A novel in vivo method to quantify slit diaphragm protein abundance in murine proteinuric kidney disease

Injury of the glomerular filter causes proteinuria by disrupting the sensitive interplay of the glomerular protein network. To date, studies of the expression and trafficking of glomerular proteins have been mostly limited to in vitro or histologic studies. Here, we report a novel in vivo biotinylation assay that allows the quantification of surface expression of glomerular proteins in mice. Kidneys were perfused in situ with biotin before harvest. Afterwards glomeruli were isolated and lyzed. The protein of interest was separated by immunoprecipitation and the amount of surface-expressed protein was quantified by Western blot analysis with streptavidin staining. As proof-of-concept, we examined the presence of nephrin in the slit diaphragm in two well-established murine models of proteinuric kidney disease: nephrotoxic nephritis and adriamycin nephropathy. In proteinuric animals, significantly less nephrin was detected in the slit diaphragm. When proteinuria decreased once again during the course of disease, the amount of surface nephrin returned to the baseline. Our present results suggest that our assay is a valuable tool to study the glomerular filter in proteinuric kidney diseases. Note that the assay is not limited to proteins expressed in the slit diaphragm, and all surface proteins that are accessible to biotin perfusion and immunoprecipitation qualify for this analysis.


Introduction
Injury to a single layer of the glomerular filter is sufficient to induce albuminuria. If podocytes and the slit diaphragm are involved, albuminuria easily reaches nephrotic ranges. Albuminuria is a main driver of the progression of kidney disease [1] and is associated with a markedly increased cardiovascular risk [2]. Several studies have suggested that altered trafficking of slit diaphragm proteins causes filter leakage. We previously demonstrated that nephrin, the backbone of the slit diaphragm, undergoes β-arrestin2-dependent endocytosis in healthy and a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 diseased states [3,4]. Meanwhile the concept of nephrin endocytosis has been confirmed and extended by other groups (reviewed in [5][6][7]). Newly identified regulatory mechanisms of nephrin endocytosis involve planar cell polarity pathways, e.g. Wnt5a and Vangl2 [8,9]. CIN85/RukL controls nephrin turnover by ubiquitination [10,11]. Some studies demonstrated dynamin-and clathrin-dependent uptake whilst others showed that raft-dependent endocytosis might be an alternative pathway [4,12,13]. Recent work suggests that reduction of endocytotic processes is also detrimental to the podocyte [14].
Studies of podocyte protein trafficking have been hampered by the lack of a suitable in vivo model of the glomerular filter. To elucidate the complex interplay of the filter structures mainly two different techniques have been used so far. First, the classical immunofluorescence labeling approach using antibodies that recognize epitopes in the extracellular domains of the receptors being studied [15]. Second, covalent modification of cell surface receptors with biotin as a valuable technique for monitoring surface protein trafficking. Biotinylation enables labeling of the cell surface protein at a particular time point, but instead of using a specific antibody, all surface proteins are tagged with biotin. The protein of interest is then separated by immunoprecipitation followed by SDS-PAGE and Western blotting. The biotinylated fraction can be visualized by staining with streptavidin and quantified by densitometry. So far biotinylation has been limited to in vitro experiments. Satoh et al. recently described a method in which glomeruli are extracted first and are then biotinylated in vitro [16]. However, it is known that the complex architecture of the glomerular filter is altered by mechanical stress. We learned from our previous work that especially the molecular composition of the slit diaphragm changes rapidly. Based on these observations we developed an in vivo biotinylation assay because we were convinced that biotinylation of glomerular extracellular proteins in situ yields more accurate results.
Applying the assay in an acute and a chronic proteinuric disease model, we demonstrate that the amount of surface nephrin can be reliably quantified with the biotinylation assay introduced.

Animal care
Mice were obtained either from an in-house breed at a local animal care facility or from Janvier Labs, France. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the local Institutional Animal Care and Use Committee: Landesamt für Natur, Umwelt und Verbraucherschutz Nordrhein-Westfalen, Recklinghausen, Germany; Permit Number: AZ: 84-02.04.2012.A099. If necessary analgesia was performed with Carprofen 5 mg/kg s.c. Animals were sacrificed by decapitation under anesthesia performed with Ketamin 100mg/kg bodyweight and Xylazin 5mg/kg bodyweight.

Induction of nephrotoxic serum nephritis (NTN)
C57Bl/6 male mice (8 weeks of age) were injected intraperitoneally with 500-800 μL NTN serum. Control animals received an equal amount of normal saline. Following the collection of urine, mice were weighed and examined on day 1 or 18 after injection.

Induction of adriamycin-induced nephropathy (ADR)
BALB/c male mice (6 weeks of age) were injected intravenously with adriamycin (10 mg/kg bodyweight; ADR, Sigma-Aldrich, Steinheim, Germany), and control animals were administered an equivalent volume of water for injection (aqua ad injectabilia). Following the collection of urine, mice were weighed and examined 7 days after injection.

Analysis of proteinuria
Twelve hour urine collection was performed in single metabolic cages. Proteinuria was assessed using a 10% polyacrylamide gel. Urinary albumin and creatinine excretion were measured using standard laboratory protocols, and proteinuria was normalized based on creatinine excretion.

In vivo biotin labeling and immunoprecipitation (IP)
First, kidneys were perfused via the abdominal aorta with 5ml ice-cold PBS (rate 2mL/min) supplemented with 1 mM MgCl 2 and 0.1 mM CaCl 2 (PBSCM). Then, perfusion was repeated with 5 mL PBSCM (rate 2mL/min) supplemented with 0.5 mg/mL EZ-Link™ Sulfo-NHS-LC-Biotin (Thermo Scientific, Rockford, USA) for surface protein labeling. Afterwards, unspecific biotin binding was quenched with 5mL PBSCM (rate 2mL/min) supplemented with 100 mM glycine. Finally, kidneys were perfused with 5 mL PBSCM (rate 2mL/min) containing 16 x 10 6 Dynabeads/mL (Invitrogen, Oslo, Norway). After perfusion, kidneys were immediately minced and digested for 40 minutes at 37˚C with 1.5 mg/mL Collagenase A (Roche, Mannheim, Germany). Cell suspensions were filtered through 100 μm cell strainers (Greiner Bio-One, Frickenhausen, Germany), separated by centrifugation (5.000 ×g for 5 min at 4˚C), and washed with PBSCM using a Dynamag magnet. When a purity of >95% glomeruli was achieved, glomeruli were collected by centrifugation (6.800 ×g for 5 min at 4˚C). Samples were immediately homogenized using a TissueRuptor (Qiagen, Hombrechtikon, Switzerland), and were lysed for 30 min on ice. Insoluble cellular material was removed by centrifugation (15.000 ×g for 30 min at 4˚C). Protein concentrations from the supernatant were measured using a BCA Protein assay kit (Thermo Scientific, Rockford, USA), and were subsequently adjusted to ensure equal total protein content.

Quantification of podocyte loss
As described before staining of the CDK-inhibitor p57 was used to measure the number of healthy podocytes [18,19]. Two μm thick paraffin sections were deparaffinized and antigen retrieval was performed by boiling for 30 min in 10 mM citrate buffer pH 6.1 at constant 98˚C in a steam cooker. Unspecific binding was blocked (5% horse serum, 30 min RT). Rabbit anti-p57 (Santa Cruz, Santa Cruz, USA) diluted 1:400 in 5% horse serum was incubated o/n at 4˚C. Color development was performed with the ZytoChem-Plus AP Polymer-Kit (Zytomed, Berlin, Germany) according to the manufacturers' instructions with neufuchsin. Stainings were evaluated under an Axioskop and photographed with an Axiocam HRc using the Axiostar software (all Zeiss, Jena, Germany). Quantification of podocyte loss was performed using a serpentine movement from cortex to medulla and vice versa. The outlines and p57-positive nuclei of 40 consecutively encountered capillary tufts were traced and counted manually at a 400-fold magnification in a blinded manner. The mean podocyte number per glomerular random cross-sectional area was calculated.

Quantification of nephrin surface expression in mice
Mice were perfused with either biotin or phosphate buffered saline (PBS). In kidneys perfused with biotin, linear staining along the glomerular capillaries was visualized. No staining was detected in controls perfused with PBS ( Fig 1A). To quantify the surface fraction of nephrin two different approaches have been used. First, nephrin was separated by immunoprecipitation with a nephrin specific antibody, followed by SDS-PAGE and western blot analyses. The biotinylated fraction of nephrin, was then visualized by staining with streptavidin. No Significance level was set to 5%, *p < 0.01, ** p<0.001, (non-significant differences (ns)).
https://doi.org/10.1371/journal.pone.0179217.g002 biotinylated nephrin was detected in PBS perfused controls. Lysate controls indicated equal amounts of protein (Fig 1B). Second, the reverse experiment has been performed. Biotinylated proteins were separated by pull-down with strepatividin agarose beads, followed by SDS-PAGE and western blot analyses. The nephrin fraction was visualized by staining with a nephrin specific antibody. To demonstrate that the assay allows the detection of other glomerular surface proteins podocalyxin was detected. The intracellular control proteins extracellular-signal regulated kinases p42 and p44 (ERK) could not be detected. Further, no biotinylated proteins were detected in PBS perfused controls. Lysate controls indicated equal amounts of protein. (Fig 1C).

Nephrin surface expression in nephrotoxic nephritis
Early nephrotoxic nephritis. Nephrotoxic nephritis (NTN) rapidly induces severe proteinuria in its heterologous phase [20]. One day after injection of the NTN serum, proteinuria in mice significantly increased 3-4 log ranks (Fig 2A and 2B). In comparison to untreated kidneys induction of NTN led to a transformation of the linear staining of nephrin to a punctate pattern. High resolution images show an enhanced colocalization of nephrin with early endosomal antigen 1 (EEA-1) (Fig 2C and 2D), suggesting increased endocytosis of the nephrin molecules. Loss of nephrin from the slit diaphragm was then quantified using the biotinylation assay (Fig 2E). At day 1, densitometric analysis of the western blot results indicated a 57% decrease of biotinylated nephrin in NTN mice (Fig 2F), but equal amounts of total nephrin were detected (Fig 2G). In line with this finding the podocyte numbers remained stable in the NTN mice compared to untreated animals (Fig 2H-2K).
Late nephrotoxic nephritis. Albuminuria decreased significantly over the course of 18 days (Fig 3A and 3B). At day 18, the results of the western blot analysis conducted after biotinylation indicated the recovery of surface nephrin (Fig 3C). At day 18, the densitometric results showed no significant differences between the percentage of cell surface nephrin in either group (Fig 3D). However, total nephrin was reduced by 25% in NTN mice at day 18 ( Fig  3E). Podocyte numbers were reduced by approximately 19% in NTN mice (Fig 3F-3J).

Adriamycin induced nephropathy
Adriamycin induced nephropathy (ADR) is a model of focal segmental sclerosis [21]. Six days after injection, a significant four log rank increase of proteinuria (compared to controls) was detected (Fig 4A and 4B). In comparison to healthy mice adriamycin induced a disruption of the linear staining pattern of nephrin, which was replaced by a fragmented and granulated cytoplasmic staining pattern. High resolution images show an enhanced colocalization of nephrin with EEA-1, suggesting endocytosis of nephrin (Fig 4C and 4D). These findings were confirmed in the in vivo biotinylation assay that was conducted using a western blot analysis (Fig 4E). Densitometric evaluation of the results at day 7 revealed that, despite a reduction of total nephrin by 71% (Fig 4F), biotinylated nephrin was still significantly decreased by 53% (Fig 4G). Podocyte numbers were reduced by approximately 40% (Fig 4H-4K).

Discussion
The in vivo biotinylation assay allowed the quantification of nephrin abundance at the glomerular slit diaphragm of mice. This method is useful as a complementary technique to immunofluorescence, and expands the possibilities of studies focusing on the expression and trafficking of glomerular surface proteins. As proof of principle the hallmark proteins of the podocyte nephrin and podocalyxin have been detected. To demonstrate the usefulness of the assay, we quantified nephrin surface abundance in two different and well-established models of proteinuric kidney disease.
NTN led to the rapid induction of albuminuria, which was nearly reversible in the course of disease. Quantification of nephrin revealed that more than half of the molecules were lost from the slit diaphragm at the peak of albuminuria, while the amount of total nephrin was preserved. In addition, the results of immunofluorescence studies showed a change of the linear distribution of nephrin to a punctate pattern. Colocalization experiments revealed that nephrin was translocated from the podocyte surface to the early endosome suggesting endocytosis of nephrin. Besides endocytosis changes in podocyte morphology and loss of podocytes presumably contribute to the observed pattern change. These findings are in line with previous results [8]. When albuminuria decreased once again, the surface abundance of nephrin was restored. However, after 18 days, nephrin abundance was diminished by approximately 25%. Histologic analysis of podocyte numbers suggested that this reduction in total nephrin was most likely caused by loss of podocytes. Our finding is also in line with previous reports of nephrotoxic nephritis [22]. Regarding ADR, albuminuric animals also showed significant reductions of surface nephrin. Immunofluorescence revealed an enhanced translocation of nephrin in vesicles of the early endosome. Moreover, there was a marked reduction of total nephrin most likely due to the loss of podocytes. This finding has been reported for ADR earlier [23], and is supported by our observation that podocyte numbers are strongly reduced in ADR at day 7. It should be kept in mind that an impaired perfusion with the magnetic particles due to advanced damage of the glomerular tuft could lead to a biased quantification of glomerular proteins. However, it might be questioned whether the inside of damaged capillary tufts would be adequately covered with biotin when extracted glomeruli are bathed in biotin like described by Satoh et al. [16]. To account for fluctuations in total protein yield and to ensure accurate quantitation and comparability between different experiments, all Western blot replicates have been normalized to beta actin.
https://doi.org/10.1371/journal.pone.0179217.g004 were reversible. Taken together this proof-of-concept study shows the potential of the in vivo biotinylation assay to gain more insight into the function of the glomerular filter.