Adriamycin Alters Glomerular Endothelium to Induce Proteinuria

The pathophysiology underlying the nephrotic syndrome is becoming clear for several inherited podocytopathies; the mechanisms of injury that lead to the acquired forms of this disease are not well understood. We explored these mechanisms using the mouse model of adriamycin-induced proteinuria. We estimated the fractional clearances for FITC-Ficolls, albumin, and neutral albumin in cooled, isolated, perfused kidneys (cIPK) in situ. Treatment with adriamycin led to a significant increase in the fractional clearance of albumin and of Ficoll with radii larger than 20 A[Combining Ring Above]. Neutral albumin (33.4 A[Combining Ring Above]) and similarly sized Ficoll behaved similarly to each other. In addition, adriamycin led to a significant loss of charge density and size selectivity of the glomerular barrier. The thickness of the glomerular endothelial surface layer (i.e., or the glycocalyx) in adriamycin-treated animals was only 20% of that in normal animals. Finally, several proteoglycans were downregulated in isolated glomeruli. In summary, adriamycin thins the glomerular glycocalyx, perhaps by downregulating proteoglycan synthesis, and alters glomerular charge- and size selectivity. These data suggest that the glomerular endothelium may play a role in the pathogenesis of proteinuric renal diseases. Nephrotic syndromes are important and dramatic clinical conditions in which the underlying mechanisms are poorly understood and the treatment is unspecific and often ineffective.1 The evidence for a highly size- and charge-selective glomerular barrier is overwhelming,2 but there are alternative views3 as is elaborated in a recent review.1 The glomerular barrier consists of a fenestrated endothelium covered with a surface layer, the glomerular basement membrane (GBM), and podocyte foot processes.1 Recent advances concerning podocyte biology4,5 have revealed the molecular mechanism behind several genetic disorders,6 but less is known about the acquired diseases.7 Studies on laminin-deficient mice suggest that the changes in the GBM may precede those in the podocyte.8 Furthermore, the glomerular endothelial surface layer (ESL) may offer significant restriction to macromolecules.9,10,7,11–13 Thus, proteinuria is likely to occur after damage to any one of the components of the glomerular barrier.1 The study presented here was undertaken to further explore the intricate mechanisms underlying nephrotic syndrome. As an experimental model, we chose adriamycin, which induces proteinuria in mice.14–16 The glomerular damage was estimated using functional, morphologic, and molecular techniques. In vivo studies of proteinuria in patients or animals have the major disadvantage that proteins can be reabsorbed, secreted, or degraded in the tubules. Consequently, the protein concentration in final urine may differ considerably from that of primary urine. To overcome these limitations, one may use the inert polysaccharides dextran or Ficoll as markers instead of albumin. A few years ago, we published the first analysis of glomerular size and charge selectivity in mice using albumin and Ficoll.13 In the study presented here, we extend this analysis using charge-modified neutral albumin as well. Electron microscopy was used to measure the thickness of the glomerular ESL.11 Moreover, gene and protein expressions of several proteoglycans and glycosaminoglycan (GAG) synthesis enzymes were analyzed with techniques established in the lab.10,12,17 RESULTS Mice with the high dose of adriamycin (AD mice, 25 mg/kg) had a significant weight-loss from day 3 and onward whereas there was no difference between controls and low-dose adriamycin mice (ADL mice, 10 mg/kg) (Figure 1). The albumin/creatinine ratio measured in vivo at day 6 from spot urine was significantly increased from 0.10 (SEM + 0.022, −0.018) in controls to 0.67 (+0.272, −0.194) in AD mice (P < 0.001). There was no significant increase for ADL mice, 0.16 (+0.030, −0.025). Functional Properties of the Glomerular Barrier The fractional clearances for all Ficoll sizes, 12 to 70 A[Combining Ring Above], are shown in Figure 2. There was a significantly increased fractional clearance of Ficoll in AD mice compared with controls and ADL mice, at least for Stokes-Einstein radii (aSE) exceeding 20 A[Combining Ring Above] (P < 0.01, Figure 2). Because there was no difference between controls and ADL mice, further analyses were done for controls and AD mice only. The GFRs were similar for controls and AD mice as shown in Table 2. The fractional clearance for albumin (HSA) was significantly increased from 0.0042 (+0.0008, −0.0007) in controls to 0.0150 (+0.0047, −0.0036) in the AD group (P < 0.01, Figure 3). Neutral Ficoll of similar size as albumin (aSE 35.5 A[Combining Ring Above]) had a fractional clearance of 0.089 (±0.0042) in controls and was significantly increased to 0.107 (±0.010) in AD mice (P < 0.01, Figure 3). Neutral HSA (nHSA, aSE 33.4 A[Combining Ring Above]) had a fractional clearance of 0.11 (±0.007) and 0.10 (+0.011, −0.010) for controls and AD mice, respectively (Figure 3). The ratio between HSA and Ficoll 35.5 A[Combining Ring Above] was found to be significantly increased from 5.2% (+0.9, −0.8) in controls to 11.8% (+2.3, −2.2) in AD mice (P < 0.01). When comparing HSA with nHSA the results were similar: an increase from 4.1% (+0.8, −0.6) in controls to 18.4% (+6.7, −4.9) in AD mice (P < 0.001, Table 2). The ratio between nHSA and neutral Ficoll of similar size (aSE 33.4 A[Combining Ring Above], see Materials and Methods) was found to be 107% (+6.1, −5.8) in control animals; that is, there was no difference in clearance between protein and Ficoll polymer of similar charge and size. According to the gel membrane model, several parameters were altered in the AD mice. The most pronounced change was the significantly decreased charge density in the glomerular barrier from 46.9 (+2.9, −2.7) mEq/L in controls to 26.5 (+5.4, −4.4) mEq/L (P < 0.01) in AD mice. There were also significant changes in size selectivity; both the pore radii and the number of large pores were increased (P < 0.05, Table 3). These changes of the functional properties of the glomerular barrier were accompanied by a significant decrease in ESL thickness from 126 nm (+39, −30) in controls to 27 nm (+13, −9) in AD mice (20% of control, P < 0.001, Figure 4). Electron micrographs also showed podocyte foot-process flattening in more than 70% of glomeruli in AD mice (Figure 5). Expression of mRNA and Proteins Real-time PCR revealed a significant regulation in isolated glomeruli of several genes in AD mice compared with controls, Figure 6. The mRNA expression of the chondroitin sulfate proteoglycan versican was found to be significantly upregulated in AD mice (P < 0.001, Figure 6). Western blot for versican showed a loss of protein (Figure 7, see Discussion). In addition, a receptor for hyaluronan, RHAMM, and hyaluronidase was significantly downregulated in AD mice (P < 0.01 and P < 0.001, Figure 6). Furthermore, heparan sulfate proteoglycans were also altered. There were significant downregulations of glypican-1 and glypican-4 (P > 0.001), but not of perlecan, agrin, syndecan-1, and syndecan-4 (Figure 6 and 8). In addition, the EXT-1 and HS2ST1 enzymes, which are important for the elongation and sulfation of heparan sulfate chains, were found to be downregulated (P < 0.001). As for the small leucine-rich proteoglycans, decorin was upregulated in AD (P < 0.05), fibromodulin was downregulated (P < 0.05), and no change was found for biglycan. The expression of podocin was downregulated in AD (P < 0.01), whereas nephrin was unchanged (Figure 9). DISCUSSION In mice made proteinuric with adriamycin, we found the glomerular ESL thickness to be only 20% of that in normal animals. This alteration seems to have been caused by downregulation of the glomerular synthesis of certain proteoglycans and enzymes required for the production of GAGs. Consequently, the glomerular charge selectivity was diminished, causing albuminuria and flattening of the podocyte foot processes. There was also an effect on glomerular size selectivity as shown by the Ficoll sieving curves in Figure 2. Furthermore, the study gives important insights into the normal glomerular barrier, which demonstrate marked size- and charge selectivity in support of the classical concept.7 The so-called “albumin retrieval hypothesis”1,3 can therefore (once again) be rejected.1,2 Also, the study shows that proteins such as albumin behave as Ficoll polymers of similar size and charge. The ESL is an intricate structure to study because its highly hydrated morphology is altered by the tracers used for its visualization.11 The ESL includes glycocalyx components attached to the endothelial cell membrane (e.g., glypican and syndecan) together with secreted proteoglycans, GAGs, hyaluronan, and certain plasma proteins that are bound to the glycocalyx mainly by charge-charge interactions. Intravital microscopy with a fluorescence tracer has been used in several organs to identify and measure an “exclusion zone,”18 but the technique is less suitable for the kidney. We have developed a technique for the glomeruli using electron microscopy and Intralipid droplets (see Figure 5) that reflects alterations of the ESL thickness; for example, in response to enzyme treatment.11 In the study presented here, we found that adriamycin-treatment drastically reduced the thickness of ESL from 126 to 27 nm; that is, to 20% of normal (see Figure 4), a finding that coincides with the development of proteinuria. Adriamycin is known to cause proteinuria in rats19–21 and BALB/c mice.16 In the latter study, 10 mg/kg of adriamycin was found to produce marked proteinuria within 1 wk, which was maintained over a period of 6 wk. However, the mice lost one-third of their body weight within 1 wk and then gradually gained weight again.16 In this study, BALB/c mice could not be used because of unforeseen, not previously reported, vasoconstriction in response to artificial perfusion. We therefore studied C57BL/6 mice, which are more commonly used in experimental work, but could not find any signs of renal damage at a dose of 10 mg/kg of adriamycin. Indeed, most investigators consider C57BL/6 to be resistant to the proteinuric effect of adriamycin, but the reason for this difference between strains is not known. Our mice were found to be sensitive to higher doses; hence, dose-response experiments were done.16 We decided to use 25 mg/kg adriamycin together with twice daily intraperitoneal injections of glucose-electrolyte solution, which induced proteinuria in the mice, but not as severe weight losses as seen in other studies using BALB at lower doses.16,22 Similar doses (26 mg/kg) have previously been reported to be required in BALB/c mice (2 × 13 mg/kg) to achieve stable proteinuria.15 Bertolatus et al.19,20 studied the proteinuric effect of adriamycin in rats using a tissue uptake technique that more recently has been reintroduced by one of the coauthors (O.T.).23,24 In rats, adriamycin reduced charge selectivity as revealed by anionic an neutral IgG19 and native and neutral bovine albumin,20 with no detectable effect on the size barrier. Also, in our study the effect on size selectivity was minimal for aSE below 40 A[Combining Ring Above], but became evident for larger Ficolls, reflecting an increased number of large pores (see Figure 2). In that figure, the effect of a low dose (10 mg/kg) of adriamycin is also shown, and the Ficoll sieving curve is close to that of the control mice. Despite differences in experimental technique, the results obtained using tissue uptake in intact rats20 are remarkably close to the data obtained in the study presented here from cooled, isolated, perfused kidneys (cIPKs), albeit that the absolute values are higher using the latter approach. Indeed, the tissue uptake technique is likely to underestimate the “true” clearance values, whereas trauma is likely to slightly increase the clearance for the cIPK model. There are surprisingly few studies of glomerular sieving of neutral albumin, most likely because of the instability of the modified protein. Thus, the charged-modified albumin will slowly regain anionic net charge, making it necessary to characterize the protein in plasma and in urine. Moreover, charged-modified proteins may alter their molecular size, which also has to be monitored and reported. In the study presented here, we have made serious efforts to characterize the tracers in all of these respects. Thus, the charge-modified HSA gave rise to bands close to the isoelectric point of 7.4; that is, neutral, with an aSE of 33.4 A[Combining Ring Above] (i.e., 2 A[Combining Ring Above] less than native albumin). Comparing the nHSA with a Ficoll of similar size gave a fractional clearance ratio of 1.07 for the control mice. These findings are in agreement with studies on isolated GBMs,25 nanopore membranes26 [for solute-over-pore radius (λ) 0.0 to 0.7, but not for λ > 0.8], and other globular proteins,27 but in contrast to recent reports by Rippe et al. 23,28 Moreover, the clearance ratio of native over nHSA of 4% in the study presented here (Table 2) is close to the value of 2% reported by Bertolatus,20 but less than the value of 10% reported by Lund et al.23 These differences may be reconciled by differences in the biochemical properties of the modified protein. We therefore conclude that neutral globular proteins and Ficolls have similar behavior in their passage across the glomerular barrier. The diminished thickness of the ESL and the changes in functional properties was coupled to several changes in glomerular gene expression. The downregulation of the heparan sulfate proteoglycans glypican-1 and glypican-4 together with the downregulation of EXT-1 and HS2ST show that in addition to the loss of core protein, there is less elongation (EXT-1) and sulfation (HS2ST) of heparan sulfate chains. To our knowledge, there is no previous report of downregulation of glypican in proteinuria. Furthermore, our lab has shown a downregulation of the chondroitin sulfate proteoglycan versican in several models of renal disease. Nonobese diabetic mice showed a downregulation of versican,12 as did puromycin-induced nephrotic syndrome in rats and puromycin in cultured glomerular endothelial cells and podocytes.10,17 To our surprise, adriamycin-induced proteinuria showed an increase of versican mRNA in the glomerulus. However, when performing Western blotting it was evident that despite the increase in mRNA there was a loss of versican protein compared with controls. The real-time PCR detects versican isoform V0 and V1, whereas the antibody detects isoforms V0, V1, V2, and V3. Because we only detected a 370-kD band in our Western blotting, this that only versican V0 is in mice glomeruli. We that the difference between versican mRNA and protein in AD mice can be to loss of in the Thus, versican is and secreted into the but because of its size with and versican has a and a to the However, in AD mice the glycocalyx thickness is decreased to 20% of that in control mice, and there are less for there may be an increased loss of protein, for which the cell by the of versican (Figure is known that versican to hyaluronan, and and In this study, we found that and hyaluronidase were we have not measured hyaluronan, this could a loss of is also that versican to other proteoglycans such as that to be In addition, it be that glomerular cells other than the endothelium produce proteoglycans and their synthesis may be as well in The expression of major components of the and agrin, that the GBM is to be in adriamycin-induced proteinuria. However, electron microscopy revealed changes of the podocyte foot which may be a effect of adriamycin, a of or In summary, adriamycin was found to glomerular cell synthesis of certain proteoglycans and GAGs, causing a of the glomerular In addition, glomerular charge selectivity was reduced and the restriction of larger causing proteinuria. Adriamycin did not to the but there was flattening of the podocyte foot processes. We conclude that adriamycin-induced proteinuria may of the glomerular endothelial were in C57BL/6 mice Mice were on and had to The the Adriamycin or was in a in animals at day we to use BALB/c mice, in which 10 mg/kg of adriamycin induces proteinuria. However, the BALB/c mice to isolated of the kidneys with a vasoconstriction and C57BL/6 mice were but in these the dose of 10 mg/kg did not renal doses were × and the dose of 25 mg/kg was day 1 to the mice were 2 daily intraperitoneal injections of 2 of a glucose-electrolyte to weight loss to low The had the The body weight was measured day was and the glomerular barrier using the cIPK model. the cIPK model, glomeruli were isolated for analysis of mRNA and protein expression. Furthermore, mice were also used at day 6 for with electron microscopy of endothelial glycocalyx thickness on the of Intralipid at was induced and by of (2 to The mouse was and the A was in the for of urine. The and were to the renal and the was in a with a to a The was of the kidneys by of a at The was then to the renal and the was for to the renal the the a period of urine were and was not to the kidneys and to with or the The of the was at to as well as and urine weight changes were monitored by using was using a modified with the addition of HSA and Ficoll in the size 12 to 70 A[Combining Ring Above]. The had the 10 Ficoll 0.16 The was from and with in of HSA The nHSA was obtained by a of the using a modified from that by and as HSA was in of at A of of was then to the to the The was at by addition of were for and to 1 of at to the The were of 10 of The was and at The effect of the was by isoelectric using a and with an isoelectric point close to was obtained by a of the nHSA was in and with using to the The of nHSA was done as close to the experiments as and any nHSA not used within a was The radius of nHSA was measured in and after the experiments using a 12 to a cell A with addition of was used as and and were used as molecular weight and albumin showed a slightly aSE (33.4 A[Combining Ring Above]) than its anionic control A[Combining Ring Above]) obtained from the for and urine from cIPK experiments were analyzed for and in a and for and The concentration of HSA was with was from the urine over concentration for urine clearance for a is by its concentration for the using or the mice from the AD group showed no signs of renal disease clearance for albumin was lower than value in and were The fractional clearance for radii of was by and urine to gel and of fluorescence using A with was used as A of from each was analyzed at an of nm and an of The and the were maintained the analysis and were 2 and albumin was using for mouse albumin to the of urine were for the concentration measured with a ESL Intralipid droplets were used as markers of the thickness of the as the Intralipid was by the layer of the after a at of droplets was obtained after at for 10 and was in the at day 6 of controls or AD animals the were to in the for 10 the renal and was and the was by of and 2% in were in a electron Electron of glomerular at a of were obtained from controls and AD mice a of glomerular for each the between the and the endothelial cell membrane surface was measured to nm from the endothelium using In the for droplets was PCR and Western was from isolated glomeruli after the cIPK using the The concentration and of the was by the synthesis of the expression of each gene (Table was by real-time PCR on the using the The detects genes in one in controls and and was and for the genes in Table were from the with the controls and The of was used to for differences in gene expression using the for was using an from and after the using a Western blotting for versican was done using protein from isolated glomeruli with in 10 3 for at to glycosaminoglycan chains. were on 3 to and to were and was done using a antibody was used to bands with in a The The properties of the glomerular barrier were using the According to the model, the glomerular barrier is of in one charge and one size The gel is in with plasma and the concentration of anionic such as albumin. The of the barrier as membrane size but no charge selectivity. The of charge selectivity was done by comparing the fractional clearance of HSA with that of its neutral nHSA and 35.5 A[Combining Ring Above], an of the charge properties are using a model with experimental fractional clearance data for Ficolls to 70 A[Combining Ring The can be using the the small pore the large pore the large pore of the glomerular pore and the are presented as were for > 2 and values were used for with that is, fractional albumin/creatinine surface layer and mRNA were made to controls using A P value less than was for the body weight for controls ADL and AD 9). analysis was made by comparing the body each day with the body weight at day P < 0.05, P < 0.01, P < for the fractional clearance of Ficolls to Ring Above]) in controls and AD mice. There is a significant difference (P < 0.01) between controls and AD for all the only value is shown for for the fractional clearance for and Ficolls in controls and AD mice. the P < 0.05, P < for the ESL in controls and AD mice P < Electron micrographs of glomerular Intralipid droplets and the podocyte foot-process flattening in AD for mRNA expression to genes and controls for hyaluronidase perlecan, agrin, and protein in AD mice. the for controls and AD mice. P < 0.05, P < 0.01 P < Western blot for versican for mRNA expression to genes and controls for the heparan sulfate proteoglycans syndecan-1, and the for controls and AD mice. P < for mRNA expression to genes and controls for the small proteoglycans and shown are enzymes important for the elongation and sulfation of on heparan sulfate proteoglycans and Table 1 for the for both controls and AD mice. P < 0.05, P < 0.01, P < of the versican mRNA and protein expression. In the normal endothelial versican is and secreted into the Because of its size with and versican has a and a to the However, in AD mice the glycocalyx thickness is decreased to 20% of that in control mice and there is a loss of for The of this will be an increased loss of protein, for which the cell to by the of genes analyzed by real-time GFRs and fractional clearance of the results from the for The the and the and the for The the and and a this

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.