A small molecule screening to detect potential therapeutic targets in human podocytes

Abstract

Widmeier E, Tan W, Airik M, Hildebrandt F. A small molecule screening to detect potential therapeutic targets in human podocytes. Am J Physiol Renal Physiol 312: F157–F171, 2017. First published October 19, 2016; doi:10.1152/ajprenal.00386.2016. Steroid-resistant nephrotic syndrome (SRNS) inevitably progresses to end-stage kidney disease, requiring dialysis or transplantation for survival. However, treatment modalities and drug discovery remain limited. Mutations in over 30 genes have been discovered as monogenic causes of SRNS. Most of these genes are predominantly expressed in the glomerular epithelial cell, the podocyte, placing it at the center of the pathogenesis of SRNS. Podocyte migration rate (PMR) represents a relevant intermediate phenotype of disease in monogenic causes of SRNS. We therefore adapted PMR in a high-throughput manner to screen small molecules as potential therapeutic targets for SRNS. We performed a high-throughput drug screening of a National Institutes of Health Clinical Collection (NCC) library (n = 725 compounds) measuring PMR by videomicroscopy. We used the Woundmaker to perform individual 96-well scratch wounds and screened compounds using a quantitative kinetic live cell imaging migration assay using IncuCyte ZOOM technology. Using a normal distribution for the average PMR in wild-type podocytes with a vehicle control (DMSO), we applied a 90% confidence interval to define “distinct” compounds (5% faster/slower PMR) and found that 12 of 725 compounds (at 10 μM) reduced PMR. Clusters of drugs that alter PMR included actin/tubulin modulators such as the azole class of antifungals and antineoplastic vinca-alkaloids. We hereby identify compounds that alter PMR. The PMR assay provides a new avenue to test therapeutics for nephrotic syndrome. Positive results may reveal novel pathways in the study of glomerular diseases such as SRNS.

the prevalence of end-stage kidney disease, which requires dialysis or transplantation for survival, has been increasing over the last few decades (27). Total healthcare cost for chronic kidney disease (CKD) now exceeds >$40 billion dollars annually (26) and continues to rise as renal replacement therapy and transplant survival have improved. Although there have been advancements in treatment regimens, there has been little to no drug development to address any of the primary causes of CKD, including steroid-resistant nephrotic syndrome (SRNS) and its histological hallmark, focal segmental glomerulosclerosis (FSGS). SRNS represents the second most frequent cause of CKD that manifests before 25 yr of age (104). Currently, there are more than 30 monogenic genes that, if mutated, cause SRNS. All of them are relevantly expressed in the glomerular epithelial cell, the podocyte, placing it at the center of the pathogenesis of SRNS (71). The pathophysiology of nephrotic syndrome is characterized by structural alteration of cytoskeleton and molecular reorganization of slit diaphragm components leading to foot process effacement (FP). Previous work has demonstrated the well-established concept that FP effacement is a migratory event, making the podocyte migration rate (PMR) an important functional assay for studying pathological condition in vitro (93). Within the podocyte, there are several pathway-specific mechanisms that are essential for disease pathogenesis. Previous work has established small Rho-like GTPase signaling to play a central role in the pathogenesis of SRNS (39–41, 102). In studying the effect of small Rho-like GTPase signaling, the PMR was demonstrated to represent a relevant intermediate phenotype of disease in monogenic causes of SRNS (38, 40, 41). As we have shown in SRNS, PMR can be increased or decreased, therefore implying that the balance of RhoA/Rac1/Cdc42 signaling, not overall PMR, is relevant to disease (9, 39, 40). In the last 10–15 yr advances in high-throughput screening techniques have opened the way for drug discovery (51, 122). In addition, several groups have recently attempted to adapt high-content screening to the field of nephrology by using podocyte based assays (68). We therefore hypothesized that using PMR in a live cell-based assay could be adapted in a high-throughput fashion to identify pathways integral to the pathogenesis of nephrotic syndrome as well as potential novel therapeutics. We have identified 12 compounds that reduced PMR.

METHODS AND STUDY DESIGN

Drug library.

The National Institutes of Health Clinical Collection (NIHCC) is a library of 725 compounds spread across ten 96-well plates (8 plates have 80 compounds, 1 plate has 45 compounds, and 1 plate has 40 compounds). The drugs represented by this library have been selected because of their purity, solubility, and commercial availability for resupply (NCC; National Center for Advancing Translational Sciences, Bethesda, MD). The compounds were prepared and shipped by Evotec as part of the National Institutes of Health Small Molecule Repository (NIHSMR). The compounds were shipped as a 10-mM stock solution diluted in DMSO solvent. Other compounds tested include Rho Activator II (CN03; Cytoskeleton), Rho Pathway Inhibitor I (ROCK Y-27632, CN06; Cytoskeleton), Rac1 inhibitor (553502; Millipore), and Rac1 Inhibitor II (553511; Millipore).

Cell culture.

The immortalized human podocyte cell line was a kind gift from M. Saleem (University of Bristol) and was cultured as previously described (96). Human podocytes were plated at 35,000 cells/well and incubated at 33°C 12 h before making the scratch wound. Cells were grown at 37°C after the scratch wound was made and allowed to migrate. Cells were tested for mycoplasma contamination on a biweekly basis.

Human podocyte cell line expressing stable nuclear mKate2 (red fluorophore).

IncuCyte NucLight Red lentivirus reagent was purchased from Essen Bioscience. The human podocyte cell line (gift from M. Saleem) was transduced according to manufacturer’s instructions per protocol, and 48 h after transduction, puromycin at a final concentration of 4 µg/ml was added to the medium for selection of transduced cells, which stably express nuclear mKate2.

Scratch wound assay.

Cells were plated on 96-well Image-lock plates (Essen Bioscience). Podocytes were examined for confluency as a monolayer via light microscopy before initiation of any scratch wound. Scratches were made by using a 96-pin tool (Woundmaker) as per protocol.

Proliferation assay.

Cells were plated on 96-well Image-lock plates (Essen Bioscience). Before initiation of any scratch wound, podocytes were examined for confluence as a monolayer via light microscopy. Scratches were made by using a 96-pin tool (Woundmaker) as per protocol. Whole well images were acquired at the beginning (t₀) and at the end (t_x) of the experiment.

Videomicroscopy.

Podocyte cell proliferation was assessed using the whole well assay format. Cells were monitored automatically via live cell imaging using the IncuCyte videomicroscopy system at the beginning (t₀) and at the end (t_x) of the experiment. Whole well images were automatically acquired and recorded by the IncuCyte software (Controller version 2015A Rev 1). Podocyte cell migration was assessed using the scratch-wound assay format. Cells were monitored automatically via live kinetic cell imaging using the IncuCyte videomicroscopy system at 60-min intervals. Wound images were automatically acquired and recorded by the IncuCyte software (Controller version 2015A Rev 1).

Data analysis.

Data processing and analysis for proliferation assay were done using the IncuCyte 96-well Basic Analyzer software module. Videomicroscopy was performed with a ×4 objective. Individual whole wells were analyzed with the IncuCyte GUI software. Data processing and analysis for migration assay were done using the IncuCyte 96-well Kinetic Cell Migration and Invasion Assay software module. Data were then exported to Excel for further analysis. Wound width is defined as the area of the wound at any time t, as determined by the processing software. Wound confluence is expressed as a percentage of the scratch wound that is filled with cells at any given time t, when compared with when the scratch was initially performed. Wound closure, as a measurement of PMR, was monitored at 60-min intervals for at least 20 h. Videomicroscopy was performed with a ×10 objective. Migration was performed at 37°C to minimize the effects of cell proliferation. Individual scratch wounds were analyzed with the Incucyte GUI software and inspected visually for cell viability. All wells that did not demonstrate a confluent monolayer after the scratch wound/washing process were discarded from analysis (Fig. 1A).

Fig. 1.

Methods to determine positive hits with videomicroscopy based scratch wound assay. Wells for the podocyte migration rate (PMR) assay were chosen for quality depending on wound width, confluence (A), and characteristics of a normal distribution curve (B). A: 2 representative images are shown of human podocytes at time point t₀, after a scratch-wound. Left: adequate scratch with a confluent monolayer above and below the wound with no debris/cellular remains within the wound to confound analysis. Right: inadequate wound width and cell confluency due to poor podocyte coverage above and below the scratch wound. B: normal distribution curve is displayed with demarcations at 1.65 SD from the mean [90% confidence interval (CI)]. In the first part of our evaluation, outliers (outside the 90% CI) in wound width were deemed inadequate and discarded from analysis. In the second part of our evaluation, wound confluence was used to identify wells with “distinct” PMR outside the 90% CI and kept as positive results.

Drug screen.

Each drug plate of the NIHCC was plated at 10 μM into an individual well for scratch wound assay analysis. Each compound was screened in triplicate to ensure reproducibility. Every plate screened also had a 16 replicates for DMSO as a vehicle control, and this was used as a control to compare wound closure rates. After the preliminary screen was completed, all preliminary hits were then screened again in quadruplicate at differing concentrations (1, 5, and 10 μM) to assess for a dosage effect.

Statistical analysis.

Student's t-test was used to determine the statistical significance between two interventions, and one-way ANOVA followed by Bonferroni’s correction was used for multiple comparisons (GraphPad Prism software). A statistically significant difference was defined as P < 0.05 and is marked as follows: **P < 0.01, ***P < 0.001, ****P < 0.0001.

RESULTS

We first performed a proliferation assay to distinguish the proliferation rate of wild-type podocytes under migration conditions using the control condition containing DMSO vehicle control (0.1%). We demonstrate that the proliferation rate of human podocytes significantly decreases over time by 23.2% over 26 h predominantly due to DMSO toxicity (Fig. 2) (88). Additionally, we performed a proof of principle experiment with RhoA and Rac1 signaling pathway effectors (Fig. 3) replicating previous data (1, 2). The proof of principle experiment with the established microtubule modulators had shown a reduction in PMR (Fig. 4 and Fig. 5) replicating previous data (111, 112). Additionally, our data showed that microtubule modulators impair the cell viability reducing the cell count, however, without compromising the PMR (Fig. 6). In these experiments we demonstrate that PMR can be reproducibly measured and that pharmaceutical compounds can be assayed in a high throughput fashion. We then extended the live cell based kinetic videomicroscopy based scratch wound assay to a National Insitutes of Health small molecule therapeutic library to assess for pharmaceutical modifiers of PMR.

Fig. 2.

Proliferation assay in immortalized undifferentiated human podocytes. Stably nuclear mKate2 expressing wild-type podocytes from an immortalized human podocytes cell line were seeded 12 h before the beginning of the experiment on a 96-well image-lock plate. Images were captured at the beginning and at the end of the experiment. A significant reduction of cell count in average of 23.2% on 240 replicates is shown over a period of 26 h in cultured media containing DMSO vehicle control (0.1%). Data are expressed as means ± SD for 3 independent experiments, 80 replicates each.

Fig. 3.

Proof of principle experiment with RhoA and Rac1 effectors. Wild-type podocytes were seeded 12 h before scratch wound analysis on a 96-well image-lock plate. After scratch wound was made in the confluent podocyte monolayer, podocytes were exposed to different compounds including ROCK inhibitor (CN06; Cytoskeleton), RhoA activator (CN03; Cytoskeleton), and Rac1 inhibitor #1 (553502; Millipore), and Rac1 inhibitor #2 (553511; Millipore). Concentrations are as follows: RhoA inhibitor: 10 μM; RhoA activator: 1 μg/ml; Rac1 inhibitor #1: 10 μM; and Rac1 inhibitor #2: 10 μM, as per dosing instructions. RhoA effectors were diluted in water. Rac1 effectors were diluted in DMSO. The volume of total media including drug in each well is 100 μl. Note that, whereas ROCK inhibitors increased PMR, RhoA activators as well as Rac1 inhibitors #1 and #2 decreased PMR. Data are expressed as means ± SD for 2 independent experiments.

Fig. 4.

Proof-of-principle experiment with established modulators of microtubule polymerization/depolymerization such as colcemid, nocodazole, paclitaxel, and vinblastine. Stably nuclear mKate2 expressing wild-type podocytes were seeded 12 h before scratch wound analysis on a 96 well image-lock plate. After scratch wound was made in the confluent podocyte monolayer, podocytes were exposed to compounds including colcemid (no. 10295892001; Roche), nocodazole (M1404; Sigma), paclitaxel (T7402; Sigma), vinblastine (V1377; Sigma), and DMSO vehicle control (0.05%). Concentrations are as follows: 0.05 μg/ml colcemid, 5 μM nocodazole, 5 μM paclitaxel, and 5 μM vinblastine. Colcemid was delivered as ready to use solution, DMSO was added accordingly. Nocodazole, paclitaxel, and vinblastine were diluted in DMSO. All drugs significantly decreased PMR at the established concentrations.

Fig. 5.

Dose response of established modulators of microtubule demonstrates decreased PMR. Stably nuclear mKate2 expressing wild-type podocytes were seeded 12 h before scratch wound analysis on a 96 well image-lock plate. After scratch wound was made in the confluent podocyte monolayer, podocytes were exposed to compounds in different concentration as indicated. A: colcemid (control, n = 47; 0.1 µg/ml colcemid, n = 47; 0.05 µg/ml colcemid, n = 32; and 0.01 µg/ml colcemid, n = 31) alters the PMR partially in a concentration-response manner. B and C: nocodazole (0.1% DMSO vehicle control, n = 45; 10 µM nocodazole, n = 48; 5 µM nocodazole, n = 3; and 1 µM nocodazole, n = 32) and paclitaxel (0.1% DMSO vehicle control, n = 46; 10 µM paclitaxel, n = 48; 5 µM paclitaxel, n = 31, 1 µM paclitaxel, n = 32) reduce PMR in a concentration-independent manner. D: whereas vinblastine (0.1% DMSO vehicle control, n = 47; 10 µM vinblastine, n = 48; 5 µM vinblastine, n = 32; and 1 µM vinblastine, n = 32) reduces PMR in a concentration-dependent manner.

Fig. 6.

Toxicity of established microtubule modulators reduce podocytes cell count without compromizing the PMR. Wild-type podocytes from an immortalized human podocytes cell line stably expressing nuclear mKate2 were seeded 12 h before the beginning of the experiment on a 96-well image-lock plate. Whole well images were captured after drug exposure at the beginning (t₀) and at the end (t_x) of the experiment to perform the quantitative analysis of total cell count. Live cell imaging captured immortalized human podocytes on an hourly basis to characterize the cellular morphology and viability of podocytes, while undergoing migration in a scratch wound assay. A, C, E, and G: all compounds at indicated concentration reduce the cell count ratio and show substantial cell morphology changes vs. control condition. B, D, F, and H: all compounds significantly affected PMR as well, however, without compromising the migratory behavior (see Fig. 5; also see Supplemental Movies 1-8; Supplemental material for this article is available at the Journal website). Data are expressed as a ratio of means ± SD for 2 independent experiments. NS = not significant; **P < 0.01, ***P < 0.001, and ****P < 0.0001 by one-way ANOVA Bonferroni’s multiple comparison test.

All 725 compounds from the NIHCC were examined by the PMR assay. Wells with significantly different PMR as compared with vehicle control (DMSO) were defined as “distinct” and were thereby identified as positive results. These “distinct” wells were isolated from the exported IncuCyte data in a five-step process:

• 1)After visual inspection, all wells that passed this initial filtering criterion were subject to a quantitative quality control step. Remaining wells were measured for wound width at time, t₀ (at initial scratch), and all outliers outside the 90% confidence interval were removed from analysis as well (Fig. 1A). The 90% confidence interval for wound width was determined from a standard deviation calculated by using the wound width for all experimental conditions per plate at t₀ (Fig. 1B). No individual compound was analyzed for wound closure rate if there were not at least two replicate conditions passing visual or wound width inspection.
• 2)Wells that passed visual and quantitative inspection then were analyzed for wound closure rates, as a measurement of PMR. Wound closure rates were calculated using wound confluence as determined by the IncuCyte software module.
• 3)For each individual compound, each replicate was averaged and wound confluence was measured over time. Assuming a normal distribution of wound closure in wild-type immortalized podocytes, the rate (slope of wound confluence over change in time) was calculated for each compound during the linear phase of migration and compared with the rate of wound closure of a control condition (DMSO).
• 4)Any compound with a wound closure rate during the linear phase of migration that was outside the 90% confidence interval was classified as a positive result (Fig. 7).
• 5)“Distinct” wells were then confirmed again in quadruplicate with different concentrations as noted previously. The 90% confidence interval was determined from a standard deviation as calculated by using the DMSO vehicle control as the standard control condition.

Fig. 7.

Positive drug screening hits determined by PMR as compared with vehicle control. Examples of positive and negative results in National Institutes of Health Clinical Collection (NIHCC) Plate #4. A: DMSO 0.1% vehicle control at time 0, 5, 10, and 15 h, respectively. B: Drug #17 (10 μM), lomerizine, at time 0, 5, 10, and 15 h, respectively. C: Drug #33 (10 μM), loxoprofen, at time 0 and 5, 10, and 15 h, respectively. D: graphical representation of wound confluence (y-axis) over time (x-axis) for DMSO, Drug #17 (lomerizine), and Drug #33 (loxoprofen) over an average of 3 replicates. E: see Supplemental Movies 9-11 for movies of wound confluence over time of vehicle control DMSO 0.1%, Drug #17 (lomerizine), and Drug #33 (loxoprofen) for 1 replicate.

Of the 725 compounds screened, 632 compounds passed the quality filtering steps of visual inspection (Fig. 1A) and wound width standardization for analysis of podocyte migration rate.

Of the 632 compounds analyzed for PMR via wound closure rate, 61 compounds were deemed as initial positive results (Table 1. Fifty-seven compounds reduced PMR in podocytes and 4 compounds increased PMR (Fig. 8).

Table 1.

Fifty-seven preliminary positive hits from drug screening with decreased PMR and 4 with increased PMR in 725 compounds NIHCC at 10-µM concentration

Name	Result	Mechanism	Pubchem ID	Reference	Drug Library Plate
Topotecan	Decreased migration	Anti-neoplastic topo I inhibitor	46386667	(64, 99)	NIHCC Plate_1_A8
SDM25N	Decreased migration	Gamma-receptor antagonist	46387008	(75)	NIHCC Plate_1_C5
Vincristine	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386588	(57, 87)	NIHCC Plate_1_G8
Vindesine	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386586	(87)	NIHCC Plate_1_H11
Adenosine, N-(2-hydroxycyclopentyl)-, (1S-trans)	Decreased migration	Activation of purine receptors A1 and A2	46387015	(110)	NIHCC Plate_2_A6
Ezetimibe	Decreased migration	inhibits absorption of cholesterol	46386640	(2, 37)	NIHCC Plate_2_A8
8-Azaspiro[4.5]decane-7,9-dione, 8-[2-][(2,3-dihydro-1,4-benzodioxin-2-yl)methyl]amino[ethyl]-,monomethanesulfonate	Decreased migration	Anticonvulsant	46387017	(84)	NIHCC Plate_2_A10
N,N′-diacetyl-1,6-diaminohexane	Decreased migration	Experimental compound; used in production of nylon	46386869	N/A	NIHCC Plate_2_B2
Cefatrizine propylene glycol	Decreased migration	Cephalosporin	46386659	(81)	NIHCC Plate_2_B3
Oxymetholone	Decreased migration	Anabolic steroid	46386778	(7, 80)	NIHCC Plate_2_B10
Anastrozole	Decreased migration	Nonsteroidal aromatase inhibitor	46386543	(79, 97)	NIHCC Plate_2_C7
Rimcazole	Decreased migration	Sigma-receptor antagonist	46387003	(42)	NIHCC Plate_2_D4
Zolmitriptan	Decreased migration	Triptan-selective agonist of serotonin 1B and 1D receptor	46386880	(74, 89)	NIHCC Plate_2_E11
Artesunate	Decreased migration	Semi-synthetic derivative of artemisinin	46386645	(52)	NIHCC Plate_2_H6
Nimetazepam	Decreased migration	Benzodiazepine	46386768	(56)	NIHCC Plate_2_H8
Ramipril	Increased migration*	ACE inhibitor	46386770	(69)	NIHCC Plate_3_B2
Ampiroxicam	Increased migration*	Prodrug of piroxicam; NSAID-reversible cox-1 inhibitor	46386688	(17, 33)	NIHCC Plate_3_E4
Glycine, N-[2-(acetylthio)methyl]-1-oxo-3-phenylpropyl-,phenylmethyl ester	Increased migration*	Amino acid; neurotransmitter in CNS	46386930	(8, 20, 65, 117)	NIHCC Plate_3_E6
Triptolide	Decreased migration	Anti-inflammatory; podocyte protective	46386571	(12, 16, 46)	NIHCC Plate_3_H2
Ethylestrenol	Decreased migration	Anabolic steroid (Pregnane steroids)-little androgenic effect	46386573	(114)	NIHCC Plate_3_H4
Midazolam HCl	Increased migration*	Benzodiazepine-GABA potentiator	46386603	(76, 103)	NIHCC Plate_3_H6
Lomerizine DiHCl	Decreased migration	Ca channel blocker	46386707	(45, 94)	NIHCC Plate_4_A4
Pancuronium	Decreased migration	Nondepolarizing muscle relaxant; competitive acetylcholine antagonist at NMJ	46386853	(29)	NIHCC Plate_4_B3
Trazodone HCl	Decreased migration	Antidepressant; binds 5HT2; selective reuptake inhibitor	46386915	(73)	NIHCC Plate_4_B8
Metronidazole	Decreased migration	Nitroimidazole antibiotic; inhibiting bacterial DNA synthesis	46386860	(121)	NIHCC Plate_4_B9
Saquinavir mesylate	Decreased migration	HIV protease inhibitor	46386596	(28)	NIHCC Plate_4_C6
Tegaserod maleate	Decreased migration	5HT4 agonist; used for IBS	46386624	(15)	NIHCC Plate_4_C7
Diphenylcyclopropenone	Decreased migration	Local irritant	46386897	(106)	NIHCC Plate_4_C8
Nifekalant HCl	Decreased migration	Class III anti-arrhythmic; inhibits hERG channel	46386697	(36, 50)	NIHCC Plate_4_C11
Bifemelane	Decreased migration	Neuroprotective; mechanism not well understood	46386922	(82, 86)	NIHCC Plate_4_D6
Loratidine	Decreased migration	Second generation H1 receptor antagonist	46386837	(14)	NIHCC Plate_4_D7
Mesoridazine	Decreased migration	Phenothiazine; adrenergic blockade; hERG cell blockade	46386921	(22, 108)	NIHCC Plate_4_D8
Irinotecan HCl	Decreased migration	Antineoplastic-topo I inhibitor	46386616	(18)	NIHCC Plate_4_D10
Rifapentine	Decreased migration	Antibiotic; inhibits RNA polymerase in bacteria	46386637	(70)	NIHCC Plate_4_D11
Vinorelbine bitatrate	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386815	(57, 87)	NIHCC Plate_4_E4
Indatraline	Decreased migration	Monoamine transporter inhibitor	46386808	(47, 115)	NIHCC Plate_4_E5
Ketorolac	Decreased migration	NSAID; COX1 and COX2 inhibition	46386614	(48)	NIHCC Plate_4_E9
Ethynylestradiol	Decreased migration	Synthetic estrogen derivative	46386858	(98)	NIHCC Plate_4_F9
Cetraxate HCl	Decreased migration	Anti-ulcer cytoprotective agent (GI tract)	46386678	(66)	NIHCC Plate_4_F10
HTMT	Decreased migration	H1/H2 agonist	46386998	(91)	NIHCC Plate_4_F11
SR 57,227A	Decreased migration	5HT3 agonist	46386849	(11)	NIHCC Plate_4_G9
Tripelennamine HCl	Decreased migration	H1 antagonist	46386917	(119)	NIHCC Plate_4_H9
5-Methoxytryptamine	Decreased migration	Tryptamine derivative (melatonin)-agonist of 5HT1,2,4,6,7 receptors)	46387021	(34)	NIHCC Plate_4_H11
Nifedipine	Decreased migration	Ca-channel blocker	46386790	(21, 77, 107a)	NIHCC Plate_5_F8
Doxapram	Decreased migration	K-channel subfamily K blocker	46386821	(4, 90, 120)	NIHCC Plate_5_H10
Nitazoxanide	Decreased migration	Pyruvate-flavodoxin oxidoreductase inhibitor	46386689	(13, 49)	NIHCC Plate_6_A6
5-Cyclopropyl-1-(2-methoxypropyl)-5-methyl-2-phenylpiperazine	Decreased migration	Unknown	104170212	N/A	NIHCC Plate_7_D5
8-tert-butyl-6-(2-methoxypropyl)-6,9-diazaspiro[4.5]decane	Decreased migration	Unknown	104170115	N/A	NIHCC Plate_7_A10
Felodipine	Decreased migration	Ca-channel blocker, inhibition of mineralocorticoid receptor, PDE1A/1B inhibitor	104170219	(21, 30, 35, 67, 100, 107a)	NIHCC Plate_7_C5
Albendazole	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition	104170113	(23, 92, 105)	NIHCC Plate_7_C9
Azathioprine	Decreased migration	Inhibitor of hypoxanthine-guanine phosphoribosyl transferase	104170112	(6, 32)	NIHCC Plate_7_D9
Griseofulvin	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition	104170114	(55, 87),	NIHCC Plate_7_D10
Miconazole	Decreased migration	Lanosterol 14-alpha demethylase inhibitor, K voltage-gated channel subfamily H inhibitor, Ca-activated K channel inhibitor	104169959	(3, 116)	NIHCC Plate_7_E3
Daunorubicin	Decreased migration	DNA-topoisomerase-2 inhibitor	104170197	(10, 123)	NIHCC Plate_7_E7
Minocycline	Decreased migration	Caspase-1 and -3 negative modulator, VEGF inhibitor, matrixmetalloproteinase-9 inhibitor, 16 rRNA inhibitor	104169958	(19, 31, 55, 87, 95, 109, 118)	NIHCC Plate_
Mobendazole	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition	104170137	(72, 85)	NIHCC Plate_8_A8
Digoxin	Decreased migration	Na-K-ATPase inhibitor	104170057	(1)	NIHCC Plate_8_G5
5-Azacytidine	Decreased migration	DNA-methyltransferase 1 inhibitor	104170170	(24, 83),	NIHCC Plate_9_A10
Podofilox	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition, DNA-topoisomerase-2 inhibitor	104170173	(54, 59)	NIHCC Plate_9_C8
Mitoxantrone	Decreased migration	DNA intercalation, DNA-topoisomerase-2 inhibitor	104170122	(21, 44)	NIHCC Plate_9_E5
1-(2-Methoxypropyl)-2,5,5-trimethyl-2-phenylpiperazine	Decreased migration	N/A	104170092	N/A	NIHCC Plate_10_B3

Preliminary hits ordered in ascending order from NIHCC plate. Fifty-seven of 61 positive results decreased migration and 4 increased (*) migration. All positive hits were screened in triplicate at 10 µM concentration and compared with a vehicle control (0.1% DMSO). NIHCC, National Institutes of Health Clinical Collection; PMR, podocyte migration rate; NMJ, neuromuscular junction; hERG, human ether-a-go-go-related gene; NSAID, nonsteroidal anti-inflammatory drugs; CNS, central nervous system; GI, gastrointestional; ACE, angiotensin-converting enzyme; COX, cyclooxygenase; Pubchem ID: https://pubchem.ncbi.nlm.nih.gov/.

Fig. 8.

Representative example of PMR assay (NIHCC Plate #8) demonstrating wound confluence over time and positive hits. Podocytes from an immortalized human podocyte cell line was seeded 12 h before scratch wound analysis on a 96-well image-lock plate. After a scratch wound was made in the confluent podocyte monolayer, podocytes were exposed to 80 different compounds from NIHCC Drug Plate #8. Each individual drug (80 compounds) was placed in 1 individual well. DMSO vehicle control (0.1%) was seeded in the 16 remaining wells of the 96-well plate. A: wound confluence over time demonstrates podocyte migration over a period of 24 h after creation of scratch wound for all 80 individual compounds. B: 2 specific compounds with PMR outside the <90% CI are shown as positive hits. Drug #49 (mobendazole) and Drug #31 (digoxin) are highlighted in comparison to DMSO vehicle control as they are significantly outside the 90% confidence interval for Δwound confluence/Δtime, a measurement of PMR. Data are expressed as means ± SD for 3 independent experiments.

All 61 compounds were then analyzed in a secondary confirmatory screen at (1, 5, and 10 μM dissolved in 0.01, 0.05, and 0.1% DMSO, respectively) to assess for a dosage effect. In the dose-response screen, 12/61 (19.7%) of the initial positive compounds were confirmed as “distinct” at 10 μM when the podocyte migration rate was compared with the DMSO control condition (Table 2).

Table 2.

Confirmed positive hits by PMR assay from small molecule screen of 725 compounds from NIHCC with decreased PMR

Name	Result on PMR	Mechanism	Pubchem ID	Reference	Drug Library Plate	Concentration, µM
Topotecan (Top)	Decreased migration	Anti-neoplastic topo I inhibitor	46386667	(64, 99)	NIHCC Plate_1_A8	10, 5, 1
Vincristine (VA)	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386588	(57, 87)	NIHCC Plate_1_G8	10, 5, 1
Vindesine (VA)	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386586	(87)	NIHCC Plate_1_H11	10, 5, 1
Digoxin (Dig)	Decreased migration	Na-K-ATPase inhibitor	104170057	(1)	NIHCC Plate_8_G5	10, 5, 1
Podofilox (Top/TI)	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition, DNA-topoisomerase-2 inhibitor	104170173	(54, 59)	NIHCC Plate_9_C8	10, 5, 1
Albendazole (TI)	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition	104170113	(23, 92, 105)	NIHCC Plate_7_C9	10, 5
Lomerizine DiHCl (CaB)	Decreased migration	Ca channel blocker	46386707	(45, 94)	NIHCC Plate_4_A4	10
Pancuronium (MR)	Decreased migration	Nondepolarizing muscle relaxant; competitive acetylcholine antagonist at NMJ	46386853	(29)	NIHCC Plate_4_B3	10
Trazodone hydrochloride (SSRI)	Decreased migration	Antidepressant; binds 5HT2; selective reuptake inhibitor	46386915	(73)	NIHCC Plate_4_B8	10
Tegaserod maleate (5HT)	Decreased migration	5HT4 agonist; used for IBS	46386624	(15)	NIHCC Plate_4_C7	10
Irinotecan HCl (Top)	Decreased migration	Antineoplastic-topo I inhibitor	46386616	(18)	NIHCC Plate_4_D10	10
Mobendazole (TI)	Decreased migration	Tubulin-alpha/beta chain polymerization inhibition	104170137	(72, 85)	NIHCC Plate_8_A8	10
Mitoxantrone (Top)	Decreased migration	DNA intercalation; DNA-topoisomerase-2 inhibitor	104170122	(21, 44)	NIHCC Plate_9_E5	5, 1
Adenosine, N-(2-hydroxycyclopentyl)-,(1S-trans)-(AR)	Decreased migration	Activation of purine receptors A1 and A2	46387015	(110)	NIHCC Plate_2_A6	5
N,N′-diacetyl-1,6-diaminohexane (Syn)	Decreased migration	experimental compound; used in production of nylon	46386869	N/A	NIHCC Plate_2_B2	5
Zolmitriptan (Trip)	Decreased migration	Triptan-selective agonist of serotinin 1B and 1D receptor	46386880	(74, 89)	NIHCC Plate_2_E11	5
Vinorelbine bitatrate (VA)	Decreased migration	Vinca-alkaloid-inhibitor of mitosis	46386815	(57, 87)	NIHCC Plate_4_E4	5
Artesunate (AM)	Decreased migration	Semi-synthetic derivative of artemisinin	46386645	(52)	NIHCC Plate_2_H6	1

Sixty-one preliminary positive results (see Table 1) were screened in serial concentrations of 1, 5, and 10 µM. Confirmed results are above with respective concentrations at which the result was confirmed. All positive hits were screened in quadruplicate at their respective concentration and compared with a vehicle control (0.01, 0.05, and 0.1% DMSO, respectively). NIHCC, National Institutes of Health Clinical Collection; PMR, podocyte migration rate; AM, anti-microbial; AR, antiarrhythmic; CaB, calcium channel blocker; Dig, digitalis glycoside; IBS, irritable bowel syndrome; MR, muscle relaxant; NMJ, neuromuscular junction; Pubchem ID: https://pubchem.ncbi.nlm.nih.gov/; SSRI, selective serotonin reuptake inhibitor; Syn, synthetic compound; Top, topoisomerase inhibitor; Trip, triptan agonist; TI, tubulin inhibitor; VA, vinca-alkaloid.

Five compounds, topotecan, vincristine, vindesine, digoxin, and podofilox, demonstrated significantly slower PMRs when compared with control at all three dosing concentrations. Two compounds, albendazole and mitoxantron, demonstrated a dosage effect with significantly slower migration at two of the three dosing concentrations.

All compounds that were confirmed reduced PMR as compared with the DMSO vehicle control. No compounds in the drug screen significantly increased PMR as compared with control.

The positive hits clustered into two classes of compounds: vinca-alkaloids (vincristine and vindesine) and the azole class of antifungals (albendazole and mobendazole) (Table 2). When inspected visually, both classes of medications alter migration rate as well as cell morphology (Fig. 9). These drugs commonly affect microtubule assembly formation as destabilizing agents. Furthermore, podofilox has microtubule depolymerization activity as well.

Fig. 9.

Podocyte cellular morphology at sequential time points after drug exposure. Human podocytes were qualitatively examined for podocyte morphology changes during the podocyte migration assay. Live cell imaging captured immortalized human podocytes at different time points (0, 5, 10, and 15 h) after drug exposure, while undergoing migration in a scratch wound assay. Compounds shown include DMSO vehicle control (0.1%), topotecan (10 μM), digoxin (10 μM), albendazole (10 μM), and podofilox (10 μM). All compounds significantly affected PMR as well (see Table 2).

Other antineoplastic agents that reduced PMR significantly include the topoisomerase inhibitors (topotecan and irinotecan). Other single compounds that decrease PMR include ligands for serotonin receptors (trazodone and tegaserod), a calcium channel blocker (lomerizine), a nondepolarizing muscle relaxant (pancuronium), and digoxin, a glycoside used for cardiac arrhythmias (Table 2).

DISCUSSION

Implications of positive results.

Recent technological advances have allowed high throughput screening of small molecules and compounds in kidney diseases, which has traditionally been lacking in comparison with other medical fields such as oncology. Although other publications have established potential screens for drug therapy in kidney diseases (68), live cell imaging has not previously been employed to screen for potential therapeutics. In this drug screen, we used PMR as a relevant surrogate phenotype of disease to confirm 12/61 (19.6%) of compounds at 10 μM that we initially screened as positive hits.

Of the five compounds that demonstrated a PMR that was significantly reduced compared with vehicle control at all concentrations (1, 5, and 10 μM), three of the compounds (vincristine, vindesine, and podofilox) alter microtubule formation by acting as depolymerization agents. Furthermore, two other compounds, albendazole and mobendazole, that confirmed at 10 μM act as microtubule inhibitors as well. These compounds demonstrate that drugs that affect microtubule assembly are robust effectors of PMR. Although many of the podocytic genes that cause SRNS alter actin regulation, microtubules are a critical component in the podocyte and functionally cooperate with the actin cytoskeleton (43, 61, 63). Made from sets of 13 protofilaments of α- and β-tubulin subunits, microtubules form the primary processes in the podocyte and have been shown to be important in supporting structural development of the podocyte cytoskeleton (5, 107). It was shown previously that podocytes express several microtubule-associated proteins (MAPs) such as MAP3 and MAP4 (60). It is also known that CHO1/MKLP1 and protein phosphatase 2a (PP2A) microtubule-associated motor proteins (MAPs-MP) play an important role in podocyte primary process formation and in establishing of cell polarity in cultured podocytes (62, 63). In addition, microtubules (MT) are involved in positioning and organization cell organelles inside the cell and are responsible for intracellular transport of vesicles and proteins between cell domains (113). It is also known that impairment of these functions results in a transient nephrotic syndrome in cell culture (58). Knowing that the MT and the actin cytoskeleton interact, we can assume that impairment of one component can lead to a dysfunction of the other. This is supported by cell culture studies showing that the microtubule-associated guanine nucleotide exchange factor GEF-H1 regulates actin cytoskeleton dynamics through activation of RhoA (78). In fact, there are multiple genes (WDR73, INF2, and TTC21B) that, if mutated, implicate abnormal microtubule assembly in the pathogenesis of nephrotic syndrome (25, 53, 101) confirming pathogenic relevance of our findings. Therefore, this screen confirms that targeting the pathway of microtubule assembly dynamics may be worth pursuing in the search for novel therapeutics in nephrotic syndrome.

Limitations.

One limitation to our assay is the fact that we only screened compounds of the NIHCC in wild-type immortalized podocytes. Drugs may have an effect as a therapy only in diseased states. Another issue in using wild-type podocytes is that identification of drugs that increase PMR significantly is difficult to detect. As immortalized cultured wild-type cells migrate rapidly at baseline, our assay may not have the sensitivity to detect drugs that increase PMR in a biologically significant manner with our defined criteria. It would be potentially interesting to use the assay we established in this study to screen compounds affecting relevant pathways in cell lines expressing different monogenic defects of SRNS.

Furthermore, the drug library obtained may not have been targeting the right pathways for finding compounds that affect nephrotic syndrome. It has been demonstrated that small Rho-like GTPases alter PMR and that their function is altered in nephrotic syndrome (39–41) and we demonstrated an effect of Rho and Rac1 activators/inhibitors on PMR (Fig. 3). The NIHCC small molecule library was curated due to its clinical availability for resupply and not for a specific pathway such as small Rho-like GTPases; hence, we expected a low hit rate in the screen. Future studies would be strengthened if specific libraries were used to target nephrotic syndrome relevant pathways in the podocyte such as those that regulate the small Rho-like GTPases RhoA/Rac1/Cdc42.

Future directions.

Confirmed hits in our screen are reliable and reproducible, and the results can be further applied to screen for pathway specific therapies of nephrotic syndrome. In establishing the scratch-wound assay as a relevant screening method for treatment of nephrotic syndrome, several potential applications are apparent. One application includes the study of pathway specific defects in nephr otic syndrome by using targeted small molecule libraries rather than large nonspecific small molecule screens. Another possibility includes screening cell lines with defects in monogenic causes of nephrotic syndrome. In the era of personalized medicine, this assay could potentially be used in conjunction with CRISPR/Cas9 technology to generate cell culture lines with genes that have allele-specific loss of function to screen for therapeutics. In summary, in this study, we have established a high-throughput assay that uses PMR and live cell videomicroscopy to identify modifiers of microtubule dynamics as a potential pathway for treatment in nephrotic syndrome. This assay also opens a new avenue for growth in research of drug discovery in kidney disease.

GRANTS

This research was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-076683 (to F. Hildebrandt). W. Tan is supported by NIH T32 Training Grant T32-DK-007726–31A1. E. Widmeier is supported by the German National Academy of Sciences Leopoldina (LPDS-2015–07).

DISCLOSURES

F. Hildebrandt receives royalties from CLARITAS.

AUTHOR CONTRIBUTIONS

E.W., W.T., M.A., and F.H. conception and design of research; E.W. and W.T. performed experiments; E.W. and W.T. analyzed data; E.W., W.T., and F.H. interpreted results of experiments; E.W. and W.T. prepared figures; E.W. and W.T. drafted manuscript; E.W., W.T., M.A., and F.H. edited and revised manuscript; E.W., W.T., M.A., and F.H. approved final version of manuscript.