The glomerular endothelial cell coat is essential for glomerular filtration

original article

http://www.kidney-international.org & 2011 International Society of Nephrology

The glomerular endothelial cell coat is essential for glomerular filtration Vincent Fride´n1, Eystein Oveland2, Olav Tenstad2, Kerstin Ebefors1, Jenny Nystro¨m1, Ulf A. Nilsson1 and Bo¨rje Haraldsson1 1

Department of Molecular and Clinical Medicine, Sahlgrenska Academy, Institute of Medicine, University of Gothenburg, Gothenburg, Sweden and 2Department of Biomedicine, University of Bergen, Bergen, Norway

The endothelial cell surface layer (ESL) is believed to contribute to the glomerular barrier, and the nature of its molecular structure is still largely unknown. The ESL consists of the membrane-bound glycocalyx and the loosely attached endothelial cell coat (ECC). A brief injection of hypertonic sodium chloride into the left renal artery was used to displace, elute, and collect non-covalently bound components of the renal ESL in rats. This procedure increased the fractional clearance of albumin 12-fold without detectable morphological changes as assessed by electron microscopy compared with the control group injected with isotonic saline. Mathematical modeling suggested a reduced glomerular charge density. Mass spectrometry of the renal eluate identified 17 non-covalently bound proteins normally present in the ECC. One of these proteins, orosomucoid, has previously been shown to be important for capillary permselectivity. Another protein, lumican, is expressed by glomerular endothelial cells and likely contributes to maintaining an intact barrier. Thus, the absence of one or more of these proteins causes proteinuria and illustrates the importance of the ECC in glomerular permselectivity. Kidney International (2011) 79, 1322–1330; doi:10.1038/ki.2011.58; published online 16 March 2011 KEYWORDS: glomerular endothelial cells; glomerular filtration barrier; proteinuria

Correspondence: Vincent Fride´n, Wallenberg Laboratory, Bruna Stra˚ket 16, Sahlgrenska University Hospital, SE-413 45, Gothenburg, Sweden. E-mail: [email protected] Received 11 October 2010; revised 14 January 2011; accepted 25 January 2011; published online 16 March 2011 1322

Accumulation of proteins in the urine is a key feature of renal disease and a consequence of an impaired glomerular filtration barrier. The glomerular filtration barrier is a biological membrane that includes a unique type of the fenestrated endothelium mainly without diaphragms covered with a glycoprotein surface layer, a glomerular basement membrane (GBM), and podocytes with foot processes bridged by slit diaphragms.1 The permselectivity of this barrier has been considered to reside principally in the GBM and in the podocyte slit diaphragm, whereas fenestrated endothelial cells and the endothelial cell surface layer (ESL) have attracted lesser attention. Nevertheless, the classical view of the GBM as a major contributor to the permselectivity of the glomerular barrier has been challenged,2–5 and a decade of extensive research regarding the podocytes has given important information about numerous proteins of the slit diaphragm that can cause proteinuria when defective.6–8 Simultaneously, experimental evidence has accumulated for the presence of a permselective cell surface layer covering the luminal face of glomerular endothelial cells.9–13 In the end of the 1980s, perfusion of the rat kidneys with cationic, anionic, or neutral ferritin demonstrated that cationic ferritin bound to, and visualized, a protein coat that covered the glomerular endothelium.14–16 Rostgaard and Qvortrup,17,18 using a different method of vascular perfusion fixation, found evidence for an extensive protein coat on the surface and in the fenestrae of the glomerular endothelium. Furthermore, we revealed a 200-nm-thick ESL stretching into the capillary lumen.19 In mice treated with adriamycin, the thickness of this layer was markedly reduced, an effect coinciding with the development of proteinuria and impaired synthesis of certain components of the ESL.20 Another study in mice demonstrated that infusion of enzymes that specifically degrade glycosaminoglycans reduced the thickness of the ESL and increased fractional clearance for albumin.21 Others have performed in vitro studies of the passage of albumin across monolayers of human glomerular endothelial cells and have observed an increased albumin flux after treatment with human heparanase.22 There is also clinical evidence that suggests that damage to the glomerular endothelium can cause proteinuria in pre-eclampsia and hemolytic uremic syndrome.23,24 Kidney International (2011) 79, 1322–1330

original article

V Fride´n et al.: Proteins in the glomerular ESL

Morphological examination with electron microscopy of kidney sections obtained from animals perfused with 0.15 mol/l NaCl (normal salt, NS) and 1 mol/l NaCl (HS) confirmed that the overall architecture of the glomerular filtration barrier remained morphologically normal during treatment (Figure 1). In addition, to quantitatively assess the condition of the filtration barrier, GBM thickness, podocyte foot process width, and width of filtration slits were measured in glomerular capillaries from animals perfused with NS and HS. The morphometric measurements revealed small but statistically significant differences between the two experimental groups (Table 1).

CL

CL EC

EC

When measured 10 and 20 min after HS perfusion, the glomerular filtration rate (GFR) was significantly reduced by 34±8.7% and 23±7.7% (P o0.001, P o0.01, n ¼ 10), respectively, relative to the pre-perfusion mean. GFR then started to recover toward the pre-perfusion level. No changes were seen in the NS group; for further details, see Figure 2. The absolute mean values of pre-perfusion points (20 and 10 min) were 1.21±0.089 and 1.19±0.073 for the NS group and 0.96±0.058 and 1.05±0.053 for the HS group (ml/min per g kidney wet weight). Functional effects on the glomerular filtration barrier

In rats perfused with HS, the fractional clearance for albumin increased B12-fold relative to both baseline and NS when measured 10 min after perfusion (Figure 3). This was followed by a slow but significant recovery over the subsequent 90 min (Po0.05, log-linear regression analysis), but the fractional albumin clearance remained significantly higher in the HS Table 1 | Summary of morphometric measurements of various structures in the glomerular filtration barrier

Glomerular basement membrane thickness (nm) Podocyte foot process width (nm) Podocyte filtration slit width (nm) Endothelial cell surface layer thickness (nm)

NS

HS

168.57±1.31

174.14±1.15**

301.49±7.54 40.63±0.37 123.42±8.06

307.38±7.58 38.23±0.41*** 123.81±8.49

Abbreviations: HS, high salt; NS, normal salt. Data are means±s.e.m. **Po0.01, ***Po0.001, n=52 glomeruli from 4 NS animals, n=60 glomeruli from 4 HS animals.

120 100

**

80

***

60 Perfusion

RESULTS Morphological evaluation of the glomerular filtration barrier

Effects on GFR

GFR in percentage of pre-perfusion

The ESL has two components. First, there is the glycocalyx, which is composed of membrane-bound proteoglycans (PGs). Second, attached to the glycocalyx, there is the endothelial cell coat (ECC), composed of secreted PGs, glycosaminoglycans, glycoproteins, and plasma proteins.1,25,26 The abundant, negatively charged PGs and glycosaminoglycans are believed to capture circulating plasma proteins and to produce a thick mesh with anionic properties in a gel-like structure.27,28 In previous studies, we have shown that the plasma protein orosomucoid is required for maintaining normal glomerular permeability in rats.29,30 The aim of this study was to provide more information on the ECC, including its permselective properties and its composition. Our approach was to establish a mild in vivo method for displacing non-covalently bound ECC components and to subsequently collect them for further characterization. A brief intrarenal high-salt (HS) perfusion through the renal artery was performed to displace anionic noncovalently bound components according to the principles of ion-exchange chromatography. A similar principle was previously used in an isolated perfused rat kidney system using low- and high-ionic strength in which we could modulate fractional clearances of albumin in a very specific and reversible manner.27,28

40 20

NS

HS

80

100

0 BM

FP US

–20

BM

US

10

20 40 Time (min)

60

FP

Figure 1 | Electron micrographs illustrating the glomerular filtration barrier after perfusion with salt solution. The rat kidneys were perfused with 0.15 mol/l NaCl (normal salt, NS) in (a) control animals and with 1 mol/l NaCl (high salt, HS) in (b) treated animals. Micrographs demonstrate a normal morphological appearance of the glomerular filtration barrier in both groups. Original magnification  8000. BM, basement membrane; CL, capillary lumen; EC, endothelial cell; FP, foot process; HS, high salt; NS, normal salt; US, urinary space. Kidney International (2011) 79, 1322–1330

–10

Figure 2 | Alterations of the glomerular filtration rate (GFR). On the x axis, 20 and 10 are pre-perfusion points, the time registration starts at perfusion and continues for 100 min. The GFR was significantly reduced in HS (high salt, 1 mol/l NaCl) animals, 10 and 20 min post perfusion relative to the pre-perfusion mean. Changes are expressed as percentages of the mean value of the two pre-perfusion points. Data represent means±s.e.m., n ¼ 10 in HS rats, except for 80 and 100 min (n ¼ 9) and in NS (normal salt, 0.15 mol/l NaCl) rats, n ¼ 10, except for 40 and 80 min (n ¼ 9) and 100 min (n ¼ 8). **P o0.01, ***P o0.001. 1323

0.005

***

0.004

NS

*** 0.003

*** **

0.002 0.001 0.000

HS

**

**

–20

–10

40 20 TIme (min)

10

60

80

100

Fractional clearance of Ficoll 35.5 Å

Figure 3 | Measurements of fractional albumin clearance in rats before and after salt perfusion. On the x axis, 20 and 10 are pre-perfusion points, and the time registration starts at perfusion and continues for 100 min. There is significant difference between HS (high salt, 1 mol/l NaCl) rats relative to baseline in all post-perfusion points. Data represent means±s.e.m., n ¼ 10 in NS (normal salt, 0.15 mol/l NaCl) rats, except for 100 min (n ¼ 7) and in HS rats, n ¼ 9, except for 80 and 100 min (n ¼ 8). **P o0.01, ***P o0.001.

0.20

NS

0.008

NS

***

***

0.004

0.002

0.000

–10

0.08

0

10

20

Time (min)

˚ Figure 5 | Analysis of the fractional clearance of Ficoll 55 A before and after salt perfusion in rats. On the x axis, 10 is a pre-perfusion point and the zero point is where perfusion occurs, followed by two post-perfusion points. Statistical analysis was made by comparing the post-perfusion fractional Ficoll clearance in HS (high salt, 1 mol/l NaCl) rats with the pre-perfusion value. Data represent means±s.e.m., n ¼ 8 in HS rats and n ¼ 8 in NS (normal salt, 0.15 mol/l NaCl) rats, except for 20 min (n ¼ 7). ***P o0.001.

Table 2 | Summary of mathematical modeling using the heterogeneous charged fiber matrix model Fiber radius (A˚) Fiber volume fraction (%) Large pore LpS (%) Fiber net charge (C/m2) Meas/calc albumin (%)

0.12

HS

0.006

HS

0.16

Normal salt

HS

P-value

5.00±0.00 5.21±0.08 0.020±0.003 0.375±0.03 25±11

5.00±0.00 4.91±0.14 0.100±0.038 0.166±0.02 49±13

NS NS o0.05 o0.001 NS

Abbreviations: C/m2, Coulomb/meter squared; HS, high salt; LpS, total hydraulic conductance; Meas/calc, measured/calculated; NS, not significant. Data are means ± s.e.m., n=8.

0.04 0.00

Fractional clearance of Ficoll 55 Å

V Fride´n et al.: Proteins in the glomerular ESL

Perfusion

Fractional clearance of albumin

original article

–10

0

10

20

Time (min)

˚ Figure 4 | Analysis of the fractional clearance of Ficoll 35.5 A before and after salt perfusion in rats. On the x axis, 10 is a pre-perfusion point and the zero point is where perfusion occurs, followed by two post-perfusion points. Data represent means±s.e.m., n ¼ 8 in HS (high salt, 1 mol/l NaCl) rats and n ¼ 8 in NS (normal salt, 0.15 mol/l NaCl) rats, except for 20 min (n ¼ 7).

Table 3 | Summary of mathematical modeling using the gel membrane model ˚) Small pore radius (A ˚) Large pore radius (A Large pore LpS (%) Charge density (mEq/l)

Normal salt

HS

P-value

49.7±0.36 79.2±2.48 0.50±0.11 86.8±3.06

50.2±1.25 81.8±7.33 1.35±0.78 64.8±3.61

NS NS o0.01 o0.001

Abbreviations: HS, high salt; LpS, total hydraulic conductance; NS, not significant. Data are means±s.e.m., n=8.

group throughout the 100-min observation period. The fractional albumin clearance remained normal in the NS group throughout the measurement period (n ¼ 7–10). The fractional clearance for Ficoll 35.5 A˚ (uncharged, size corresponding to albumin) was measured in HS and NS rats, and there were no significant alterations after salt perfusion (Figure 4). The fractional clearance for Ficoll 55 A˚ (a measure of the large pore fraction) was significantly increased 4-fold in HS-perfused rats 10 min after perfusion compared with the pre-perfusion value (P o0.001, n ¼ 8, Figure 5). Ficoll 55 A˚ clearance was unaffected in NS-perfused rats.

The most advanced model is the heterogeneous charged fiber model according to which there was a 56% reduction in fiber net charge (P o0.001, n ¼ 8) after a short perfusion with HS compared with NS (Table 2).1,21 The number of large pores increased five-fold (P o0.05, n ¼ 8), whereas the reduction in fiber volume fraction was non-significant. The gel membrane model showed that the kidneys of HS-perfused rats had an increase in large pores (P o0.01, n ¼ 8) and a reduced charge density (P o0.001, n ¼ 8); see Table 3.31

Results of mathematical modeling

Effects on the thickness of the ESL

Alterations to the barrier were analyzed using two separate models, and similar results were found with both models.

The thickness of ESL was estimated using Intralipid droplets as indirect markers (Figure 6).20 Measurement of the distance

1324

Kidney International (2011) 79, 1322–1330

original article

V Fride´n et al.: Proteins in the glomerular ESL

NS eluate

1 3

ILD

23 EC CL

40

US

39

FP BM 19 HS eluate

Figure 6 | A magnification of the glomerular capillary wall, indicating the procedure for estimations of various structures in the filtration barrier. The inset shows an overview of the glomerular capillary from which the magnification was taken. The white lines from left to right indicate the measuring position for podocyte filtration slit width, basement membrane thickness, podocyte foot process width, and thickness of the endothelial cell surface layer (ESL). The width of podocyte filtration slits was assessed by measuring the slit diaphragms interconnecting the foot processes. The thickness of the basement membrane was measured perpendicular to the podocyte slit diaphragm and the plasmalemma of adjacent endothelial cells. The width of the podocyte foot processes was measured linearly at the base of the foot processes, following interface with the basement membrane. The thickness of the ESL was determined by measuring the distance of the lipid droplet closest to and within 500 nm from the endothelial cell membrane. BM, basement membrane; CL, capillary lumen; EC, endothelial cell; FP, foot process; ILD, intralipid droplet; US, urinary space.

between lipid droplets and the endothelial cell membrane in rats perfused with HS and NS demonstrated that there was no significant difference of the thickness of the ESL between the groups (Table 1). Mass spectrometry analysis of renal eluates

To examine displaced proteins from the ESL, eluates from both groups were immunodepleted and analyzed with liquid chromatography-mass spectrometry using an LTQ-Orbitrap. Two independent runs, using different settings but the same original material identified in total 67 proteins in the NS eluate, 93 proteins in the HS eluate, and 93 proteins in the rat plasma. The total number of unique peptides was 1993, 1847, and 2933 in the HS eluate, NS eluate, and plasma, respectively. The overlap of identified proteins is shown in a Venn diagram (Figure 7). Peptide counting was used as a semi-quantitative method. Proteins were selected from the total number of identified proteins in Figure 7 according to the following inclusion criteria and are presented in Tables 4–6: proteins found in the HS eluate in both runs, in which the number of identified peptides assigned to the protein was at least 2 times higher compared with the NS eluate (Table 4) (8 proteins out of 63); proteins found in the Kidney International (2011) 79, 1322–1330

11

Plasma

Figure 7 | Venn diagram showing the distribution of the total number of identified proteins in the plasma, HS eluate, and NS eluate, using liquid chromatography-mass spectrometry. Merged plasma samples, merged HS (high salt, 1 mol/l NaCl) eluates, and merged NS (normal salt, 0.15 mol/l NaCl) eluates from three rats in each group were processed, run on an LTQ-Orbitrap, and protein identities were analyzed using ProteinCenter (Thermo Fisher Scientific, Waltham, MA).

HS eluate and the plasma in both runs, but not found in the NS eluate (Table 5) (3 proteins out of 11); and proteins found exclusively in the HS eluate and in both runs (Table 6) (6 proteins out of 19). The molecular weight range of proteins displayed in Tables 4–6 was 5–190 kDa. An absolute quantification was performed with unique peptides for lumican using selected reaction monitoring with respective AQUA peptides. The peptide ITNIPDEYFNR produced an HS/NS ratio of 2.3 based on peak area when compensating for the spike in the AQUA peptide (7278/1174781)/(2967/ 1089557), and the peptide FTGLQYLR produced an HS/NS ratio of 3.2 (15915/1356854)/(4610/1155172) (Figure 8). A third peptide (ILGPLSYSK) produced an HS/NS ratio of 3.4 without AQUA peptide compensation, resulting in an average HS/NS ratio for lumican of 3.0. Supplementary Figure S1 online illustrates the spectrum signatures of the AQUA peptides (A, B) and the quantification of the ILGPLSYSK peptide (C). Two other examples for selected reaction monitoring quantification without using AQUA peptides are shown for glutathione peroxidase 3 and gelsolin, which produced HS/NS ratios of 3.7 and 1.2, respectively (see Supplementary Table S1 online). Immunohistochemical analysis of lumican

To confirm that lumican is present in the glomerular capillary endothelium, we performed immunohistochemical analysis in cryosections of the human kidney tissue. For additional evidence about the localization, we performed colocalization studies with the endothelial-specific lectin rhodamine Ulex europaeus agglutinin I. There was a clear colocalization of lumican with this endothelial-specific marker in the glomerular capillary endothelium (Figure 9). 1325

original article

V Fride´n et al.: Proteins in the glomerular ESL

Table 4 | Proteins found in the HS eluate in two runs, in which the number of identified peptides assigned to the protein was at least two times higher compared with the NS eluate Name

Accession number

nP_HS

nP_NS

nP_plasma

nP_HS/np_NS

P51886 P04041 P62329 P63102 P10111 P23764 Q68FP1 P08649

11 4 3 5 7 22 12 8

3 1 1 2 3 3 4 4

— — — — — 21 17 62

3.7 3.7 2.8 2.5 2.3 7.3 3.0 2.0

Lumican Glutathione peroxidase 1 Thymosin b-4 14-3-3 protein zeta/d Peptidyl-prolyl cis-trans isomerase A Glutathione peroxidase 3 Gelsolin Complement C4

Abbreviations: HS, high salt; nP, number of unique peptides assigned to displayed proteins; NS, normal salt. Merged plasma, merged HS, and merged NS eluate from three rats processed and run on an LTQ-Orbitrap, followed by analysis by Proteome Discoverer (Thermo Fisher Scientific), ProteinCenter, and MS Excel 2007 (Microsoft AB, Kista, Sweden).

Table 5 | Proteins found in the HS eluate and the plasma in two runs, not found in the NS eluate Name a-1-Acid glycoprotein (orosomucoid) Complement component C9 Hepatic triacylglycerol lipase

Accession number

nP_HS

nP_plasma

P02764 Q62930 P07867

5 5 3

9 10 5

Table 6 | Proteins found in the HS eluate in two runs, not found in the NS eluate or the plasma Name

Abbreviations: HS, high salt; nP, number of unique peptides assigned to displayed proteins; NS, normal salt. Merged plasma, merged HS, and merged NS eluate from three rats processed and run on an LTQ-Orbitrap followed by analysis by Proteome Discoverer, ProteinCenter, and MS Excel 2007.

Peroxiredoxin-1 a-Actinin-1 Heat-shock cognate 71-kDa protein Profilin-1 Cytochrome b–c1 complex subunit 2 Annexin A5

Accession number

nP_HS

Q63716 Q9Z1P2 P63017 P62962 P32551 P14668

4 3 3 3 2 2

Abbreviations: HS, high salt; nP, number of unique peptides assigned to displayed proteins; NS, normal salt. Merged plasma, merged HS, and merged NS eluate from three rats processed and run on an LTQ-Orbitrap, followed by analysis by Proteome Discoverer, ProteinCenter, and MS Excel 2007.

DISCUSSION

The main contribution of this study is the identification of novel components of the ECC, which seem to be required for normal glomerular barrier function. The tool for acquiring the data was a new in vivo method, developed for studying the ECC without using any noxious chemical agents. The method, consisting of a short perfusion with 1 mol/l NaCl, uses the basic principles of ion-exchange chromatography. Non-covalently and loosely attached components of the ESL are displaced when the kidney is perfused briefly with the high-ionic strength NaCl solution. In this study, we focused on medium-sized plasma proteins and small PGs as the perfusion time was too short to allow for diffusion of large proteins or large PGs. HS perfusion gave rise to an increased fractional albumin clearance, indicating a reduced charge selectivity of the glomerular filtration barrier. Size selectivity also was affected to some extent. The proteinuria following HS perfusion was reversible, as seen by the 50% recovery of fractional albumin clearance in the first hour after perfusion, but the fractional albumin clearance did not return to the pre-perfusion level during the observation period. The observed recovery of the fractional albumin clearance in HS-perfused rats could depend on re-enrichment of circulating plasma proteins in the ESL when the ionic strength normalizes. Potentially re-enriched proteins in the ESL are those that we found in the HS eluate and plasma (Table 5), including the last three proteins in Table 4. One of these proteins, orosomucoid, has previously been shown to be essential for the maintenance of normal permeability in peripheral capillaries.29,30,32–34 1326

A reason for the fractional albumin clearance remaining high 1 h after HS injection could be that components lost from the ECC need to be synthesized and secreted by glomerular endothelial cells. The PG lumican may be one such component; it was identified together with four other proteins, in significantly higher concentration in the HS eluate compared with the NS eluate, but not present in the plasma (Table 4). Lumican was also immunolocalized and found to be colocalized with an endothelial-specific marker in the glomerular capillary endothelium of kidney sections (Figure 9). Lumican belongs to the family of small leucinerich repeat PGs that constitutes an important fraction of non-collagenous extracellular matrix proteins.35 The expression of lumican in the kidney has been observed by our group (unpublished array data) and reported elsewhere.36 We identified 11 peptides assigned to lumican in the HS eluate versus 3 in the NS eluate (HS/NS ratio of 3.7), with a ratio of 3.0 with the selected reaction monitoring assay. We also performed selected reaction monitoring on other identified proteins without applying AQUA peptides, and the HS/NS ratios were similar as with the Orbitrap analysis shown in Tables 4–6. Two examples are glutathione peroxidase 3 and gelsolin (Supplementary Table S1 online). This supports the semi-quantitative approach of counting peptides identified by the LTQ-Orbitrap and validates the peptide counting ratios for other proteins. Furthermore, we identified proteins that were found only in the HS eluate (Table 6). A review of a substantial number Kidney International (2011) 79, 1322–1330

original article

V Fride´n et al.: Proteins in the glomerular ESL

HS

NS 940.5

m/z

Counts (10×e3)

Counts (10×e3)

940.5

m/z

Area (940.5) = 2967

Area (940.5) = 7278 1.0

1.0 8.6 9.0 RT(min)

8.6 9.0 RT(min)

HS

NS

m/z

1.0

Area (749.5) = 15915

8.8 9.2 RT(min)

749.5

Counts (10×e4)

Counts (10×e4)

749.5

m/z

1.0

Area (749.5) = 4610

8.8 9.2 RT(min)

Figure 8 | Absolute quantification of lumican using selected reaction monitoring (SRM) assay. Two peptides unique for lumican were quantified in merged HS (high salt, 1 mol/l NaCl) and merged NS (normal salt, 0.15 mol/l NaCl) eluates using SRM assay designed on the basis of respective AQUA peptide standards. The unique SRM spectra are shown in the upper right corners of the figure and the corresponding SRM spectra from the AQUA peptides are shown in Supplementary Figure S1 online. The most abundant transition was used for quantification based on the peak area and the other transitions for peptide genuineness and assay verification. (a) The peptide ITNIPDEYFNR with m/z ¼ 691.4 (z ¼ 2) and its transition to m/z ¼ 940.5 (z ¼ 1) was used for quantification, and five other transitions (m/ z ¼ 720.4, 843.4, 1053.6, 1167.6, and 1268.7) were used for verification. (b) FTGLQYLR, with m/z ¼ 499.3 (z ¼ 2) to m/z ¼ 749.5 (z ¼ 1), was used for quantification, and three other transitions (m/z ¼ 579.4, 692.5, and 850.5) were used for verification. AQUA, absolute quantification; counts, peptide counts; HS, high salt; m/z, mass to charge; NS, normal salt; RT, retention time.

of comparative proteomic studies by Wang et al.37 revealed that proteins from 28 different protein families were frequently detected, regardless of species, in vivo or in vitro conditions, tissues and organs, and experimental objective. The authors suggested that cellular stress response can be the common reason why particular proteins are generally expressed. Three proteins (namely Peroxiredoxin-1, Heatshock cognate 71-kDa protein, and Annexin A5) from Table 6 belong to the group of frequently detected protein families. Kidney International (2011) 79, 1322–1330

When we performed the morphological evaluation, we found no ultrastructural abnormalities that would suggest a leaky barrier after HS perfusion. On the contrary, electron microscopy revealed a 3% increase in GBM thickness and a 5% reduction in podocyte filtration slit width and, if anything, these changes would be expected to make the barrier slightly tighter and reduce albumin clearance. Possibly, the HS solution induced a contraction of podocytes and/or affected the GBM per se. Furthermore, we looked at the glomerular ESL thickness with a method developed in our laboratory using electron microscopy and Intralipid droplets.19 The thickness of the ESL was expected to diminish after HS perfusion as seen after enzymatic digestions of the ESL in mice.21 However, no difference was seen between the two groups. A likely explanation is that enzymes used in previous studies digest the membrane-bound PGs that constitute the framework of the ESL, whereas salt perfusion rather elutes specific non-covalently bound molecules leaving a more intact ESL meshwork. In addition, the HS solution was injected as a short-lasting (15 s) 1-ml bolus compared with the much longer enzyme treatment. The total vascular surface area in the kidney is completely dominated by the capillaries, whereas the arteries and veins constitute only a minor fraction. Thus, proteins eluted in this study are considered to come from the ESL of the glomerular and/or the peritubular capillaries. The temporal coincidence of increased albumin leakage with the elution of certain proteins and the colocalization of lumican with the glomerular endothelium strongly support that we study effects specifically in the glomerular ESL. The mathematical modeling is consistent with the interpretation that injection of 1 ml of 1 mol/l NaCl eluted ‘charged fibers’ from the barrier. Thus, according to both the gel-membrane model and the more advanced charged fiber model, HS injection reduced the glomerular charge density while leaving the size selectivity largely unaffected, except for the largest solutes. The perfusion procedure used here is mild as revealed by the modest effects on GFR and the reversibility of albuminuria. The kidney wet weights were identical in both groups, and the transient reduction in GFR during the first 20 min after HS perfusion is most probably caused by vasoconstriction of afferent arterioles, due to HS concentration.38 One could argue that reduced GFR would elevate albumin clearance per se, as the lower rate of fluid transport across the barrier allows time for more albumin to diffuse into the urinary space for a given amount of fluid filtrated. However, this phenomenon has been studied for neutral Ficoll in the same size of albumin and the clearance increased B10% in the GFR interval of this study.39,40 For anionic albumin, this effect would be even smaller. HS perfusion increased albumin clearance by more than one order of magnitude; hence, the impact of transiently reduced GFR on albumin clearance is insignificant in this study. To conclude, using a new, ion-exchange-based in vivo technique for analyzing the composition of the capillary ESL and in particular its cell coat, we have, so far, identified 1327

original article

V Fride´n et al.: Proteins in the glomerular ESL

Lumican

Ulex europaeus

Merge

Lumican

Ulex europaeus

Merge

Figure 9 | Lumican is localized in the glomerular capillary endothelium and colocalizes with an endothelial-specific marker in the healthy human kidney tissue. (a) Confocal microscopic images of cryosections show the localization of lumican in the glomerular capillary endothelium (green) and the endothelial-specific lectin rhodamine Ulex europaeus agglutinin I (red). The colocalization of these proteins is shown in Merge (yellow). (b) Magnifications (  10) of one glomerular capillary loop from the image in panel a. (c) Omission of the primary antibody caused no staining. Bar ¼ 25 mm in panel a and bar ¼ 2.5 mm in panel b.

17 proteins from the ECC. Some of these, like lumican, are probably secreted by the endothelium, whereas others, such as orosomucoid, are plasma proteins. Indeed orosomucoid has previously been shown to be important for normal capillary permeability.29,30 Owing to the nature of the intervention, that is, perfusion of the entire renal vasculature with HS solution, at least some of the eluted proteins could originate from the peritubular capillary ESL, downstream of the glomerular capillaries, and it is presently not possible to pinpoint the origin of individual proteins. However, the much thicker glomerular ECC, which is notoriously easy to lose during fixation for histology, makes it likely that a substantial fraction of the eluted material is of glomerular origin. In addition, in vivo albumin clearance increased by one order of magnitude after injection of 1 ml HS solution, suggesting that one or more of the eluted ECC proteins are essential for preserving an intact glomerular barrier, and such proteins must reasonably be of glomerular origin. In any case, the results presented here 1328

clearly convey the message that even morphologically subtle alterations of the glomerular ESL can cause clinically significant proteinuria. MATERIALS AND METHODS General animal preparation Female Sprague–Dawley rats (Harlan, Horst, The Netherlands) kept on standard fodder and free access to water were used. Animal experiments were endorsed by the local ethics review committee. Anesthesia was induced and maintained by inhalation of isoflurane (1.5–3% v/v, Isobavet, Schering-Plough Animal Health, Buckinghamshire, UK) mixed with air (2.0 l/min) and oxygen (0.3 l/ min) in an isoflurane vaporizer (Ohmeda Isotec 5, Simtec Engineering, Askim, Sweden). Body temperature (37 1C) was preserved using a thermostatically controlled heating pad, together with a heating lamp. Experimental protocols Protocol 1: Measurement of GFR. The left internal jugular vein was used for administration of infusion solution with and without tracers. The left carotid artery served as a route for perfusion of the Kidney International (2011) 79, 1322–1330

original article

V Fride´n et al.: Proteins in the glomerular ESL

kidney, for collection of arterial blood, and for registration of arterial pressure using an electronic transducer (CODAN pvb Critical Care GmbH, Forstinning, Germany). Animals were eviscerated and the left ureter was cannulated for urine collections. Arterial pressure and urine weight were monitored using AcqKnowledge v3.7.3 (Biopac Systems, Goleta, CA) software. A priming dose of 0.055 MBq/ml 51Cr-EDTA and 700 mg/l FITC (fluorescein isothiocyanate)-Ficoll in the infusion solution was injected into the jugular vein, followed by a continuous infusion of 0.015 MBq/ml 51 Cr-EDTA and 340 mg/l FITC-Ficoll in the infusion solution at a rate of 150 ml/min throughout the experiment. Tracers were equilibrated for 15 min in the circulation, whereupon two sequential collections of arterial blood and urine in 10-min durations were performed for the estimation of GFR and fractional clearance for albumin and Ficoll. A catheter in the carotid artery was advanced through the abdominal aorta and fixed in the renal artery to restrict perfusion to include only the kidney. The renal vasculature system was perfused during 10–15 s with 1 ml of 1 mol/l NaCl (HS) or 0.15 mol/ NaCl (NS) in control animals, followed by a brief flush with 0.2 ml of 0.15 mol/l NaCl to remove all HS solutions. The catheter was withdrawn and blood flow to the kidney restored. Two sequential portions of blood and urine were collected in 5-min durations 10 and 20 min after perfusion, followed by 4 sequential collections in 5-min durations and at 20-min intervals. Protocol 2: Collection of the renal eluate for protein identification. The left carotid artery again served as a route for perfusion of the kidney and to maintain fluid balance. Animals were eviscerated, small capillaries extending the renal vein were ligated, and ligatures were prepared around both the renal artery and the renal vein for closure and fixation of a collecting catheter. The catheter in the carotid artery was then advanced through the abdominal aorta and the following steps were performed consecutively and within B1.5 min. The catheter in the abdominal aorta was advanced and fixed in the renal artery. The ligature around the renal vein in proximity of the vena cava was ligated after a catheter was positioned for collection of the renal vein eluate. The renal vasculature system was rinsed with 1 ml of 0.15 mol/l NaCl, perfused with 1 ml of 1 mol/l (HS group) or 0.15 mol/l (NS group) NaCl, and finally rinsed with 1 ml of 0.15 mol/l NaCl. The effluent was collected through the venous catheter as three separate eluates p1, p2, and p3. The hemoglobin concentration in each was measured with a blood gas analyzer Radiometer ABL 725 (Radiometer Medical Aps, Brønshøj, Denmark) to assess blood contamination. Eluates p2 from three animals in each group were pooled for further analysis with mass spectrometry. Protocol 3: Assessment of renal morphology. The experimental procedure was identical to protocol 1, with the exception that functional measurements were not performed in these animals. After perfusion of the kidney with HS or NS, 1.5 ml of intralipid solution was injected into the left jugular vein and allowed to equilibrate in the circulation for 10 min. The left renal artery and vein was clamped and the kidney was fixed by subcapsular injection of Karnovsky’s fixative and sliced as described earlier.19,41

Indirect immunofluorescence and confocal microscopy Cryosections of the nephrectomized healthy human kidney tissue were blocked with 2% fetal calf serum and 2% bovine serum albumin in phosphate-buffered saline. To detect lumican, we used an anti-lumican antibody (R&D Systems, Minneapolis, MN) at a 1:100 dilution. The secondary antibody was anti-goat Alexa Fluor 488 (Invitrogen, Carlsbad, CA) diluted 1:2000. For endothelial colocalization studies, we used the endothelial-specific lectin rhodamine Ulex europaeus agglutinin I (Vector Laboratories, Burlingame, CA) at a dilution of 1:2000. The sections were mounted using ProLong Gold antifade reagent (Invitrogen) and analyzed using a confocal microscope, that is, Leica TCS SP5 (Leica Microsystems GmbH, Wetzler, Germany). Statistics SPSS statistics software (SPSS, Chicago, IL) was used for all analyses. Mann–Whitney’s U-test was used to investigate statistical differences. P o0.05 was considered to indicate statistical significance. DISCLOSURE

All the authors declared no competing interests. ACKNOWLEDGMENTS

We thank Katarina Askerlund, Johnny Tran, Frode S Berven, and Magnus Arntzen for technical assistance. We have used the Proteomics Unit at the University of Bergen (PROBE), partly supported by the National Program for Research in Functional Genomics (FUGE) funded by the Norwegian Research Council. This study was supported by the Swedish Medical Research Council 9898 and 14764, the Inga-Britt and Arne Lundberg Research Foundation, the John and Brit Wennerstro¨m Research Foundation, and the Sahlgrenska University Hospital Grant LUA/ALF.

SUPPLEMENTARY MATERIAL Appendix 1. Residual materials and methods. Figure S1. SRM spectrum signatures pf AQUA peptides for lumican. Table S1. SRM quantification of glutathione peroxidase 3 and gelsolin. Supplementary material is linked to the online version of the paper at http://www.nature.com/ki REFERENCES 1. 2.

3.

4.

5.

Analysis of albumin, Ficoll, and 51Cr-EDTA in urine and plasma samples Albumin concentration was measured using a rat albumin enzymelinked immunosorbent assay according to the manufacturer’s protocol (Bethyl Laboratories, Montgomery, TX). Determination of FITC-Ficoll of different radii and g-radiation of 51Cr-EDTA were analyzed according to previous work.20 Kidney International (2011) 79, 1322–1330

6. 7.

8.

Haraldsson B, Nystrom J, Deen WM. Properties of the glomerular barrier and mechanisms of proteinuria. Physiol Rev 2008; 88: 451–487. Goldberg S, Harvey SJ, Cunningham J et al. Glomerular filtration is normal in the absence of both agrin and perlecan-heparan sulfate from the glomerular basement membrane. Nephrol Dial Transplant 2009; 24: 2044–2051. Harvey SJ, Jarad G, Cunningham J et al. Disruption of glomerular basement membrane charge through podocyte-specific mutation of agrin does not alter glomerular permselectivity. Am J Pathol 2007; 171: 139–152. Chen S, Wassenhove-McCarthy DJ, Yamaguchi Y et al. Loss of heparan sulfate glycosaminoglycan assembly in podocytes does not lead to proteinuria. Kidney Int 2008; 74: 289–299. van den Hoven MJ, Wijnhoven TJ, Li JP et al. Reduction of anionic sites in the glomerular basement membrane by heparanase does not lead to proteinuria. Kidney Int 2008; 73: 278–287. Roselli S, Gribouval O, Boute N et al. Podocin localizes in the kidney to the slit diaphragm area. Am J Pathol 2002; 160: 131–139. Ruotsalainen V, Ljungberg P, Wartiovaara J et al. Nephrin is specifically located at the slit diaphragm of glomerular podocytes. Proc Natl Acad Sci USA 1999; 96: 7962–7967. Tryggvason K, Patrakka J, Wartiovaara J. Hereditary proteinuria syndromes and mechanisms of proteinuria. N Engl J Med 2006; 354: 1387–1401.

1329

original article

9. Ciarimboli G, Hjalmarsson C, Bokenkamp A et al. Dynamic alterations of glomerular charge density in fixed rat kidneys suggest involvement of endothelial cell coat. Am J Physiol Renal Physiol 2003; 285: F722–F730. 10. Deen WM. What determines glomerular capillary permeability? J Clin Invest 2004; 114: 1412–1414. 11. Haraldsson B, Sorensson J. Why do we not all have proteinuria? An update of our current understanding of the glomerular barrier. News Physiol Sci 2004; 19: 7–10. 12. Ohlson M, Sorensson J, Haraldsson B. Glomerular size and charge selectivity in the rat as revealed by FITC-ficoll and albumin. Am J Physiol Renal Physiol 2000; 279: F84–F91. 13. Pries AR, Secomb TW, Gaehtgens P. The endothelial surface layer. Pflugers Arch 2000; 440: 653–666. 14. Avasthi PS, Koshy V. The anionic matrix at the rat glomerular endothelial surface. Anat Rec 1988; 220: 258–266. 15. Avasthi PS, Koshy V. Glomerular endothelial glycocalyx. Contrib Nephrol 1988; 68: 104–113. 16. Koshy V, Avasthi PS. The anionic sites at luminal surface of peritubular capillaries in rats. Kidney Int 1987; 31: 52–58. 17. Rostgaard J, Qvortrup K. Electron microscopic demonstrations of filamentous molecular sieve plugs in capillary fenestrae. Microvasc Res 1997; 53: 1–13. 18. Rostgaard J, Qvortrup K. Sieve plugs in fenestrae of glomerular capillaries–site of the filtration barrier? Cells Tissues Organs 2002; 170: 132–138. 19. Hjalmarsson C, Johansson BR, Haraldsson B. Electron microscopic evaluation of the endothelial surface layer of glomerular capillaries. Microvasc Res 2004; 67: 9–17. 20. Jeansson M, Bjorck K, Tenstad O et al. Adriamycin alters glomerular endothelium to induce proteinuria. J Am Soc Nephrol 2009; 20: 114–122. 21. Jeansson M, Haraldsson B. Glomerular size and charge selectivity in the mouse after exposure to glucosaminoglycan-degrading enzymes. J Am Soc Nephrol 2003; 14: 1756–1765. 22. Singh A, Satchell SC, Neal CR et al. Glomerular endothelial glycocalyx constitutes a barrier to protein permeability. J Am Soc Nephrol 2007; 18: 2885–2893. 23. Besbas N, Karpman D, Landau D et al. A classification of hemolytic uremic syndrome and thrombotic thrombocytopenic purpura and related disorders. Kidney Int 2006; 70: 423–431. 24. Karumanchi SA, Maynard SE, Stillman IE et al. Preeclampsia: a renal perspective. Kidney Int 2005; 67: 2101–2113. 25. Van Teeffelen JW, Brands J, Stroes ES et al. Endothelial glycocalyx: sweet shield of blood vessels. Trends Cardiovasc Med 2007; 17: 101–105.

1330

V Fride´n et al.: Proteins in the glomerular ESL

26. 27.

28.

29.

30.

31.

32.

33.

34.

35. 36.

37.

38. 39.

40.

41.

Weinbaum S, Tarbell JM, Damiano ER. The structure and function of the endothelial glycocalyx layer. Annu Rev Biomed Eng 2007; 9: 121–167. Sorensson J, Ohlson M, Haraldsson B. A quantitative analysis of the glomerular charge barrier in the rat. Am J Physiol Renal Physiol 2001; 280: F646–F656. Sorensson J, Ohlson M, Lindstrom K et al. Glomerular charge selectivity for horseradish peroxidase and albumin at low and normal ionic strengths. Acta Physiol Scand 1998; 163: 83–91. Haraldsson BS, Johnsson EK, Rippe B. Glomerular permselectivity is dependent on adequate serum concentrations of orosomucoid. Kidney Int 1992; 41: 310–316. Johnsson E, Haraldsson B. Addition of purified orosomucoid preserves the glomerular permeability for albumin in isolated perfused rat kidneys. Acta Physiol Scand 1993; 147: 1–8. Ohlson M, Sorensson J, Haraldsson B. A gel-membrane model of glomerular charge and size selectivity in series. Am J Physiol Renal Physiol 2001; 280: F396–F405. Haraldsson B, Rippe B. Orosomucoid as one of the serum components contributing to normal capillary permselectivity in rat skeletal muscle. Acta Physiol Scand 1987; 129: 127–135. Haraldsson B, Rippe B. Serum factors other than albumin are needed for the maintenance of normal capillary permselectivity in rat hindlimb muscle. Acta Physiol Scand 1985; 123: 427–436. Curry FE, Rutledge JC, Lenz JF. Modulation of microvessel wall charge by plasma glycoprotein orosomucoid. Am J Physiol 1989; 257: H1354–H1359. Nikitovic D, Katonis P, Tsatsakis A et al. Lumican, a small leucine-rich proteoglycan. IUBMB Life 2008; 60: 818–823. Schaefer L, Grone HJ, Raslik I et al. Small proteoglycans of normal adult human kidney: distinct expression patterns of decorin, biglycan, fibromodulin, and lumican. Kidney Int 2000; 58: 1557–1568. Wang P, Bouwman FG, Mariman EC. Generally detected proteins in comparative proteomics–a matter of cellular stress response? Proteomics 2009; 9: 2955–2966. Oberleithner H, Kusche-Vihrog K, Schillers H. Endothelial cells as vascular salt sensors. Kidney Int 2010; 77: 490–494. Ohlson M, Sorensson J, Lindstrom K et al. Effects of filtration rate on the glomerular barrier and clearance of four differently shaped molecules. Am J Physiol Renal Physiol 2001; 281: F103–F113. Rippe C, Asgeirsson D, Venturoli D et al. Effects of glomerular filtration rate on Ficoll sieving coefficients (theta) in rats. Kidney Int 2006; 69: 1326–1332. Jeansson M, Haraldsson B. Morphological and functional evidence for an important role of the endothelial cell glycocalyx in the glomerular barrier. Am J Physiol Renal Physiol 2006; 290: F111–F116.

Kidney International (2011) 79, 1322–1330