritis. Clin J Am Soc Nephrol 2: 58–67, 2007. doi: 10.2215/CJN.02400706

roteinuria is a hallmark of diabetic nephropathy (DN) reactive oxygen species, such as superoxide, and, via the gen- (1). It strongly affects renal prognosis (2) and is the eration of hydrogen peroxide, hydroxyl radicals (11). P most powerful predictor of associated cardiovascular Besides these potentially harmful hemodynamic and nonhe- disease (3). The renin-angiotensin-aldosterone system (RAAS) modynamic effects, AngII attenuates both production and sul- is a key mediator of renal damage in DN. Interference with the fation of heparan sulfate proteoglycans (HSPG) in cultured RAAS both by angiotensin-converting enzyme (ACE) inhibitors human mesangial cells (12). AngII also reduces HSPG expres- (4) and angiotensin II (AngII) type 1 receptor (AT1R) blockers sion in human podocytes, as could be demonstrated by a recent (5,6) decreases proteinuria and retards progression to ESRD in study from our group (13). In line with these in vitro findings patients with DN that is caused by type 1 and 2 diabetes. are in vivo data showing that the ACE inhibitor enalapril im- AngII increases glomerular pressure predominantly proves albuminuria by preventing glomerular loss of HS in at the efferent , which is believed to be pivotal in the diabetic rats (14). Therefore, the antiproteinuric effect of ACE progression of renal damage (7). AngII stimulates extracellular inhibition in DN may be explained at least partially by struc- matrix protein synthesis through potent induction of TGF-␤,a tural effects on the glomerular basement membrane (GBM); i.e., suspected key promoter of glomerulosclerosis (8). The nonhe- qualitative and quantitative preservation of HSPG improving modynamic effects of AngII also include the induction of proin- glomerular permselectivity. flammatory cytokines (9,10). Recent studies suggested that Not only RAAS-intervention strategies but also heparins may AngII can induce oxidative stress, resulting in the generation of influence DN favorably. Heparins prevent and cure experimen- tal DN in streptozotocin-induced diabetic rats (15,16). Also in humans, several phase 2 studies showed that heparins lower

Received July 12, 2006. Accepted October 26, 2006. proteinuria in both type 1 (17) and type 2 diabetes (18). Several hypotheses to explain the renoprotective effects of heparins Published online ahead of print. Publication date available at www.cjasn.org. have been discussed: Downregulation of proteases and TGF-␤ Address correspondence to: Dr. Urs Benck, V. Department of Medicine, Univer- cascade, modulation of mesangial matrix synthesis and resto- sity Hospital Mannheim, Theodor-Kutzer-Ufer 1-3, D-68167 Mannheim, Ger- many. Phone: ϩ49-170-945-1428; Fax: ϩ49-621-383-73-3525; E-mail: urs.benck@ ration of GBM anionic charges (reviewed in reference [19]). med5.ma.uni-heidelberg.de Relatively little is known about a putative interaction be-

Copyright © 2007 by the American Society of Nephrology ISSN: 1555-9041/201–0058 Clin J Am Soc Nephrol 2: 58–67, 2007 Proteinuria-Lowering Effect of Heparin on RAAS 59

tween heparins and the RAAS. Heparins are known to inhibit Protocol aldosterone production (20). In 1990, Zaragoza et al. (21) dem- The protocol was approved by the local medical ethics committee. onstrated that heparins suppress AngII-stimulated intracellular We performed a prospective, controlled, nonrandomized trial with the calcium mobilization in vascular smooth muscle cells. Another three study populations and a total study duration of 11 wk. The trial in vitro study from our group showed that heparins modulate is illustrated in Figure 1. In the treatment period, we used enoxaparin ϩ (Clexane; Aventis Pharma, Berlin, Germany) as low molecular weight Ca2 -dependent AngII signaling in human mesangial cells (22). heparin (LMWH) in a dosage of 4000 IU/d subcutaneously. Blood Because heparins are renoprotective in DN and the RAAS is samples were drawn 5 and 10 d after beginning of the treatment period a key mediator in its pathogenesis, we wanted to clarify to exclude LMWH-induced hyperkalemia and low platelet count and to whether there is an interaction between heparins and the RAAS check compliance to therapy by determination of anti–factor Xa activ- in vivo. Therefore, we studied the influence of heparin therapy ity. Normal salt diet was defined as liberal sodium intake of 200 on aldosterone levels, on the acute AngII-induced systemic and mmol/d. Salt restriction was defined as low salt intake of 50 mmol/d. renal hemodynamic effects, and on proteinuria in patients with All participants were instructed by an experienced dietician. Compli- DN and nondiabetic renal disease. ance to the diet was checked twice 5 and 3 d before each study day by 24-h urine collections to ensure inclusion only when dietary compliance Materials and Methods was good. Patients On study days, participants were advised to drink approximately 700 A total of 30 men and women were recruited for the study. Patients ml of liquids before and 400 ml during the examination. At 8 a.m., they with hypertensive BP values Ͼ160/90 mmHg at rest, instable coronary emptied their bladders and then rested in a supine position in a quiet heart disease, or diabetes with poor glycemic control (glycosylated room. After completion of the baseline clearance period of 50 min, hemoglobin Ͼ 8.5%) were excluded from the study. Pregnancy and age AngII (Clinalfa AG, Switzerland) was administered intravenously in a ␮ Ͻ18 or Ͼ80 yr were additional exclusion criteria. The participants were dosage of 8 g/kg per min for 60 min. This dosage is known to cause studied in three groups and admitted sequentially. a moderate systemic and renal hemodynamic pressure response, as Group 1. Group 1 consisted of 10 patients with diabetes (three of described previously (23). This was followed by a 20-min recovery 10 type 1, seven of 10 type 2) and DN. Diagnosis of DN was based on period. Mean arterial pressure (MAP) and heart rate (HR) were re- the presence of macroalbuminuria or overt proteinuria (confirmed by corded during examination by an automatic recording device (Di- three consecutive 24-h urine collections) together with nonproliferative namap; Critikon, Tampa, FL) at 5-min intervals. diabetic retinopathy. They ranged in age from 43 to 61 yr (55.2 Ϯ 5.55). The duration of diabetes ranged from 8 to 21 yr (15.1 Ϯ 4.65). Age of Measurements of Renal Hemodynamics onset of diabetes ranged from 19 to 52 yr (38.5 Ϯ 9.40). Body mass index Using a single-compartment dual-tracer infusion clearance and the (BMI) ranged from 21 to 41 kg/m2 (30.1 Ϯ 5.61). They had acceptable tracers 99mTc-diethylentriamine penta-acetic acid (DTPA) and 131I-hip- metabolic control by the inclusion criterion of glycosylated hemoglobin purate, we measured both GFR and effective renal plasma flow (ERPF) Ͻ8.5% (range 5.30 to 8.50; 7.34 Ϯ 1.07). Serum creatinine ranged from simultaneously in the supine position using the method of Pedersen et 0.81 to 2.78 mg/dl (1.42 Ϯ 0.68), reflecting renal function that ranged al. (24). The clearance examination of participants began 40 min after from normal to moderately impaired. Arterial hypertension had been simultaneous intravenous injection of 4.4 MBq 99mTc-DTPA and 4.4 diagnosed in all 10 patients previously. All of them were on long-term MBq 131I-hippurate. ACE inhibitor treatment for at least 6 mo, except for one patient who A superficial of the right arm was used for a continuous infusion was on stable ACE inhibition therapy for exactly 8 wk before the study. clearance. Both tracers were monitored by two scintillation probes Mean systolic office cuff BP was 135 Ϯ 15 mmHg. Mean diastolic office placed over the right and the left shoulder of each individual. This site cuff BP was 77 Ϯ 6.80 mmHg. None had instable cardiovascular dis- is selected because of the large vessels in the probe’s field of view. Each ease. detector monitored one of the two radioisotopes by means of energy Group 2. Group 2 consisted of 10 patients without diabetes, seven discrimination. The signals that were monitored by the probes acti- with biopsy-confirmed IgA glomerulonephritis (IgA-GN) and three vated an infusion pump system via feedback control. Two pumps were with membranous GN. They ranged in age from 28 to 71 yr (45.6 Ϯ used; one contained 99mTc-DTPA, and the other contained 131I-hippu- 11.7). BMI ranged from 19 to 39 kg/m2 (28.1 Ϯ 6.11). Serum creatinine ranged from 0.73 to 2.39 mg/dl (1.33 Ϯ 0.55), reflecting renal function that ranged from normal to moderately impaired. There were six of 10 in whom arterial hypertension had been diagnosed previously. All of them were on stable ACE inhibitor treatment for at least 8 wk. Mean systolic office cuff BP was 126 Ϯ 16.8 mmHg. Mean diastolic office cuff BP was 77 Ϯ 7.29 mmHg. None of them had instable cardiovascular disease. Group 3. Group 3 consisted of 10 healthy control subjects. They ranged in age from 27 to 66 yr (49.8 Ϯ 10.3). BMI ranged from 20 to 34 kg/m2 (25.5 Ϯ 3.75). Serum creatinine ranged from 0.66 to 1.07 mg/dl Figure 1. Study protocol. All participants were studied on four (0.89 Ϯ 0.16), reflecting normal renal function. Arterial hypertension occasions: Twice before and twice after an 8-wk treatment had been diagnosed previously in three of 10 patients; these individuals period on low-dosage, low molecular weight heparin (LMWH) were on stable ␤ blocker treatment for at least 8 wk, and none of them and on normal salt diet and on salt restriction, respectively. was on ACE inhibitor treatment. Mean systolic office cuff BP was 131 Ϯ Participants were studied after week 1 (A; normal salt, before 5.59 mmHg. Mean diastolic office cuff BP was 82 Ϯ 6.24 mmHg. All heparin), after week 2 (B; low salt, before heparin), after week participants gave written informed consent for the study before enroll- 10 (C; normal salt, after heparin), and after week 11 (D; low salt, ment. after heparin). Scr, screening. 60 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 2: 58–67, 2007 rate. A separate step motor drove the pumps, whereby steady-state analyzed with one-way ANOVA and Tukey multiple comparisons. For conditions were reached. The data from the first 10 min of the clearance nonparametric data, Wilcoxon test for paired differences was used. examination were discarded because this time was needed to equili- Statistical significance was defined as P Ͻ 0.05. Power analysis that brate the feedback control system. After completion of the 50-min compared ERPF before and after AngII infusion was performed with baseline clearance period, 10 ml of blood was drawn from the cubital ␣ ϭ 0.05 and 1 Ϫ ␤ ϭ 90% to define the sample size of each group to vein of the contralateral arm to obtain a plasma sample. A probe from detect a 20% difference in the AngII-induced change of ERPF. each pump that contained the respective isotope served as standard. A microcomputer registered the motor step rates, documented the serum activity level of each isotope in the probe’s field of view, and calculated Results Baseline characteristics of the participants are listed in Table the clearance after the activity of the standard and the serum sample had been registered using the equation: 1. According to power analysis, a sample size of 10 patients in each group was required to meet the criteria mentioned previ- N ϫ A CI ϭ st ously. Apl where Cl is the clearance (ml/min), N is the number of motor Hormonal Parameters Aldosterone levels on baseline were not significantly differ- steps per minute, Ast is the activity pumped per motor step ent between patients with diabetes, patients with GN, and (Bq), and Apl is specific activity of plasma (Bq/ml). The clearance was calculated for 2-min time intervals. These healthy control subjects. AngII induced a significant rise in separate GFR and ERPF values then were used to compute a aldosterone levels on every occasion except in patients who had mean clearance for any appropriate time interval studied, such diabetes and were on low salt and heparin (measurement D). as baseline (50 min), AngII infusion (60 min), and recovery (20 As expected, aldosterone levels were significantly higher on min). low salt (measurements B/D) compared with normal salt in- Filtration fraction was calculated as ratio of GFR and ERPF. take (measurements A/C) on every occasion in all groups. This Renal vascular resistance was calculated as the ratio of MAP reflects RAAS activation on salt restriction. It is interesting that and ERPF multiplied with (1 Ϫ hematocrit) ϫ 1000. heparin therapy never lowered aldosterone levels. Figure 2 Single-compartment infusion clearance measurements are depicts aldosterone levels in patients with diabetes. highly reproducible. Phantom studies showed that clearance Baseline active renin values were significantly higher in pa- values varied Ͻ5% in consecutive measurements. Head-on tients with diabetes and patients with GN compared with comparison between classical clearance that was based on urine healthy control subjects, reflecting RAAS activation in renal collection and infusion clearance in the same individual had disease. In all groups, active renin was significantly attenuated shown that the measured clearance values differ on average by by AngII. As expected, active renin values were significantly 8.4% and no more than 11% in individuals with a GFR Ͼ50 higher on low salt intake compared with a normal salt diet ml/min (25). This was reconfirmed in 1998: Clearance values (data not shown). (GFR) that were based on urine collection (without catheteriza- AngII levels increased on average by a factor of 950 during tion relying on spontaneous voiding) and single-compartment AngII infusion. There were no significant differences in AngII infusion clearance were compared in five individuals for 75 levels between any of the groups (data not shown). min. Clearance that was based on urine collection was 105 Ϯ 6.51 ml/min, and infusion clearance 96.2 Ϯ 10.2 ml/min, the Systemic Hemodynamics former being slightly but consistently higher by an average of AngII infusion was followed by a marked systemic pressor 10% (2 to 20%) (26). response with significant increases in MAP in all groups (for patients with diabetes, see Figure 3A). In patients with diabetes, Laboratory Procedures salt restriction resulted in a BP reduction, which reached sta- Venous blood samples were taken from a catheter that was inserted tistical significance only after heparin therapy (from 97.0 Ϯ 5.41 into the left antecubital vein at the end of the baseline and AngII to 85.5 Ϯ 4.1 mmHg). Notably, in patients with DN, the pressor periods, respectively. Blood samples were collected on ice and centri- response was more pronounced (29.5 Ϯ 9.86 versus 24.3 Ϯ 10.5 fuged immediately at 4°C, and the samples were stored at Ϫ20°C before mmHg in patients without diabetes and 24.5 Ϯ 9.51 mmHg in assay. control subjects), but this trend failed to reach statistical signif- Active renin was measured by an immunoradiometric assay (27). icance. However, in four patients with diabetes, dosage reduc- Aldosterone and AngII were determined by RIA (28,29). tions of AngII were necessary in seven of 16 measurements to Proteinuria was determined from 24-h collections with the pyrogallol avoid increases in BP above 200/100 mmHg. In contrast, pa- red-molybdate method (30). The intra-assay coefficient of variation of this method is Ͻ3.3%, and the interassay coefficient of variation is Ͻ3%. tients with GN and healthy control subjects did not require any Each value of urine protein was validated with concurrent creatinine dosage reductions. This finding suggests that patients with DN and expressed as the ratio between protein and creatinine excretion in are more susceptible to the hemodynamic actions of AngII. urine. Heparin treatment did not lower baseline BP values on any occasion. Statistical Analyses AngII infusion induced a moderate increase in HR, reaching Data are means Ϯ SD, unless stated otherwise. Within-group differ- statistical significance in all groups (in patients with diabetes ences were analyzed with paired t test. Between-group differences were 76.6 Ϯ 4.47 versus 70.0 Ϯ 2.7 bpm, in patients with GN 70.8 Ϯ Clin J Am Soc Nephrol 2: 58–67, 2007 Proteinuria-Lowering Effect of Heparin on RAAS 61

Table 1. Baseline characteristicsa

Parameters DN GN C P (ANOVA)

No. of patients 10 10 10 Age (yr) 55.2 Ϯ 5.55b 45.6 Ϯ 11.7 49.8 Ϯ 10.3 P Ͻ 0.05 (43 to 61) (28 to 71) (27 to 66) (DN versus GN) Gender distribution (M/F) 7/3 6/4 6/4 Disease distribution — 7 IgA-GN, 3 — membranous GN Body mass index (kg/m2) 30.1 Ϯ 5.61c 28.1 Ϯ 6.11 25.5 Ϯ 3.75 P ϭ 0.05 (21 to 41) (19 to 39) (20 to 34) (DN versus C) Duration of diabetes (yr) 15.1 Ϯ 4.65 —— (8 to 21) Age of onset of diabetes (yr) 38.5 Ϯ 9.40 —— (19 to 52) Ϯ HbA1c (mmol/L) 7.34 1.07 —— (5.30 to 8.50) BP (mmHg) 135 Ϯ 15.1/ 126 Ϯ 16.8/ 131 Ϯ 5.59/ NS 77 Ϯ 6.80 77 Ϯ 7.29 82 Ϯ 6.24 Serum sodium (mmol/L) 139 Ϯ 6.47 142 Ϯ 2.66 139 Ϯ 3.17 NS (125 to 146) (139 to 147) (134 to 143) Serum potassium (mmol/L) 4.36 Ϯ 0.31 4.47 Ϯ 0.45 4.57 Ϯ 0.61 NS (3.77 to 4.75) (3.78 to 4.96) (3.80 to 5.50) Serum creatinine (mg/dl) 1.42 Ϯ 0.68d 1.33 Ϯ 0.55d 0.89 Ϯ 0.16 P Ͻ 0.05 (0.81 to 2.78) (0.73 to 2.39) (0.66 to 1.07) (DN versus C) P Ͻ 0.05 (GN versus C) Proteinuria (mg/24 h) 800 Ϯ 508d 1958 Ϯ 1951d 112 Ϯ 27.5 P Ͻ 0.005 (194 to 1816) (259 to 6651) (72 to 154) (DN versus C) P Ͻ 0.01 (GN versus C)

aData are means Ϯ SD (range).C, healthy controls; DN, diabetic nephropathy; GN, nondiabetic renal disease (IgA or membranous glomerulonephritis); HbA1c, glycosylated hemoglobin. bSignificant differences between DN and GN. cSignificant differences between DN and C. dSignificant differences between DN/GN and C.

4.74 versus 66.4 Ϯ 4.25 bpm, in control subjects 67.5 Ϯ 2.70 infusion induced hyperfiltration (for patients with diabetes, see versus 63.3 Ϯ 2.34 bpm, AngII versus baseline; P Ͻ 0.05, respec- Figure 4C) and a marked increase in renal vascular resistance tively). Figure 3B shows HR in patients with diabetes. (data not shown), as expected and described previously (31). Heparin therapy did not alter the AngII-induced changes in Renal Hemodynamics renal and systemic hemodynamics during normal salt intake or Mean GFR at baseline was 83.8 Ϯ 20.5 ml/min per 1.73 m2 in during salt restriction. All groups showed identical renal he- patients with diabetes, 80.4 Ϯ 21.3 ml/min per 1.73 m2 in modynamic responses to AngII before and after heparin ther- patients with GN, and 107 Ϯ 26.9 ml/min per 1.73 m2 in apy, as demonstrated for ERPF in Figure 4D. healthy control subbjects, indicating mild impairment of renal function in renal patients with and without diabetes and nor- Proteinuria mal renal function in healthy control subjects. As expected, Heparin therapy caused a significant decrease in proteinuria GFR remained unchanged after AngII infusion in all groups only in patients with DN: On normal salt intake from 961 Ϯ 830 and all measurements (for patients with diabetes, see Figure to 561 Ϯ 493 mg/g creatinine per 24 h and even after salt 4A). restriction from 720 Ϯ 709 to values of 494 Ϯ 555 mg/g creat- Mean ERPF at baseline was 521 Ϯ 110 ml/min per 1.73 m2 in inine per 24 h (P ϭ 0.01 by Wilcoxon test; Figure 5A). There was patients with diabetes, 564 Ϯ 140 ml/min per 1.73 m2 in pa- a trend to a decreased proteinuria on low salt compared with tients with GN, and 583 Ϯ 152 ml/min per 1.73 m2 in control normal salt intake, but this trend failed to reach statistical subjects. AngII caused an average 35% reduction of the ERPF in significance. all groups (for patients with diabetes, see Figure 4B). AngII Figure 5B denotes the individual decrease of each patient 62 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 2: 58–67, 2007

Figure 2. Aldosterone levels in patients who had diabetes and had measurement A (before heparin on normal-salt diet), C (after heparin on normal-salt diet), B (before heparin on low-salt diet), and D (after heparin on low-salt diet). Baseline values (Ⅺ) and after angiotensin II (AngII) infusion (f). *P Ͻ 0.05 after AngII versus baseline by two-sided paired t test; #P Ͻ 0.05 baseline values on low-salt versus normal-salt diet by ANOVA. with DN on a normal salt diet. Note that patients with the most aparin in particular, a small, placebo-controlled crossover trial severe proteinuria experienced the greatest reduction in protein in patients with type 1 diabetes suggested that this LMWH excretion. In striking contrast, proteinuria remained unchanged might have antiproteinuric properties. However, the demon- in patients with GN: On normal salt 1632 Ϯ 1481 versus 1842 Ϯ strated trend to a reduced proteinuria failed to reach statistical 1744 mg/g creatinine per 24 h and on low salt 1640 Ϯ 1752 significance as a result of the very small group of six patients, versus 2059 Ϯ 1175 mg/g creatinine per 24 h, before versus after who had therapy for only 4 wk (17). Furthermore, two clinical heparin, respectively. studies in which sulodexide was used in microalbuminuric type 2 diabetes showed that this GAG formulation of fast- Discussion moving heparin and dermatansulfate that offers oral bioavail- This study shows that treatment with enoxaparin effectively ability also decreases proteinuria (35,36). It should be empha- lowers proteinuria in DN. The antiproteinuric effect of enox- sized that a great pharmacologic gap separates heparins from aparin persists even after lowering proteinuria by salt restric- GAG, because these polysaccharides have different degrees of tion, which is known to improve the beneficial effect of ACE sulfation, molecular weights, and diverse biologic activities. inhibitor treatment (32). The treatment with enoxaparin did not Nonetheless, our results suggest that these diverse compounds lower proteinuria in the other nephropathies tested. Because with heterogeneous structure have a common antiproteinuric patients with the most severe proteinuria had the greatest effect in DN. reduction in protein excretion, we postulate that patients with The results in DN differ from those reported by Rostoker et diabetes and overt proteinuria will experience the greatest ben- al. (37), who examined patients with primary GN. They found efit from heparin therapy. There was no evidence that enox- no decrease in proteinuria even in patients with nephrotic- aparin treatment modulated proteinuria by interfering with the range disease when they were treated with enoxaparin. These RAAS; aldosterone levels remained unchanged as did systemic contrasting results in patients with GN support our hypothesis and renal hemodynamic reactivity after AngII infusion. Enox- that the proteinuria-lowering effect of enoxaparin might be aparin did not lower BP values on any occasion, although BP specific for DN (37). However, there is no evidence from clinical control is an important determinant of proteinuria in treatment studies yet that heparins or GAG preserve renal function in any of DN (33). This finding is in contrast to animal models of entity of glomerular disease. How can the results of this study hypertension, in which unfractionated heparin lowered BP in be explained in the context of what is known about the inter- spontaneously hypertensive rats and Goldblatt hypertensive actions between proteoglycans and the RAAS? rats (34). First, there is substantial evidence that heparins may reduce Our results confirm and extend previous findings, demon- aldosterone levels (20). However, the effect of heparins on strating an antiproteinuric effect of heparins and glycosamino- aldosterone secretion may depend on their molecular weight or glycans (GAG) in DN (for review, see reference [19]). For enox- sulfation pattern. Most studies that described hyperkalemia Clin J Am Soc Nephrol 2: 58–67, 2007 Proteinuria-Lowering Effect of Heparin on RAAS 63

Figure 3. Systemic hemodynamics. Mean arterial pressure (MAP; A) and heart rate (HR; B) in patients with diabetes and with measurement A (before heparin on normal-salt diet), C (after heparin on normal-salt diet), B (before heparin on low-salt diet), and D (after heparin on low-salt diet). Baseline values (Ⅺ) and after AngII infusion (f). *P Ͻ 0.0001 (MAP) and P Ͻ 0.05 (HR) after AngII versus baseline by two-sided paired t test. #P Ͻ 0.05 after salt restriction versus normal-salt diet by ANOVA. and decreased aldosterone levels after heparin treatment used is mediated through modulation of AngII signaling in intrare- unfractionated heparins. Second, heparins and GAG modulate nal smooth muscle cells or mesangial cells. Third, AngII mod- calcium fluxes in smooth muscle cells and mesangial cells after ulates HSPG production in cultured human podocytes (13). To AngII stimulation in vitro (21,22). These effects seem to depend what extent a preservation of HSPG synthesis after ACE inhi- on the sulfation pattern of the heparins used and on the sulfa- bition or ATR1 blockade contributes to their antiproteinuric tion pattern of the GAG side chains in the HSPG on the surface effects in DN remains to be elucidated. ACE inhibition has been of the investigated cells (22). Our in vivo data do not support the shown to preserve staining of the GBM for HSPG core protein hypothesis that the antiproteinuric effect of enoxaparin in DN and GAG side chains in the adriamycin model in the rat (35). 64 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 2: 58–67, 2007

Figure 4. Renal hemodynamics. GFR (A), effective renal plasma flow (ERPF; B), and filtration fraction (FF; C) in patients with diabetes and with measurement A (before heparin on normal-salt diet), C (after heparin on normal-salt diet), B (before heparin on low-salt diet), and D (after heparin on low-salt diet). Baseline values (Ⅺ) and after AngII infusion (f). *P Ͻ 0.01 after AngII versus baseline by two-sided paired t test. (D) Percentage decrease of ERPF in patients with diabetes (DN), patients with glomerulonephritis (GN), and healthy controls (C) after AngII infusion versus baseline with measurement A (before heparin on normal-salt diet), (C) after heparin on normal-salt diet, B (before heparin on low-salt diet, and D (after heparin on low-salt diet).

However, it has been shown convincingly by several groups inhibits heparanase-1 activity, an endoglycosidase that spe- (36,37) that nephrin expression is reduced in experimental and cifically degrades HSPG. Furthermore, heparanase-1 seems human nephropathy and that AngII antagonist treatment or to be overexpressed in patients with DN, but, in contrasting, ACE inhibition can prevent these changes. We are not aware of not in patients with nondiabetic nephropathies (40). These any studies on the effect of heparins on nephrin expression in findings provide an intriguing hypothesis to explain, podocytes. If not by interaction with the RAAS, then how do why—in our study—heparins have antiproteinuric effects in heparins reduce proteinuria in DN? DN but not in patients with GN. The exact mode of action of heparins and sulodexide in DN In evaluating the results of our study, we point to specific is not clear at present. As reviewed by Striker et al. (38), in vitro limitations of our protocol: The patients were not studied in studies demonstrate that HSPG and GAG modulate extracellu- a randomly assigned sequence for reasons of feasibility. lar matrix composition (e.g., by increasing fibronectin and However, we do not believe that a random order could have thrombospondin synthesis and by regulating the assembly of influenced our results. Furthermore, a crossover design with laminin and collagen IV via specific receptors). HSPG and GAG washout periods of long duration was not considered nec- influence mesangial cell turnover by acting on both production essary to exclude carryover effects of ACE inhibition on and effects of several growth factors, such as TGF-␤1 and basic proteinuria. Relevant interference of ACE inhibition with the fibroblast growth factor (38), but they also affect vascular en- effects of enoxaparin was believed to be minimized by the dothelial growth factor and have, depending on dosage, anti- long-term ACE inhibitor pretreatment for 6 mo in all patients proliferative rather than proliferative effects (19). with diabetes except one. Second, the method used—which An interesting hypothesis on the mechanism of the pro- is applicable in human volunteers—to investigate the influ- teinuria-lowering effect of heparins emerged recently: Xu et ence of heparin therapy on the systemic and renal hemody- al. (39) demonstrated in an in vitro study that sulodexide namic effects of AngII cannot rule out that a modulation of Clin J Am Soc Nephrol 2: 58–67, 2007 Proteinuria-Lowering Effect of Heparin on RAAS 65

Figure 5. (A) Proteinuria in patients with diabetes validated with concurrent creatinine before heparin (Ⅺ) and after heparin (f). A, normal-salt diet before heparin; C, normal-salt diet after heparin; B, low-salt diet before heparin; D, low-salt diet after heparin. *P ϭ 0.01 after heparin versus before heparin by Wilcoxon test for paired differences. (B) Individual values of proteinuria with concurrent creatinine in 10 patients who had diabetes and were on a normal-salt diet and had measurements A (before heparin) and C (after heparin). the local RAAS occurred. Third, we were not able to obtain Conclusion data on proteinuria after AngII infusion. However, we be- Our results show that treatment with the LMWH enoxaparin lieve that this limitation is of minor importance, because it effectively lowers proteinuria in patients with DN, who were on ACE was demonstrated convincingly by Gansevoort et al. (41) that inhibition therapy. The efficacy of this treatment cannot be explained the antiproteinuric effects of RAAS blockade occur long after by the effects of enoxaparin on aldosterone levels. Our study on the the acute renal hemodynamic alterations. effects of heparin on AngII-induced renal hemodynamic changes We believe that the mechanisms of heparin action in renal does not support the notion that the known heparin-induced mod- disease are of substantial clinical relevance. The beneficial effect ulation of AngII signaling, in mesangial or smooth muscle cells, can of sulodexide suggested by two preliminary studies (42,43) was explain its antiproteinuric effect. These findings suggest that heparin confirmed recently by a larger randomized trial. In the treatment may have a beneficial effect in DN independent from and Di.N.A.S. trial, an oral sulodexide dosage dependently reduced in addition to ACE inhibition or ATR1 blockade. albumuria in 223 patients with micro- and macroalbuminuria and DN that was caused by type 1 as well as type 2 diabetes. Acknowledgments The antiproteinuric effect of sulodexide was long lasting and This study was supported in part by a grant from Aventis Pharma. seemingly additive to the ACE inhibitory effect (44). We gratefully acknowledge Irene Sellger for technical assistance. 66 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 2: 58–67, 2007

Disclosures 15. Gambaro G, Cavazzana AO, Luzi P, Piccoli A, Borsatti A, None. Crepaldi G, Marchi E, Venturini AP, Baggio B: Glycosami- noglycans prevent morphological renal alterations and al- buminuria in diabetic rats. Kidney Int 42: 285–291, 1992 References 16. Gambaro G, Venturini AP, Noonan DM, Fries W, Re G, 1. Deckert T, Feldt-Rasmussen B, Borch-Johnsen K, Jensen T, Garbisa S, Milanesi C, Pesarini A, Borsatti A, Marchi E, et Kofoed-Enevoldsen A: Albuminuria reflects widespread al.: Treatment with a glycosaminoglycan formulation ame- vascular damage. Diabetologia 32: 517–523, 1989 liorates experimental diabetic nephropathy. Kidney Int 46: 2. Parving HH, Osterby R, Anderson PW, Hsueh W: Diabetic 797–806, 1994 nephropathy. In: The Kidney, 5th Ed., edited by Brenner 17. Tamsma JT, van der Woude FJ, Lemkes HHPJ: Effect of BM, Philadelphia, London, Toronto, Saunders Company, sulfated glycosaminoglycans on albuminuria in patients 1996, pp 1864–1892 with overt diabetic (type-1) nephropathy. Nephrol Dial 3. Yusuf S, Sleight P, Pogue J, Bosch J, Davies R, Dagenais G: Transplant 11: 182–185, 1996 Effects of an angiotensin-converting-enzyme inhibitor, 18. Solini A, Carraro A, Barzon I, Crepaldi G: Therapy with ramipril, on cardiovascular events in high-risk patients. glycosaminoglycans lowers albumin excretion rate in non- The Heart Outcomes Prevention Evaluation Study Inves- insulin dependent diabetic patients with microalbumin- tigators. N Engl J Med 342:145–153, 2000 uria. Diabetes Nutr Metab 7: 304–307, 1994 4. Lewis EJ, Hunsicker LG, Bain RP, Rohde RD: The effect of 19. Gambaro G, van der Woude FJ: Glycosaminoglycans: Use angiotensin-converting-enzyme inhibition on diabetic ne- in treatment of diabetic nephropathy. J Am Soc Nephrol 11: phropathy. N Engl J Med 329: 1456–1462, 1993 359–368, 2000 5. Lewis EJ, Hunsicker LG, Clarke WR, Berl T, Pohl MA, 20. Oster JR, Singer I, Fishman LM: Heparin-induced aldoste- Lewis JB: Renoprotective effect of the angiotensin-receptor rone suppression and hyperkalemia. Am J Med 98: 575–586, antagonist irbesartan in patients with nephropathy due to 1995 type 2 diabetes. N Engl J Med 345: 851–860, 2001 21. Zaragoza R, Battle-Tracy KM, Owen NE: Heparin inhibits 6. Brenner BM, Cooper ME, de Zeeuw D, Keane WF, Mitch NaϩHϩ exchange in vascular smooth muscle cells. Am J WE, Parving HH: Effects of losartan on renal and cardio- Physiol 258: C46–C53, 1990 vascular outcomes in patients with type 2 diabetes and 22. Koppel H, Yard BA, Christ M, Wehling M, van der Woude nephropathy. N Engl J Med 345: 861–869, 2001 FJ: Modulation of angiotensin II-mediated signalling by 7. Hostetter TH, Rennke HG, Brenner BM: The case for intra- heparan sulphate glycosaminoglycans. Nephrol Dial Trans- renal hypertension in the initiation and progression of plant 18: 2240–2247, 2003 diabetic and other glomerulopathies. Am J Med 72: 375–380, 23. Luik PT, Kerstens MN, Hoogenberg K, Navis GJ, Dullaart 1982 RPF: Low plasma aldosterone despite normal plasma renin 8. Kagami SH, Border WA, Miller DA, Noble NA: Angioten- activity in uncomplicated type I diabetes mellitus: Effects sin II stimulates extracellular matrix protein synthesis of RAAS stimulation. Eur J Clin Invest 33: 787–793, 2003 through induction of transforming growth factor-beta ex- 24. Pedersen JF, Pedersen MF, Madsen PO: Clinical renal test- pression in rat glomerular mesangial cells. J Clin Invest 93: ing by external double isotope monitoring technique. Nucl 2431–2437, 1994 Med 10: 16–24, 1971 9. Wagner M, Mascareno E, Siddiqui M: Cardiac hypertro- 25. Mo¨hring K: Radiolabeled molecules for the comparative phy: Signal transduction, transcriptional adaptation and evaluation and clinical application of a catheter-free, uni- altered growth control. Ann N Y Acad Sci 874: 1–10, 1999 laterally differentiated and simultaneous determination of 10. Diep QN, El Mabrouk M, Cohn S, Endemann D, Amiri F, glomerular filtrate and renal plasma flow. Heidelberg, Ha- Virdis A, Neves MF, Schiffrin EL: Structure endothelial bilitationsschrift, 1973 function, cell growth and inflammation in blood vessels of 26. Ritz E, Benck U, Franek E, Keller C, Seyfarth M, Clorius J: angiotensin II infused rats: Role of peroxisome prolifera- Effects of smoking on renal hemodynamics in healthy vol- tor-activated receptor gamma. Circulation 105: 2295–2302, unteers and in patients with glomerular disease. JAmSoc 2002 Nephrol 9: 1798–1804, 1998 11. Wolf G: Free radical production and angiotensin. Curr 27. Rosmalen F, van Hoek M, Martens M: Development of an Hypertens Res 2: 167–173, 2000 IRMA for the quantification of active renin. In: Abstract 12. Van Det NF, Tamsma JT, van den Born J, Verhagen NAM, Book. Ninth International Congress of Endocrinology 1992, ed- van den Heuvel LPWJ, Lowik WGM, Berden JHM, Bruijn ited by Mornex R, Jaffiol C, Leclere J, Bull Cancer 79[Suppl JA, Daha MR, van der Woude FJ: Differential effects of 1]: 1–8, 1992 angiotensin II and transforming growth factor beta on the 28. Ito T, Woo J, Haning R, Horton R: A radioimmunoassay for production of heparan sulfate proteoglycan by mesangial aldosterone in human peripheral plasma including a com- cells in vitro. J Am Soc Nephrol 7: 1015–1023, 1996 parison of alternate techniques. J Clin Endocrinol Metab 34: 13. Brinkkoetter PT, Holtgrefe S, van der Woude FJ, Yard BA: 106–112, 1972 Angiotensin II type 1-receptor mediated changes in hepa- 29. Suzuki H, Ferrario CM, Speth RC, Brosnihan KB, Smeby ran sulfate proteoglycans in human SV40 transformed RR, De Silva P: Alterations in plasma and cerebrospinal podocytes. J Am Soc Nephrol 15: 33–40, 2004 fluid norepinephrine and angiotensin II during the devel- 14. Reddi A, Ramamurthi R, Miller M, Dhuper S, Lasker N: opment of renal hypertension in conscious dogs. Hyperten- Enalapril improves albuminuria by preventing glomerular sion 5: 139–148, 1983 loss of heparan sulphate in diabetic rats. Biochem Med 30. Watanabe N, Kamei S, Ohkubo A, Yamanake M, Ohsawa Metab Biol 45: 119–131, 1991 S, Makino K, Tokuda K: Urinary protein as measured with Clin J Am Soc Nephrol 2: 58–67, 2007 Proteinuria-Lowering Effect of Heparin on RAAS 67

a pyrogallol red-molybdate complex, manually and in a nephrin expression in a model of diabetes and hyperten- Hitachi 726 automated analyzer. Clin Chem 32: 1551–1554, sion. Diabetologia 44: 874–877, 2001 1986 38. Striker LJ, Peten EP, Elliot SJ, Doi T, Striker GE: Mesangial 31. Heeg JE, de Jong PE, van der Hem GK, de Zeeuw D: cell turnover: Effect of heparin and peptide growth factors. Angiotensin II does not acutely reverse the reduction of Lab Invest 64: 446–456, 1991 proteinuria by long-term ACE inhibition. Kidney Int 40: 39. Xu X, Rao G, Maxhimer JB, Prinz RA, Lewis EJ: Mechanism 734–741, 1991 of action of sulodexide-mediated control of diabetic pro- 32. Van der Kleij FG, Schmidt A, Navis GJ, Haas M, Yilmaz N, teinuria: inhibition of heparanase-1 activity [Abstract]. de Jong PE, Mayer G, de Zeeuw D: Angiotensin converting J Am Soc Nephrol 16: 673A, 2005 enzyme insertion/deletion polymorphism and short-term 40. Maxhimer JB, Prinz RA, Lewis EJ, Xu X: Increased hepara- renal response to ACE inhibition: Role of sodium status. nase-1 expression in the kidneys of patients with diabetic Kidney Int 63[Suppl]: 23–26, 1997 nephropathy [Abstract]. J Am Soc Nephrol 16: 407A, 2005 33. Gilbert RE, Krum H, Wilkinson-Berka J, Kelly DJ: The 41. Gansevoort RT, de Zeeuw D, de Jong PE: Dissociation renin-angiotensin system and the long-term complications between the course of the hemodynamic and antiprotein- of diabetes: Pathophysiological and therapeutic consider- uric effects of angiotensin I converting enzyme inhibition. ations. Diabet Med 20: 607–621, 2003 Kidney Int 44: 579–584, 1993 34. Mandal AK, Lyden TW, Saklayen MG: Heparin lowers 42. Solini A, Vergnani L, Ricci F, Crepaldi G: Glycosaminogly- blood pressure: Biological and clinical perspectives. Kidney cans delay the progression of nephropathy in NIDDM. Int 47: 1017–1030, 1995 Diabetes Care 20: 813–817, 1997 35. Wapstra FH, Navis GJ, van Goor H, van den Born J, Berden 43. Velussi M, Cernigoi AM, Dapas F, De Monte A: Glycos- JHM, de Jong PE, de Zeeuw D: ACE inhibition preserves aminoglycans oral therapy reduces microalbuminuria, heparan sulfate proteoglycans in the glomerular basement blood fibrinogen levels and limb arteriopathy clinical signs membrane of rats with established adriamycin nephropa- in patients with non-insulin dependent diabetes mellitus. thy. Exp Nephrol 9: 21–27, 2001 Diabetes Nutr Metab 9: 53–58, 1996 36. Doublier S, Salvidio G, Lupia E, Ruotsalainen V, Verzola D, 44. Gambaro G, Kinalska I, Oksa A, Pontuch P, Hertlova M, Deferrari G, Camussi G: Nephrin expression is reduced in Olsovsky J, Manitius J, Fedele D, Czekalski S, Perusicova human diabetic nephropathy: Evidence for a distinct role J, Skrha J, Taton J, Grzeszczak W, Crepaldi G: Oral for glycated albumin and angiotensin II. Diabetes 52: 1023– sulodexide reduces albuminuria in microalbuminuric 1030, 2003 and macroalbuminuric type 1 and type 2 diabetic pa- 37. Bonnet F, Cooper ME, Kawachi H, Allen TJ, Boner G, Cao tients: The Di.N.A.S. randomized trial. J Am Soc Nephrol Z: Irbesartan normalises the deficiency in glomerular 13: 1615–1625, 2002

It is well known that blockade of the renin-angiotensin system lowers protein excretion in diabetic nephropathy. Benck et al. show that heparin can also lower proteinuria without affecting the renin-angiotensin system. In this month’s issue of JASN, Sasser et al. (pp. 143–154) show that endothelin receptor blockade can also reduce diabetic renal injury via an anti-inflammatory mechanism that does not involve the renin-angiotensin system, which gives us the possibility that other therapeutic strategies and pathways can reduce proteinuria.