Increased glomerular albumin permeability in old spontaneously hypertensive rats

Nephrol Dial Transplant (2004) 19: 1724–1731 DOI: 10.1093/ndt/gfh276 Advance Access publication 21 April 2004 Original Article Increased glomerular albumin permeability in old spontaneously hypertensive rats Omran Bakoush1, Jan Tencer1, Ole Torffvit2, Olav Tenstad3, Ingela Skogvall1 and Bengt Rippe1 1 Department of Nephrology and 2Department of Medicine, Lund University Hospital, Sweden and Department of Physiology, University of Bergen, Bergen, Norway 3 Abstract Background. Severe long-standing hypertension is associated with an increased urinary protein excretion. Methods. To investigate the mechanisms of this proteinuria, we measured the glomerular clearances and calculated the glomerular sieving coefficients (y) for neutral albumin (yo-alb) and for native albumin (yalb) in spontaneously hypertensive rats (SHR) at the ages of 3, 9 and 14 months, in comparison with age-matched normal control Wistar rats (NCR). The hypothesis was that increases in the glomerular permeability of both negatively charged and neutral albumin would indicate a preferential size-selective dysfunction of the glomerular capillary wall (GCW), while an increased permeability to negatively charged albumin, as compared with neutral albumin, predominantly would indicate a charge-selectivity dysfunction of the GCW. A tissue (renal) uptake technique together with urinary sampling was used to assess y. The glomerular filtration rate was assessed using the plasma to urine clearance of 51Cr-EDTA. Results. The yalb in SHR increased 2.6 times at 14 months of age as compared with at 3 months, while there was no significant change of yalb in NCR with age. Furthermore, the increased yalb in old SHR correlated significantly with an increase in yo-alb (r ¼ 0.86, P<0.001), suggesting that albuminuria in old SHR primarily results from an increased number of rather unselective (‘large’) pores in the glomerular filter. Conclusions. In old age, but not at a young age, hypertensive rats develop proteinuria as a result of dysfunction of the glomerular capillary filter, affecting primarily its size-selectivity. The changes are functionally compatible with the appearance in the glomerular barrier of an increased number of more unselective pores. Correspondence and offprint requests to: Omran Bakoush, MD, Department of Nephrology, Lund University Hospital, S-221 85 Lund, Sweden. Email: [email protected] Keywords: albuminuria; charge-selectivity; glomerular capillary wall; hypertension; macromolecular transport Introduction Renal insufficiency due to nephrosclerosis is frequently seen in patients with essential hypertension. Proteinuria is a major sign of hypertensive renal damage [1]. Glomerular filtration of proteins is governed by several factors, including the size- and charge-selectivity of the filtration barrier and the size, shape and charge of the transported protein [2–4]. In addition, transport of molecules across the glomerular filter is also influenced by plasma flow and intraglomerular pressure [3,4]. A widely used description of macromolecular transport across the glomerular capillary wall (GCW) is based on the concept of hypothetical, water-filled cylindrical pores perforating the glomerular filter at the right angle. In a ‘two-pore with a shunt’ model, the permeable structures consist of two major populations of pores [2–4]. The vast majority of the pores are ‘small pores’ of uncharged radius
38 Å, but being negatively charged [3]. The remaining pore population consists of a very small number of ‘large pores’ of radius
100 Å, also being negatively charged [2]. The small pores are essentially impermeable to (negatively charged) macromolecules of the size of albumin or larger. Such molecules are transported by convection across the large pores [2,3]. In addition to the two populations of pores, the intact GCW exhibits very sporadic physiological ‘membrane defects’ or ‘shunts’, large enough to allow transport of very large proteins and red blood cells [4]. In order to examine the mechanisms of proteinuria in hypertension, especially with respect to capillary charge-selectivity, we studied the glomerular permeability of native (negatively charged) and chargemodified, neutral albumin in aging hypertensive rats using age-matched normotensive Wistar rats (NCR) as controls. Nephrol Dial Transplant Vol. 19 No. 7 ß ERA–EDTA 2004; all rights reserved Albumin glomerular permeability in hypertensive rats 1725 Previously, micropuncture techniques have been used extensively for the collection of tubular fluid and for assessing glomerular permselectivity. These techniques are compromised by the fact that proteins can bind to the glass pipette and by the possibility of contamination of the tubular fluid by extratubular proteins. To avoid contamination by extratubular proteins, Tojo and Endou [5] developed a new fractional double-barrel micropuncture technique for the collection of tubular fluid, which yielded an albumin sieving coefficient of 0.00062 in NCR. This is almost identical to the y obtained by this group in a recent study using the present tissue uptake technique (0.00066) [3]. To our knowledge, no assessment of albumin sieving in spontaneously hypertensive rats (SHR) using novel micropuncture techniques has been reported. This study will, thus, be the first to report a y-value for albumin in young and old SHR in vivo, seemingly unaffected by the drawbacks of the classical micropuncture procedures. At the same time, by simultaneously assessing the clearance of native (negatively charged) and neutralized albumin, having similar size and shape, we obtained an opportunity to examine the controversial issue of the impact of negative charge of the GCW on the glomerular sieving of macromolecules. Neutralized human serum albumin (nHSA) was obtained by a graded modification of the carboxyl groups using a procedure modified from that described by Hoare and Koshland [6], as follows. HSA (1.5 g) was dissolved in 15 ml of 0.133 M glycine methyl ester at pH 4.75 (at room temperature). A solution of 5 ml of 0.04 M EDC [N-ethyl-N0 (3-dimethylaminopropyl) carbodiimide hydrochloride] was then added to the mixture to initiate the reaction. The pH was continuously recorded and kept at 4.75 by addition of 0.1 M NaOH. Aliquots (1 ml) were removed every 5 min until 60 min and immediately added to 1 ml of 4.0 M acetate buffer at pH 4.75 to quench the reaction. After standing for a few minutes at room temperature, these solutions were dialysed overnight against two changes of 10 l distilled water and the dialysate was freeze-dried and stored at 20 C. The effect of the reaction was evaluated by isoelectric focusing using a vertical minigel system (CBS Scientific Company, Inc., CA, USA) and NOVEXTM (Novel Experimental Technology, San Diego, CA, USA) pre-cast gels. A reaction time of 45 min produced albumin with an average isoelectric point close to 7.4 without any significant change in hydrodynamic radius, as measured by high-performance liquid chromatography (Superdex 75 HR and Superose 12 HR). During storage at 20 C, the neutralized albumin was found to be maintaining its isoelectric point for up to 6–7 months. Subjects and methods Tracers and labelling procedures Thirty-four male SHR and 27 male Wistar NCR were obtained from the animal house (Mollegaard, Stensved, Denmark) at the age of 3 months. The rats were kept on standard chow and had free access to water before the experiments. The animal experiments were approved by the Animal Ethical Committee at Lund University. The studies were performed according to two study protocols. In protocol 1, the glomerular clearances of native albumin and of 51CrEDTA [glomerular filtration rate (GFR)] were simultaneously measured in six SHR at the age of 3 months and 11 age-matched NCR. In protocol 2, measurements of native and neutral albumin clearances were conducted simultaneously in seven SHR at the age of 9 months, nine SHR at the age of 14 months and in six and five age-matched NCR controls, respectively. In parallel experiments, successful simultaneous measurements of GFR were performed in seven SHR at the age of 9 months, five SHR at the age of 14 months and in two and three age-matched NCR controls, respectively. All rats were anaesthetized by an intraperitoneal injection of pentobarbital (50 mg/g body weight). Supplementary injections of 10–20% of the initial dose were given intravenously when necessary. The animals were tracheotomized to ensure free airways. The tail artery was cannulated for continuous recording of the mean arterial pressure and for subsequent administration of drugs. The right carotid artery and the left jugular vein were cannulated for infusion and blood sampling purposes. A catheter was placed in the urinary bladder via an abdominal incision for continuous urine sampling. The external urethral opening was closed with Histoacryl (Melsungen, Germany) to prevent urine losses. The rats were allowed to stabilize after surgery for
0.5 h before the clearance measurements. Neutralization of human serum albumin Neutralized albumin was labelled with 131I using Iodo-Gen. Briefly, 0.1 mg 1,3,4,6-tetrachloro-3a,6a-diphenylglycouril (product no. T0656; Sigma-Aldrich Co.) dissolved in 0.1 ml chloroform was dispersed in a 1.8 ml Nunc vial (NuncKamstrup, Roskilde, Denmark). A film of the virtually water-insoluble Iodo-Gen was formed in the Nunc vial by allowing the chloroform to evaporate to dryness under nitrogen. Then, 1 ml of 0.05 M phosphate-buffered saline (pH 7.5, containing 1–2 mg protein to be labelled), 5 MBq 131I (Institute for Energy Technique, Kjeller, Norway) and 15 ml of 0.01 M NaI were added and the iodinating tube gently agitated for 10 min before the reaction was terminated by removing the solution from the Iodo-Gen tube. Unincorporated iodine-isotope accounting for <10% of the total radioactivity, as estimated by trichloroacetic acid (TCA) precipitation, was removed by dialysing the tracer against 1000 ml of 0.9% saline containing 0.02% azide. The stock solution was stored in the dark at 4 C and dialysed for 24 h before use. The nHSA had an isoelectric point (pkI) of 7.8, a molecular weight of 70 kDa and a SE radius of 35 Å. The pre-labelled native albumin (125I-HSA; Institute for Energy Technique, Kjeller, Norway) had an isoelectric point of 4.9, a molecular weight of 69 kDa and a SE radius of 35.5 Å. The level of free (unbound) 125I and 131I was always checked before use by TCA precipitation and was kept below 1.5% (usually <0.5%). Tissue uptake technique The technique has been described in detail and validated by Tenstad et al. [7]. When a tracer protein is added to the plasma compartment, it will mix with the plasma, dissipate within the extracellular space and, depending on size, filter 1726 across the glomerular barrier. After appearing in the Bowman’s space, it will be reabsorbed, more or less completely, by the renal proximal tubules to be processed by the tubular cells. During the first 8–12 min of tracer infusion, the protein breakdown, and the subsequent reabsorption to the plasma of split products, will be negligible, while a tiny fraction of the tracer will appear in the urine, which can be collected continuously. The total kidney uptake during 8–12 min plus the urinary excretion of the filtered protein will, thus, closely reflect the total protein clearance from the plasma to the Bowman’s capsule space during this limited period of time. This is the principle utilized in the present experiment. In the present study, we infused simultaneously 125I-native albumin (125I-HSA) and 131I-neutral albumin (131I-nHSA) into the jugular vein over a period of 10–12 min. Blood sampling from the carotid artery (20 ml at a time), approximately every 2 min, was performed using microcapillaries, while urine sampling was performed for the entire infusion period. After
8 min of tracer infusion, a whole body vascular washout was started by rapidly infusing (at 20 ml/min) a half-and-half mixture of 0.9% saline and heparinized horse serum (SVA, Uppsala, Sweden), containing 1 mg/l papaverine (P 3510; Sigma, St Louis, MO, USA) as a vasodilator, via the jugular vein, after having cut the inferior vena cava open. This usually occurred at
10 min after the start of the tracer infusion, because the laparatomy usually lasted 2–3 min. During the subsequent 8 min of washout, the animals usually expired within the first 2 min. During the last 4 min of this procedure, the tracer concentration in the effluent rinse fluid was <0.1% of the initial concentration. The two kidneys were then removed, blotted, weighed and assessed for radioactivity. TCA (10%)-precipitable urine radioactivity was assessed and included in the clearance measurements. The inner renal medulla (rich in interstitial tissue) was dissected away from the rest of the kidney and was not included in the radioactivity measurements. All radioactivity measurements were performed in a gamma-scintillation counter (Wizard 1480; LKP Wallac, Turku, Finland). Appropriate corrections for radioactive decay and spillover from the 51Cr or 131I into the 125I channel were performed. GFR measurements and histology For GFR measurements, 51Cr-EDTA (Amersham Biosciences, Amersham, UK) was given in priming dose i.v. (0.09 MBq in 0.2 ml), followed by a constant infusion (0.005 MBq/min) for repeated measurements of the plasma to urine 51Cr-EDTA clearance during 20 min intervals throughout the study. Before the onset of measurements, an infusion of 4 ml horse serum was given to elevate the GFR in a standardized manner in both NCR and SHR. After the GFR measurements, the animals were sacrificed using i.v. KCl and the kidneys were removed and promptly sectioned and postfixed in 10% buffered formalin. The histological examination was by light microscopy at the Institution of Pathology, Herlev Hospital, University of Copenhagen, Copenhagen, Denmark. Semi-quantitative scoring of glomerular and interstitial injury was done blindly. Urine albumin was analysed by enzyme-linked immunosorbent assay. Rat albumin (A6414; Sigma) and rabbit anti-rat antibodies (batch 4961; Nordic Immunology Laboratories, O. Bakoush et al. Tillberg, the Netherlands) were used. The latter was diluted 1:65 000 (v/v). The detection limits were 16 mg/l or 1.6 ng. The conjugate used was swine anti-rabbit alkaline phosphatase (306; DAKO, Copenhagen, Denmark), diluted 1:40 (v/v). Enzyme activity was determined with phosphatase substrate tablets (104; Sigma). Urine was diluted 1:10 000–1:80 000 times. Urine creatinine was determined enzymatically using a Kodak Ektachem 700 XR-C system. Calculations Renal tracer protein clearance (Cl) was calculated from the amount of tracer radioactivity accumulated in the two kidneys þ the TCA-precipitable urine tracer activity (collected during the tracer infusion period) divided by the average venous plasma tracer concentration and by the effective plasma tracer time until sacrifice (
11 min), i.e. by the plasma tracer ‘area under the curve’. Protein sieving coefficients (y) were calculated by dividing the measured protein Cl by the GFR. GFR was calculated from: GFR ¼ CuE  Vu Cpw 1 where CuE represents the urine concentration of Cr-EDTA and Vu represents the urine flow (per min) and where the Cpw is the plasma water concentration of Cr-EDTA. Cpw was obtained from: Cpw ¼ CpE 0:984  0:000718  Cprot 2 where CpE represents the plasma concentration of Cr-EDTA and Cprot represents the plasma concentration of total protein. Theoretical analysis Data on yalb from the rat groups studied were analysed using the two-pore model of membrane permeability [2,3]. Although a number of highly sophisticated glomerular sieving models have been published recently [8], we chose the pore model, because this model has been applied widely over the past few decades. Furthermore, the pore model is the simplest model that can adequately describe glomerular transport data [3]. Since native albumin is negatively charged, it should be completely excluded from the small-pore pathway in the glomerular filter and, thus, be confined to convective transport through large pores [3]. Based on these considerations, the large-pore volume flow (JvL) could be determined from the clearance of native albumin at any level of GFR. Furthermore, if more large pores were formed during increases in glomerular permeability, the impact of the concomitantly increased JvL upon the relationship between the sieving coefficients for native and neutral albumin could be analysed. These analyses are presented in the appendix. Statistical analysis Values are given as means±SE. Statistical comparison between the groups was performed with the non-parametric Albumin glomerular permeability in hypertensive rats 1727 Mann–Whitney test. Correlation was tested using Spearman’s correlation coefficient. The statistical package for social science (SPSS, version 10) was used. (P < 0.001) .0015 SHR Results As shown in Table 1, the body weight of the SHR at 9 and 14 months was lower than that of the age-matched NCR (P ¼ 0.001) and, similarly, the kidney weight was lower in SHR than in NCR at 9 and 14 months of age (P<0.05). Mean arterial blood pressure was higher in SHR at 3, 9 and 14 months compared with that of agematched NCR (P<0.01). Urinary albumin excretion (mg/mmol creatinine) was markedly increased in SHR in comparison to NCR at 14 months of age (1036 vs 142 mg/mmol; P<0.01) (Table 1). The calculated glomerular sieving coefficient for native albumin (yalb) in SHR increased by 2.6 times at the age of 14 months compared with at 3 months (Figure 1). It, thus, increased from 5.0 (±0.5)  104 at 3 months to 7.6 (±0.8)  104 at 9 months and to 12.9 (±0.9)  104 at 14 months of age (P<0.001), while yalb did not change significantly with age in NCR, remaining at 7.0 (±0.5)  104 at 3 and 9 months and at 7.2 (±0.9)  104 at 14 months of age, respectively [not significant (NS); Figure 1]. At the same time, the yo-alb in SHR increased from 50.8 (±2.4)  104 at 9 months to 72.9 (±3.8)  104 at 14 months of age (P ¼ 0.001), while it was stable with age in NCR, namely remaining at 47.3 (±2.7)  104 and 56.9 (±4.8)  104, respectively (NS; Figure 2). The increase in yalb in old SHR was significantly correlated with the increase in yo-alb (r ¼ 0.86, P<0.001; Figure 3) with a regression coefficient of 0.213±0.034 (Figure 3). The percentual tubular albumin reabsorption, calculated as the albumin radioactivity in the cortex divided by the total albumin radioactivity excreted (cortex þ urine), Sieving coefficient of native albumin .0013 (P = 0.85, ns) n=9 .001 NCR n=5 .0008 n=7 n=11 n=6 .0005 n=6 .0003 0 SHR 3 months SHR 14 months NCR 9 months SHR 9 months NCR 3 months NCR 14 months Fig. 1. Values of native albumin sieving coefficient (yalb) in animals studied (SHR and NCR) at 3, 9 and 14 months of age. averaged 76 (±1.8)% for native albumin and 96 (±0.4)% for neutral albumin. There was no statistically significant difference in percentual tubular reabsorption of albumin between SHR and NCR at any age. The GFR (ml/min/g kidney weight) decreased in SHR by 33%, i.e. from 1.26 (±0.048) at 3 months to 1.15 (±0.059) at 9 months and to 0.84 (±0.08) at 14 months of age (P<0.001; Figure 4). In NCR, however, the GFR did not change significantly with age, being 1.39 (±0.083) ml/min/g kidney weight at 3 months of age, 1.31 (±0.019) ml/min/g kidney weight at 9 months of age and 1.13 (±0.051) ml/min/g kidney weight at 14 months of age (P ¼ 0.13, NS). Histological examination of the kidneys demonstrated moderate to severe degrees of fibrotic changes Table 1. Mean value (±SE) of weight, kidney weight, mean arterial blood pressure, urinary albumin excretion and ratio between glomerular clearance of neutral and native albumin in SHR groups compared with age-matched NCR SHR Weight (g) 3 months 346 (±19) 9 months 409 (±11) 14 months 398 (±9) Kidney weight (g) 3 months 2.0 (±0.09) 9 months 2.6 (±0.06) 14 months 2.6 (±0.07) Mean arterial blood pressure (mmHg) 3 months 177 (±4) 9 months 169 (±5) 14 months 170 (±4.3) Urinary albumin creatinine index (mg/mmol) 9 months 213 (±30) 14 months 1036 (±169) Ratio between neutral and native albumin glomerular clearance 9 months 7.2 (±0.76) 14 months 5.7 (±0.23) P-values refer to the statistical difference between the groups at each age. NCR P-value 305 (±11.7) 540 (±23) 581 (±10) 0.098 0.001 0.001 2.1 (±0.05) 2.88 (±0.07) 2.9 (±0.13) 0.4 0.014 0.029 133.6 (±5) 126 (±4.5) 133 (±5) 0.001 0.001 0.002 347 (±89) 142 (±52) 0.18 0.006 6.8 (±0.38) 7.2 (±0.3) 0.84 0.06 1728 O. Bakoush et al. 0.008 Sieving coefficient of albumin 0.007 n=9 0.006 n=5 0.005 n=7 n= 6 0.004 0.003 Native albumin 0.002 Neutral albumin n=9 0.001 n=7 0 n=5 n=6 SHR 9 months NCR 9 months SHR 14 months NCR 14 months Fig. 2. Comparison between native and neutralized albumin sieving coefficient in animals studied (SHR and NCR) at 9 and 14 months of age. 1.6 (P=0.13, ns) (P<0.001) 0.0016 NCR 1.4 0.0014 SHR n=11 r = 0.86, p<0.001 n=2 SHR 14 months 0.0012 0.0010 0.0008 0.0006 SHR 9 months 0.0004 0.0002 0.003 0.004 0.005 0.006 0.007 0.008 GFR (ml/min/g kidney weight) Sieving coeficient of native albumin 0.0018 1.2 n=6 n=3 n=7 1 .8 n=5 .6 .4 0.009 Sieving coefficient of neutral albumin Fig. 3. Correlation between native (yalb) and neutral (yo-alb) albumin sieving coefficient in 9- and 14-month-old SHR. The equation for the regression line is: yalb ¼ 0.0003 þ 0.213  yo-alb. .2 0 Months 3 9 SHR in the 14-month-old SHR and no-to-mild fibrotic changes in the 9-month-old SHR and the 9- and 14month-old NCR (P<0.02). Partial glomerulosclerotic changes were found in 9 (±3)% of glomeruli in 14month-old SHR and in only 0.5% in 9-month-old SHR and in 9- and 14-month-old NCR (P<0.02). Large amounts of protein casts were seen in all 14month-old SHR, while there were mild to moderate amounts of casts in 9-month-old SHR and no or only mild amounts in 9- and 14-month-old NCR (P<0.02). Discussion The major result of this study is that the glomerular sieving coefficients (y) for both native and neutral albumin in SHR were normal during the first 9 months 14 3 9 14 NCR Fig. 4. Values of GFR (ml/min/kidney weight) in animals studied (SHR and NCR) at 3, 9 and 14 months of age. of hypertension, but significantly increased in old animals, as compared with age-matched NCR. Thus, in SHR, it takes more than half a rat lifetime of hypertension to develop proteinuria and kidney damage. Furthermore, the glomerular disturbance developed during long-standing hypertension is of size-selective nature and not represented by a primary chargeselective defect of the GCW. The tissue uptake method described here confirms and extends previous measurements of glomerular albumin permselectivity performed by micropuncture techniques in Wistar rats at 3 months of age. The albumin sieving coefficient in 3-month-old NCR assessed using the present technique (0.0007) was Albumin glomerular permeability in hypertensive rats almost identical to that assessed by a careful fractional micropuncture technique (0.00062) [5]. Also, our value of the sieving coefficient of neutral albumin in NCR (0.005) is remarkably close to that measured for Ficoll 37 Å reported by Oliver et al. [9] in intact MunichWistar rats (0.006), but it is nearly one order of magnitude lower than values obtained in (cooled) isolated kidneys perfused with red cell free perfusates having low oncotic pressure (0.08) [10]. The albumin sieving coefficient obtained by our tissue uptake or earlier micropuncture studies, along with a recent report by Norden et al. [11] of the absence of a significant amount of albumin degradation products in normal urine, contradict findings that the glomerulus is normally highly leaky to albumin, and reconfirms the old concept of the GCW as a highly size-selective barrier to the passage of macromolecules of the size of albumin or larger [12]. All the animals studied (NCR and SHR) showed a generally higher clearance of neutral albumin (6–7-fold) than of native (negatively charged) albumin, indicating a normally marked influence of charge on transglomerular protein transport (Figure 2). The data in the present study, along with previous and recent investigations comparing molecules of similar size but different charge [10,13], thus, strongly contradict recent reports of the absence of charge-selectivity of the GCW [12]. Since native (negative) albumin has a molecular radius of 36 Å, it should, unlike neutral albumin, be incapable of passing through the negatively charged small pores of the GCW and would, thus, be confined to the large-pore pathway (r  100 Å) to reach the Bowman’s space [2,3]. Theoretically, an increased number of large pores will cause an increase in the transport of both native and neutral albumin, even if the charge-selectivity of the GCW is maintained intact [2,3] (Appendix). However, if the charge-selectivity of the GCW were selectively impaired, then the sieving coefficient of the negatively charged native albumin would have approached that of neutral albumin, so that eventually the yo-alb/yalb ratio would have approached unity. This was not the case in the old SHR of this study. Instead, the ratio between neutral and negative albumin was maintained high throughout the lifespan of the rats (Table 1). This can be explained by an increase in the number of both small and large pores or, more likely, by the creation of an increased number of rather unselective pores of ‘intermediate’ radius (64.2 Å) (Appendix, Figures 2 and 3). This was predicted to occur in the total absence of any changes in the charge selective properties of the GCW. The present results in old SHR are in essential agreement with a previous study on uninephrectomized fawn-hooded rats displaying glomerular hypertension with glomerulosclerosis early in life [14]. However, although the large-pore fraction of the GFR increased very markedly (one order of magnitude) in these rats, this marked reduction in size-selectivity could not be picked up by Ficoll molecules of sizes comparable to that of albumin (36 Å) in the experiment, because Ficoll (36 Å) showed a high value of y (
0.09). The authors, 1729 therefore, preferred to interpret the marked increase in urinary albumin excretion in uninephrectomized fawnhooded rats to be mainly due to a charge-selective defect in the GCW. However, the data interpretation is indeed cumbersome, since the Ficoll sieving curves assessed were actually not very different from those determined earlier using neutral dextrans, which are flexible, linear sugar polymers showing abnormally high membrane permeability. Since the uncharged probe in the present study had a low fractional clearance, comparable to that measured for 37 Å Ficoll in normal rats [9], theoretically it would be much more suitable for assessing changes in glomerular size-selectivity than the non-protein molecular probes in fawn-hooded rats. Although a number of previous studies of albumin fractional clearance (y) have been performed in SHR using micropuncture studies, these studies have been confined to superficial nephrons. Using the present tissue uptake technique, it was possible, however, to obtain an average y for all nephrons, i.e. superficial and deep ones, the latter not normally being accessible by micropuncture studies. Indeed, the deep nephrons, rather than the superficial ones, are the likely source of the initial proteinuria in hypertension [15]. This may, at least partially, explain the higher albumin sieving coefficient obtained in the present study in 14-monthold SHR (0.0013) than that measured previously by micropuncture techniques (0.00029) [15]. As reported elsewhere, young SHR compared with age-matched NCR apparently show a lower glomerular ultrafiltration coefficient (Kf) and lower plasma flow, accounting for a lower GFR, than NCR [16]. The lower Kf (despite a lower GFR) will, according to the twopore model, result in a lower albumin sieving coefficient in SHR than NCR, which was actually seen in this study (0.0005 compared with 0.0007; P<0.05). In young SHR, increased vascular resistance and vasoconstriction of the afferent arteriole actually seems to protect the glomeruli from the transmission of a high systemic blood pressure to the glomerular capillary level [17]. The increased albumin glomerular permeability in old SHR was associated with significantly increased urinary albumin excretion, along with the appearance of severe nephrosclerotic changes on histological examination of kidneys of 14-month-old SHR. Thus, during long-standing hypertension, reduced endogenous activity of the renin–angiotensin system and a relative nitric oxide deficiency might have led to the transmission of the elevated systemic pressure to the glomeruli, inducing glomerular hypertension and an accelerated renal injury [18]. The renal vascular changes might have caused glomerular ischaemia, glomerular structural changes and a size-selective dysfunction of the GCW [1,18]. The present tissue uptake technique has been validated for small proteins in previous publications [7]. For proteins with low renal clearance, such as albumin, it is crucial that the kidneys are completely washed free of intravascular tracer and that the bulk of interstitially accumulated tracer (and free iodine) is, to a large 1730 extent, cleared by back-diffusion to the rinse fluid during washout. The washout procedure is, thus, crucial to the success of the technique. Even though we consider the washout to have been more or less complete, we cannot completely rule out that some tracer remained either intravascularly or extracellularly after tracer washout. From that point of view, the present y for native albumin may represent an overestimate. However, the value obtained is in excellent agreement with the recent careful micropuncture study by Tojo and Endou [5] referred to above and, therefore, we consider this overestimate to be moderate. Previously, Bertolatus et al. [19] found no evidence of plasma protein binding of labelled albumin. Also, renal deiodinase did not significantly affect the iodinelabelled proteins. Thus, substantial tissue binding or tracer deiodination is unlikely in the present study. Furthermore, with a dual isotope technique, in which one albumin tracer was used as an extravascular marker and the other as an intravascular one, we obtained essentially the same y as in the present study (unpublished data). In conclusion, this study indicates that, in untreated long-standing hypertension in rats, proteinuria results from a dysfunction of the GCW, affecting primarily its size-selectivity. This conceivably occurs by the appearance of an increased number of more unselective pores in the glomerular filtration barrier. Acknowledgements. The authors wish to thank Anna Rippe and Nermina Jagansac for their technical assistance and Prof. Steen Olsen, MD, PhD, Department of Pathology, University of Copenhagen, Denmark, for help with examining kidney biopsies. Grants from Riksförbundet för Njursjuka, Swedish Society for Medical Research and the Swedish Medical Research Council (08285) supported this study. Conflict of interest statement. None declared. References 1. Luke RG. Hypertensive nephrosclerosis: pathogenesis and prevalence. Essential hypertension is an important cause of end-stage renal disease. Nephrol Dial Transplant 1999; 14: 2271–2278 2. Tencer J, Frick IM, Oquist BW, Alm P, Rippe B. Sizeselectivity of the glomerular barrier to high molecular weight proteins: upper size limitations of shunt pathways. Kidney Int 1998; 53: 709–715 3. Lund U, Rippe A, Venturoli D, Tenstad O, Grubb A, Rippe B. Glomerular filtration rate dependence of sieving of albumin and some neutral proteins in rat kidneys. Am J Physiol Renal Physiol 2003; 284: F1226–F1234 4. Bakoush O, Torffvit O, Rippe B, Tencer J. High proteinuria selectivity index based upon IgM is a strong predictor of poor renal survival in glomerular diseases. Nephrol Dial Transplant 2001; 16: 1357–1363 5. Tojo A, Endou H. Intrarenal handling of proteins in rats using fractional micropuncture technique. Am J Physiol 1992; 263: F601–F606 6. Hoare DG, Koshland DE,Jr. A method for the quantitative modification and estimation of carboxylic acid groups in proteins. J Biol Chem 1967; 242: 2447–2453 O. Bakoush et al. 7. Tenstad O, Roald AB, Grubb A, Aukland K. Renal handling of radiolabelled human cystatin C in the rat. Scand J Clin Lab Invest 1996; 56: 409–414 8. Deen WM, Lazzara MJ, Myers BD. Structural determinants of glomerular permeability. Am J Physiol Renal Physiol 2001; 281: F579–F596 9. Oliver JD,III, Anderson S, Troy JL, Brenner BM, Deen WH. Determination of glomerular size-selectivity in the normal rat with Ficoll. J Am Soc Nephrol 1992; 3: 214–228 10. Hjalmarsson C, Ohlson M, Haraldsson B. Puromycin aminonucleoside damages the glomerular size barrier with minimal effects on charge density. Am J Physiol Renal Physiol 2001; 281: F503–F512 11. Norden AG, Sharratt P, Cutillas PR, O’Driscoll E, Gardner SC, Unwin RJ. Quantitative amino acid and proteomic analysis of normal and fanconi urine: no evidence for significant excretion of plasma protein fragments in normal urine. J Am Soc Nephrol 2003; 14 (Abstracts): 284A 12. Russo LM, Bakris GL, Comper WD. Renal handling of albumin: a critical review of basic concepts and perspective. Am J Kidney Dis 2002; 39: 899–919 13. Rennke HG, Patel Y, Venkatachalam MA. Glomerular filtration of proteins: clearance of anionic, neutral, and cationic horseradish peroxidase in the rat. Kidney Int 1978; 13: 278–288 14. Oliver JD,III, Simons JL, Troy JL, Provoost AP, Brenner BM, Deen WM. Proteinuria and impaired glomerular permselectivity in uninephrectomized fawn-hooded rats. Am J Physiol 1994; 267: F917–F925 15. Feld LG, van Liew JB, Galaske RG, Boylan JW. Selectivity of renal injury and proteinuria in the spontaneously hypertensive rat. Kidney Int 1977; 12: 332–343 16. Dilley JR, Stier CT,Jr, Arendshorst WJ. Abnormalities in glomerular function in rats developing spontaneous hypertension. Am J Physiol 1984; 246: F12–F20 17. Kett MM, Bergstrom G, Alcorn D, Bertram JF, Anderson WP. Renal vascular resistance properties and glomerular protection in early established SHR hypertension. J Hypertens 2001; 19: 1505–1512 18. Komatsu K, Frohlich ED, Ono H, Ono Y, Numabe A, Willis GW. Glomerular dynamics and morphology of aged spontaneously hypertensive rats. Effects of angiotensin-converting enzyme inhibition. Hypertension 1995; 25: 207–213 19. Bertolatus JA, Hunsicker LG. Glomerular sieving of anionic and neutral bovine albumins in proteinuric rats. Kidney Int 1985; 28: 467–476 20. Mason JC, Wendt RP, Bresler EH. Similarity relations (dimensional analysis) for membrane transport. J Memb Sci 1980; 6: 283–298 Received for publication: 24.11.03 Accepted in revised form: 12.03.04 Appendix According to the two-pore model, the glomerular clearance of any protein larger than the assumed small-pore radius is determined by convective transport across large pores. Thus, their large-pore clearances (ClL) are just products of the large-pore volume flow (JvL) and the reflection coefficients (sL), or actually (1–sL), of the solutes: ClL ¼ JvL ð1  sL Þ A1 Since the transport of native (negative) albumin through the glomerular filter is solely dependent upon Albumin glomerular permeability in hypertensive rats 1731 convective transport through large pores, its sieving coefficient (yalb) is: yalb ¼ JvL ð1  sL Þ GFR A2 In a previous study, we determined the large-pore radius to be
100 Å [2]. According to this assumption, 1  sL can be determined from pore theory (equation 3 in [2]) and from the Debye–Hückel theory of ion–ion interactions, predicting the pore radius to decrease by 8 Å (to 92 Å) and the solute radius to increase by 8 Å (to 44 Å) to account for the negative charge of native albumin and the GCW [20]. Hence, from yalb, JvL [or (JvL/GFR)] can be calculated at any known level of GFR measured. Charge-selectivity defect If charge-selectivity is lost, then native albumin, which is normally excluded from the small-pore pathway, will be able to penetrate the GCW also through the small pores. This leads to a selective increment in yalb, so that it will eventually approach yo-alb, the latter staying unaltered. This pattern was not seen in the present study, in which yalb and yo-alb were altered in a parallel fashion in old SHR. Size-selectivity defect If more of the enlarged, less selective (negatively charged) pores are formed when permeability is increased, then charge- and size-selectivity will be changed in parallel, leading to increases in both yalb and yo-alb, which was noted in the old SHR in the present study. The apparent reduction of charge-selectivity occurring when more large pores are formed is due to the fact that any increases in large-pore number will affect yalb relatively much more than yo-alb. If yalb and yo-alb during control conditions are denoted yalb,0 and yo-alb,0, respectively, then an increased (convective) transport through large pores (
JvL) in old SHR will change yalb and yoalb as follows: yalb ¼ yalb,0 þ
JvL ð1  sL Þalb GFR yo-alb ¼ yo-alb,0 þ
JvL ð1  sL Þo-alb GFR A3 A4 Rearranging equations A3 and A4 and dividing equation A3 by equation A4 and again rearranging, yields: yalb ¼ yalb,0 þ ð1  sL Þalb ðyo alb  yo-alb,0 Þ ð1  sL Þo-alb - A5 As seen from equation A5, the expression (1  sL)alb/ (1  sL)o-alb represents the regression coefficient of the (linear) correlation between yalb and yo-alb when solute (albumin) transport through large pores increases. If the large-pore radius is set at 100 (Å), the regression coefficient is calculated to be 0.670. Note that the reflection coefficients have been calculated assuming charge interactions to occur for native albumin according to the Debye–Hückel theory. Thus, increasing JvL will, due to the presence of negative charge in the large pores, cause smaller increases in yalb, in absolute terms, than in yo-alb. This is why the regression coefficient is smaller than one (0.67). In the present study, the experimentally determined regression coefficient between yalb and yo-alb was, however, much lower than 0.67, namely 0.213±0.034. This indicates that the loss of size-selectivity in old SHR cannot be unambiguously described according to the two-pore theory by just assuming the creation of additional large pores of radius 100 Å. The present data are either compatible with an increased large-pore transport of native albumin occurring in parallel with a slight increment in small-pore (and large-pore) transport of neutral albumin or, more likely, with the creation of additional unselective pores of radius <100 Å, namely 64.2 Å. This pore radius exactly matches the regression coefficient of 0.213.