Establishment of Nephrin Reporter Mice and Use for Chemical Screening

Nephrin is a critical component of glomerular filtration barrier, which is important to maintain glomerular structure and avoid proteinuria. Downregulation of nephrin expression is commonly observed at early stage of glomerular disorders, suggesting that methods to increase nephrin expression in podocytes may have therapeutic utility. Here, we generated a knockin mouse line carrying single copy of 5.5 kb nephrin promoter controlling expression of enhanced green fluorescent protein (EGFP) at Rosa26 genomic locus (Nephrin-EGFP mouse). In these mice, EGFP was specifically expressed in podocytes. Next, we isolated and cultivated glomeruli from these mice, and developed a protocol to automatically quantitate EGFP expression in cultured glomeruli. EGFP signal was markedly reduced after 5 days of culture but reduction was inhibited by vitamin D treatment. We confirmed that vitamin D increased mRNA and protein expression of endogenous nephrin in cultivated glomeruli. Thus, we generated a mouse line converting nephrin promoter activity into fluorescence, which can be used to screen compounds having activity to enhance nephrin gene expression.


Introduction
Slit membrane synthesized by podocytes plays an essential role to form glomerular filtration barrier in the kidney [1]. Loss in the function and numbers of podocytes is a key event in various renal disorders, leading to proteinuria, glomerulosclerosis and, eventually, end-stage renal disease [2]. To alleviate podocyte injury is one of major strategies to treat chronic kidney disease, but still efficient treatment options are limited [3]. Nephrin is a critical component of glomerular filtration barrier [4]. Downregulation of nephrin expression is commonly observed at early stage of glomerular disorders [3,5], suggesting that methods to increase nephrin expression in podocytes may have therapeutic utility.
Vitamin D has been known to increase nephrin expression in cultured mouse podocytes [6]. In renal precursor cells derived from human kidney cortex [7] or human amnion [8], vitamin D upregulates nephrin expression and induces differentiation into podocytes. Furthermore, recent reports have shown that vitamin D ameliorates proteinuria and glomerular lesions in murine models of diabetic nephropathy [9,10] or puromycin-induced nephrosis [11]. These findings suggest that screening of compounds having activity to enhance nephrin expression may lead to identification of a new therapeutic reagent to combat glomerular disorders.
Screening chemicals using gene promoter activity as an index is often performed using cultured cells transfected with a reporter plasmid encoding a fluorescent protein or an enzyme at the downstream of promoter sequence. However, the copy numbers or genomic locations of inserted gene cannot be controlled by gene transfer methods such as lipofection or electroporation. Therefore, it is important to analyze several independent clones to verify reproducibility. Transgenic (Tg) animals are quite useful to analyze a role of gene in vivo, but they also have the same problems of Tg copy numbers and integration sites and also a concern about unexpected disruption of unrelated gene. Indeed, introducing too much reporter DNA into cells may consume up transcription factors, resulting in unphysiological responses. Therefore, to strictly study developmental, spatio-temporal or pathophysiologic regulation of gene expression in vivo, gene knockin method by homologous recombination has been used to insert reporter gene under control of endogenous promoter [12], but these experiments are time consuming.
We invented a method to insert a single copy of DNA cassette, uni-directionally, into a specific genomic locus by direct injection of DNA construct into fertilized eggs utilizing the Cre-loxP system-mediated gene arrangement. This method, termed pronuclear injection-based targeted transgenesis (PITT), bypasses a use and screen of embryonic stem cells [13,14], and can markedly reduce potential interference from enhancer or silencer sequences surrounding the inserted gene.
In this study, we report establishment of a mouse cell line expressing enhanced green fluorescence protein (EGFP) under nephrin promoter using PITT method, and its use to screen compounds with activity to enhance nephrin expression.

Chemicals and media
Chemical reagents and media were purchased from Nacalai Tesque (Kyoto, Japan) unless otherwise described.
Animal care and procedures were approved by Institutional Animal Care and Use Committees of Tokai University (permit numbers 121007, 132013) and Kyoto University Graduate School of Medicine (Med Kyo 13116). Mice were maintained on a 12 hour light/dark cycle with free access to standard diet (F-2, Oriental BioService, Kyoto, Japan) and water. All mice were examined twice a week and were physically healthy. Overall mortality rate of mice was 2%. Cervical dislocation was used for euthanasia.
For genotyping of EGFP-Nephrin mice, genomic DNA extracted from tail was amplified with Thunderbird Probe qPCR Mix (Toyobo, Osaka, Japan) and analyzed by StepOnePlus Real-Time PCR System (Thermo Fisher Scientific, Waltham, MA, USA). Quantitation was carried out by ΔΔCT method. Amount of EGFP DNA was normalized by that of Mafb DNA (which has no introns) as internal control. PCR primers and probes used are shown in Table 1.

Reverse transcription-PCR for gene expression analysis
Total RNA was extracted from samples using RNeasy Plus Mini Kit (Qiagen, Hilden, Germany) and cDNA was synthesized by ReverTra Ace qPCR RT Master Mix with gDNA Remover (Toyobo, Osaka, Japan). Gene expression levels of nephrin (Nphs1) and podocin (Nphs2) were examined by quantitative PCR (qPCR) with Thunderbird Probe qPCR Mix and StepOnePlus Real-Time PCR System, and normalized by 18S ribosomal RNA levels (Table 1) [17,18].

Quantitative assay of fluorescence
After cultivation, primary culture of glomerular tissue was fixed with 4% PFA at 4°C for 30 min and stained by immunofluorescence, if appropriate, and also stained with 4',6-diamidino-2-phenylindole (DAPI, 500 ng/ml, Dojindo, Kumamoto, Japan). Fluorescence intensities of tissues immersed in 100 μl PBS were measured by ArrayScan VTI HCS Reader (Thermo Fisher Scientific) as described previously [21] using 5x objective (numerical aperture 0.25) through 3 channels: channel 1 (BGRFR 386-23 for DAPI), channel 2 (485-20 for EGFP) and channel 3 (549-15 for Alexa Fluor 568). Images were analyzed by HCS Studio 2.0 Cell Analysis Software with Cell Health Profiling Algorithm (Thermo Fisher Scientific). Glomeruli were identified as clusters of DAPI signals whose sizes were larger than a threshold value described in the Result section. Regions of interest (ROI) were set to glomeruli and signals of respective glomeruli at channel 2 or 3 were recorded. Data from 4 fields covering majority of a single well were combined and the mean value among glomeruli of each well was calculated.

Statistical analysis
Results are expressed as mean±SEM. Data were analyzed by 2-tailed Student's t test or ANOVA. Statistical significance was defined as P<0.05.

Results
Generation of nephrin reporter mice using PITT method A 5.5 kb nephrin promoter sequence has been shown to target gene expression specifically into podocytes of Tg mice [15], and nephrin-Cre mice have been successfully used to make podocyte-specific conditional knockout animals [22,23]. A donor vector carrying Nephrin+EGFP expression cassette and a Cre expression vector were co-injected into fertilized eggs isolated from the seed mice to obtain knockin mice carrying a single copy of Nephrin+EGFP cassette at the Rosa26 locus in the opposite direction to Rosa26 transcription [13,14] (Fig 1). Of 53 pups obtained, two mice showed expected gene recombination at the Rosa26 locus in their tail DNA. These two lines were mated with FLPe mice to remove extra sequence [16] and EGFP expression was confirmed in both lines preferentially at the periphery of glomeruli by immunohistochemistry (Fig 2A). Line 1 was named Nephrin-EGFP mice, and further analyzed. By genomic qPCR, wild-type, heterozygous and homozygous Nephrin-EGFP mice could be distinguished (Fig 2B). Heterozygotes were used to study regulation of EGFP expression.

Renal localization of EGFP expression
By immunofluorescence, EGFP expression co-localized well with endogenous nephrin and podocin, which are podocyte-specific molecules (Fig 3). On the other hand, cells stained by antibodies against PDGFR β (mesangial cell marker) or PECAM-1 (endothelial cell marker) did not express EGFP. Outside of glomeruli, no EGFP expression was observed. These findings verified that EGFP was specifically expressed in podocytes of Nephrin-EGFP mice.

Cultivation of glomeruli from Nephrin-EGFP mice
By magnetic bead method, approximately 10,000-20,000 glomeruli were obtained per mouse (95% purity by light microscopy) as previously described [20]. When glomeruli from Nephrin-EGFP mice were cultivated, EGFP fluorescence persisted at least for 2 days on the surface of glomeruli, but was markedly reduced within 5 days (Fig 4). On the other hand, after 1 or 2 days, nephrin mRNA level was decreased to 4.9% of the initial level on day1, further decreased to 0.8% on day 2, and partially recovered to 14.9% on day 5 (Fig 5). Apparent difference in the time course between EGFP fluorescence and nephrin mRNA was likely caused by much longer half life of EGFP molecule compared to that of endogenous nephrin mRNA. Podocin mRNA of cultured glomeruli showed similar changes: 1.8%, 0.9% and 17.6% on days 1, 2 and 5, respectively (Fig 5).
Previous reports showed that nephrin expression in cultured podocytes or renal precursor cells was increased by treatment with 10-100 nM vitamin D and 1 μM retinoic acid [6][7][8]. In glomeruli from Nephrin-EGFP mice, addition of vitamin D+retinoic acid increased EGFP fluorescence (Fig 4), thus validating our reporter system.

Automated quantitation of EGFP signals
To detect glomeruli, cultivated material was fixed and stained with a nuclear marker DAPI. Blue signals by DAPI were scattered in podocyte outgrowth, but were clustered in glomeruli ( Fig 6A). The sizes of DAPI clusters were automatically calculated. When a threshold of > 2000 μm 2 was selected, recognition of glomeruli from migrated podocytes was best performed and matched well with visual judgement (Fig 6B). This threshold seems reasonable since it is consistent to a diameter of > 45 μm used in classical sieving method for mouse glomerulus isolation [24]. After setting ROI to glomerular areas, EGFP signals in glomeruli were quantitated as green fluorescence (Fig 6C).
After 5 days of culture, some glomeruli showed preserved structure, while other glomeruli appeared destructed or melted by escape of cells from glomeruli ( Fig 6A). As a consequence, total EGFP signals ( Fig 7A) and the numbers of DAPI-positive cells (Fig 6A) of individual glomeruli were highly variable even in a single well. Consistently, when EGFP signal per DAPI signal was calculated for each glomerulus, distribution of histogram was much concentrated ( Fig  7B). When normalization of EGFP signal by area and that by DAPI signal were compared, they gave very similar patterns (Fig 7C and 7D). Therefore, simple normalization by area was  chosen, and the average of EGFP signal/glomerular area among whole glomeruli from each well was used as a representative value for that well.
Next, FBS concentration for chemical screening was optimized (Fig 8). After isolation from Nephrin-EGFP mice, glomeruli were seeded at a density of 100-150 glomeruli per well and cultivated for 3 days in DMEM containing 0%, 0.5% or 10% FBS (Days 0-3). By changing medium to fresh DMEM with 0%, 0.5% or 5% FBS on day 3, approximately half of glomeruli were removed. On day 5, cultured materials were fixed and stained with DAPI. At this point, approximately one third of glomeruli of the initial number were retained on plates (with some variations) and EGFP fluorescence was quantitated. Culturing glomeruli with 10% FBS for the first 3 days (days 0-3) increased the recovery of glomeruli by 30% compared to 0% or 0.5% FBS (Fig 8A). Cultivation with 5% FBS during the last 2 days (days 4-5, in the presence of vitamin D plus retinoic acid) further increased recovery by 20% compared to 0% or 0.5% FBS during days 4-5. When EGFP signals were measured as EGFP/Area, cultivation with 5% FBS during days 4-5 reduced the signals by 20% compared to 0.5% FBS during days 4-5 (after incubation with 0.5% or 10% FBS during days 0-3), presumably due to enhanced outgrowth and escape of podocytes from glomeruli by 5% FBS during days 4-5 ( Fig 8B). As a consequence,  10% FBS on days 0-3 plus 0.5% FBS on days 4-5 gave the largest difference between vitamin D+retinoic acid vs vehicle treatment in EGFP/Area values (by more than 2-fold, Fig 8C), and these FBS conditions were selected.  Performance of Nephrin-EGFP mouse glomeruli in culture system First, the effects of various vitamin D concentrations were tested. EGFP signals were dosedependently increased by vitamin D: 2.6-fold by 1 nM, 3.9-fold by 50 nM and 4.5-fold by 1250 nM (Fig 9). Since addition of retinoic acid (1 μM) to 50 nM vitamin D did not further increase expression of EGFP, we chose 50 nM vitamin D without retinoic acid as a positive control for glomerular EGFP assay. Treatment of glomeruli during the last two days with vitamin D in the presence of 0.5% FBS increased nephrin mRNA expression by 3.6-fold on day 5 compared to vehicle treatment, without significant effects upon podocin mRNA (Fig 5). Dose-response curves of EGFP signals and nephrin mRNA expression by vitamin D were considerably proportional (Fig 9).
We studied effects of DMSO upon chemical screening, since all reagents were first dissolved in DMSO. DMSO at 2% mildly decreased EGFP signals both in vehicle-and 50 nM vitamin Dtreated glomeruli by 20% (Fig 10), but 1% DMSO showed no significant effects. Therefore, we judged that presence of 0.1-1% DMSO during screening is tolerable.
Endogenous nephrin protein expression in cultured glomeruli and its change by vitamin D was examined. By immunofluorescence and quantitation, nephrin protein expression was increased by 2.7-fold with vitamin D treatment, while EGFP signal was increased by 4.1-fold ( Fig  11). On the other hand, podocin protein expression was only mildly increased (1.3-fold) by vitamin D, suggesting predominant effects of vitamin D upon nephrin expression. In a time course analysis, EGFP intensity was slightly elevated on day 2, and gradually decreased towards day 5.
An example of screening in a 96-well format is shown in Fig 12. Vitamin D-treated wells showed constantly and clearly increased EGFP fluorescence signals by excitation compared to vehicle-treated wells (Fig 12). Positive wells with similar signal levels or much stronger signals compared to vitamin D-treated wells were also found. Next, we investigated dose-response of some positive chemicals as to EGFP intensity (Fig 13). Very strong EGFP signals by several reagents were often reduced by 90% when the reagent concentrations were reduced by 90%, which was not the case with vitamin D (Fig 9). We realized that, in such cases, chemicals presumably gave very strong fluorescence in glomeruli because of the colors of reagents (which are visible even under standard room light, Fig 13), especially when they had high cell permeability to live cells. These findings suggested that studying dose-response and the color of compounds may be efficient ways to exclude false positive hits from the primary screening.

Discussion
In the present study, we generated Nephrin-EGFP mice, which specifically express EGFP in podocytes under the control of nephrin promoter inserted at the Rosa26 locus. To our  Quantitative evaluation of changes in EGFP, nephrin and podocin protein expression in glomeruli by vitamin D. Endogenous nephrin and podocin were visualized by immunofluorescence using primary antibodies and Alexa Fluor 568-conjugated secondary antibodies. Fluorescence intensity was measured on days 1-5 (D1-D5). Glomeruli were treated with 50 nM vitamin D (VD) or vehicle (Veh) for the last 2 days and examined on day 5 (D5), and the difference was studied by unpaired t test. For EGFP quantitation, mean background signal of wild-type glomeruli was subtracted and the level of Veh-treated Nephrin-EGFP glomeruli on D5 was defined as 1.0 unit. For nephrin and podocin, mean background signal of wild-type glomeruli incubated with secondary antibody alone was subtracted, respectively. Difference between EGFP signals among days was examined by one-way ANOVA with Bonferroni's multiple comparison test. N = 5.
knowledge, this strain is the first nephrin reporter line to be established by knockin strategy, instead of conventional random transgene insertion [25]. Next, we isolated glomeruli from these mice, and set up cultivation and assay which can be used to screen reagents enhancing nephrin transcriptional activity.
We previously showed that, unlike traditional random-insertion transgenesis approaches, the targeted transgenesis approach using PITT (insertion of a gene of interest into the Rosa26 locus) enables to obtain Tg mice with stable, reliable, predictable, and reproducible transgene  expression when constitutive CAG promoter was used [13,14]. In the present study, we applied the PITT strategy for generation of Tg mice showing tissue-specific expression. Because both of two Nephrin-EGFP founder mouse lines identically exhibited highly podocyte-specific expression, the PITT method is useful also for production of tissue-specific Tg mice with reproducible and expected transgene expression.
Since podocytes play an essential role for maintenance of structure and function of glomeruli [1][2][3][4][5], cultured podocytes have been successfully used to screen chemicals possessing activities to prevent morphological change or damage of podocytes [26]. Upon cultivation of glomeruli on plates covered with type I collagen, podocytes start to migrate out as single cells [27] but substantial amount of podocytes remain upon glomeruli. For assessment of EGFP expression, we uniquely focused on podocytes attached to glomeruli, since such cells are ligated to a native scaffold, glomerular basement membrane, and located close to endogenous mesangial and endothelial cells, which might help mutual exchange of soluble factors such as vascular endothelial growth factor [25].
Our system is also unique in a point that we used freshly isolated glomeruli and podocytes without any passage. Since podocytes are terminally differentiated cells with very small proliferating capacity especially in vivo [28], primary podocytes cannot keep on growing and require immortalization procedure for propagation [29], which might alter the nature of podocytes. Further studies are required to judge whether features in our screening method provides advantage compared to methods reported so far [6,10,26].
There are several limitations in this work. In our culture system, certain number of podocytes are lost by day 5 (when EGFP intensity is evaluated) by podocyte outgrowth and medium change. Furthermore, we describe here that colors of the original compounds may affect the results of fluorescence-based assay. Therefore, findings in nephrin reporter assay have to be verified by evaluation of endogenous nephrin mRNA and protein expression as we did for vitamin D (Figs 5 and 11).
Vitamin D has been known to upregulate nephrin expression both in vitro and vivo, and to inhibit glomerular injury [6,[9][10][11]. Using Nephrin-EGFP glomeruli, we could show that vitamin D increases EGFP fluorescence intensity as a positive control essential for chemical library screening.