Differentiation of human iPSCs into functional podocytes

Podocytes play a critical role in glomerular barrier function, both in health and disease. However, in vivo terminally differentiated podocytes are difficult to be maintained in in vitro culture. Induced pluripotent stem cells (iPSCs) offer the unique possibility for directed differentiation into mature podocytes. The current differentiation protocol to generate iPSC-derived podocyte-like cells provides a robust and reproducible method to obtain podocyte-like cells after 10 days that can be employed in in vitro research and biomedical engineering. Previous published protocols were improved by testing varying differentiation media, growth factors, seeding densities, and time course conditions. Modifications were made to optimize and simplify the one-step differentiation procedure. In contrast to earlier protocols, adherent cells for differentiation were used, the use of fetal bovine serum (FBS) was reduced to a minimum, and thus ß-mercaptoethanol could be omitted. The plating densities of iPSC stocks as well as the seeding densities for differentiation cultures turned out to be a crucial parameter for differentiation results. Conditionally immortalized human podocytes served as reference controls. iPSC-derived podocyte-like cells showed a typical podocyte-specific morphology and distinct expression of podocyte markers synaptopodin, podocin, nephrin and WT-1 after 10 days of differentiation as assessed by immunofluorescence staining or Western blot analysis. qPCR results showed a downregulation of pluripotency markers Oct4 and Sox-2 and a 9-fold upregulation of the podocyte marker synaptopodin during the time course of differentiation. Cultured podocytes exhibited endocytotic uptake of albumin. In toxicological assays, matured podocytes clearly responded to doxorubicin (Adriamycin™) with morphological alterations and a reduction in cell viability after 48 h of incubation.


Introduction
Podocytes, also known as visceral epithelial cells, play a key role in the glomerular filtration barrier and the maintenance of glomerular function [1,2]. Podocyte injury is involved in many a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 human kidney diseases like membranous glomerulopathy and diabetic nephropathy [3]. Due to their relevance in the initiation and progression of nephropathies, podocytes have gained increased attention for their potential role in chronic kidney diseases [4,5].
Podocytes are highly specialized, terminally differentiated cells playing a pivotal role in maintaining the glomerular filtration barrier and producing growth factors for surrounding cells, namely mesangial and endothelial cells [6,7]. They sustain their filtration barrier by wrapping around the glomerular capillaries with interdigitated foot processes, which are connected through intercellular junctions, known as the slit diaphragm. The slit diaphragm facilitates the passing of the primary urinary filtrate and is not traversable for high-molecularweight plasma proteins [8]. At their apical side podocytes face the primary urine and at their basolateral side they are connected to the glomerular basement membrane via integrins and dystroglycans. The glomerular basement membrane is mainly composed of type IV collagen, laminin, and sulfate proteoglycans which is completed by a fenestrated endothelium [9].
In their mature, fully differentiated state, podocytes have a limited capacity to proliferate in vivo, making them vulnerable to various insults leading to glomerular sclerosis [10]. The limited proliferation capacity is further reflected in vitro, making it difficult to obtain adequate quantities of podocytes at a differentiated stage for basic research [11][12][13]. Due to these facts, only a limited number of podocyte cell lines of animal and human origin are available [14,15]. The use of primary podocyte cultures is also limited due to dedifferentiation over time [16]. However, there is a clear need for meaningful podocyte culture systems [17][18][19].
A conditionally immortalized human podocyte cell line has been established by transfection of podocyte primary cells with a temperature-sensitive SV40 large-T antigen gene construct [15]. However, differentiation of cells at the non-permissive temperature of 37˚C bears phenotypic variations within different podocyte cell isolates [20]. The development of new technologies to reprogram somatic cells into the pluripotent state, so-called induced pluripotent stem cells (iPSCs) [21], opened new avenues to re-differentiate iPSC into tissue cells of all 3 germ layers [22]. Thus, iPSCs generated from adult somatic cells provide an almost unlimited supply of human cells without ethical concerns known for embryonic stem cells. iPSCs could be used in biochemical research for disease modelling, in the development and screening of drugs, and in toxicity testing, which may lead to promote personalized medicine and future applications in regenerative medicine for cell replacement therapy [23][24][25][26][27][28][29]. Furthermore, iPSCs offer exciting opportunities to create kidney organoids that enable in vitro podocyte research [30][31][32][33].
In this respect, the directed differentiation of human iPSCs into glomerular podocytes was described in two recent studies [34,35]. Song et al. [35] used a 10 day directed differentiation with an intermediate suspension culture of mechanically dissociated cells, while Ciampi et al. [34] applied a three-stage protocol including induction into intermediate mesoderm, commitment towards nephron precursors, and specification into podocytes. iPSC-derived podocytes were characterized by the expression of podocyte-specific markers, the endocytic internalization of albumin, and the disappearance of pluripotent markers Oct3/4 and Sox-2.
The aim of the current study was to evaluate the reproducibility and robustness of currently available podocyte differentiation protocols and to optimize the protocols accordingly. Here, we report on a direct differentiation of human iPSCs into functional podocytes, based on the protocols of Ciampi et al. [34] and Song et al. [35]. A modified, robust and reproducible differentiation protocol is described, tailored to different human iPSC lines, generated from adult and neonatal donors, that were reprogrammed by Sendai virus [36,37] during the course of the IMI-funded StemBANCC project (http://stembancc.org) [23]. iPSC-derived podocytes exhibited distinct morphological features of podocyte foot processes, and expressed the podocyte-specific markers synaptopodin and nephrin. Beyond recent studies [34,35]

Results
The differentiation protocol developed in the present study is based on a direct path to mature podocyte-like cells without intermediate progenitor stages, using a combination of BMP7, activin A and retinoic acid. In modification of earlier protocols [34,35], the amount of fetal bovine serum (FBS) in culture media was reduced to a minimum, and thus β-mercaptoethanol (β-ME) could be omitted. In studies on chemically defined culture conditions for human iPSCs it was shown, that β-ME is toxic when high molecular weight protein components, like albumin (BSA), are removed from the culture medium [38]. Without β-ME, BSA was no longer necessary for human iPSC culture, suggesting that β-ME is inactivated by unspecific protein binding. Therefore, by decreasing the content of FBS down to 1.25% (v/v) in the present culture media, β-ME was no longer added. Further modifications that were made to the protocol include the omittance of the suspension cultures on days 1-3. In the current protocol the transfer of cells from ultra-low cluster plates to gelatin-coated dishes was no longer needed. This enabled a more standardized procedure since the seeding of a precise number of cells was now possible. The seeding density was actually one of the main factors contributing to the reproducibility of the present protocol.

The initial seeding density is crucial for differentiation efficiency
In systematically elaborated culture experiments it was found, that the plating densities of precultures of undifferentiated iPSCs as well as the seeding density of the differentiation cultures are absolutely crucial for a stable and reproducible differentiation of iPSCs into podocyte-like cells. Several initial iPSC plating densities of undifferentiated cells as well as several cell seeding densities for differentiation cultures were tested. Initial plating of undifferentiated iPSC density was tested at 50-100,000, 200,000 and 400,000 cells/cm 2 . After 4 days in culture, iPSCs were detached by Accutase to yield single cell suspensions to start differentiation into podocyte-like cells. Cells were counted and 9,000, 15,000 and 40,000 cells/cm 2 were then plated on Geltrex-coated multi-well plates for differentiation experiments. The results of the initial plating experiments are shown in Fig 1. The most efficient seeding densities for generating iPSCderived podocytes were 50-100,000 cells/cm 2 for the undifferentiated iPSC pre-cultures, and after 4 days in culture, 9000 cells/cm 2 for efficient differentiation into podocyte-like cells, respectively. The elaborated plating protocol ensured optimal growth conditions for a maximum of evenly distributed cells at appropriate density for cell enlargement and development of podocyte-like structures, respectively. It also ensured sufficient cell numbers and biomass to be obtained for down-stream experiments. These seeding densities were further critical for the reproducibility and robustness of the protocol, as higher seeding densities of undifferentiated iPSCs at 200.000 and 400.000 cells/cm 2 resulted in significantly lower numbers of podocytelike cells or no differentiation into podocyte-like cells, since cells did not attach on day 1 (Fig  1). This protocol was successfully repeated with three additional iPSC lines (SBAD2, SBNEO and SFC018; see below).
The time course of differentiation of iPSCs into podocyte-like cells was monitored by determination of mRNA expression profiles of the pluripotency markers Oct4 and Sox-2, and of synaptopodin, a podocyte-specific marker, by quantitative real-time PCR (qPCR). As depicted in Fig 2, characteristic time courses of the specific markers were found. While Oct4 and Sox-2 gradually decreased to zero during the first 9 days of differentiation, synaptopodin mRNA expression levels gradually increased 9-fold after 17 days of cells in medium M2.

Morphology of iPSC-derived podocyte-like cells
The temporal changes of iPSC during differentiation towards a podocyte-like morphology were assessed by phase-contrast microscopy (Fig 3). Cells changed from a typical iPSC colony morphology, consisting of small rounded cells, to enlarged cells with the characteristic foot processes. After ten days of differentiation, cells clearly showed long spindle like projections, short rounded projections, fine processes, and interdigitations between cells. Differentiated podocyte-like cells could be maintained in culture for at least further 10 days. Conditionally immortalized human podocytes [15] served as reference controls. When these cells were expanded at the permissive temperature of 33˚C, they remained in undifferentiated, epithelial-like state (not shown); however, when transferred to 37˚C, they started to differentiate into distinct podocyte morphology with well-preserved foot processes and synaptopodin expression (Fig 4). Thus, both iPSC derived podocyte-like cells and conditionally immortalized human podocytes showed comparable podocyte-specific morphologies after terminal differentiation.

Expression of podocyte-specific markers
Aside from the distinct podocyte-specific morphology, the protein expression of podocytespecific markers by differentiated iPSC-derived podocyte-like cells and human immortalized  podocytes, respectively, was determined by immunofluorescence and Western blot anaylsis. Synaptopodin, an actin-associated protein (see also Fig 2), podocin and nephrin, both raftassociated components of the glomerular slit diaphragm, and the transcription factor WT-1 could be detected by immunofluorescency microscopy (Fig 5, S1 and S2 Figs). While cellular location of synaptopodin, podocin and WT1 were as expected, nephrin staining appeared in the cytoplasm and was not localized within the slit diaphragm of the podocytes. Since synaptopodin binds specifically to actin in the foot processes of glomerular podocytes [14], F-actin cytoskeletal bundles were visualized in differentiated podocyte-like cells with fluorescentlabeled phalloidin [14,39]. Representative micrographs are shown in Fig 5 and S1 and S2 Figs. For WT1, which is also expressed in undifferentiated iPSC, expression levels were maintained during the differentiation time course. Oct3/4 protein expression declined to undetectable levels at day 6. This was comparable with mRNA levels as measure via PCR (Fig 2).
In order to test for podocyte-specific functions of differentiated cells, two functional assay regimes were applied: (a) the endocytic uptake of albumin and (b) the susceptibility of podocyte-like cells to doxorubicin (Adriamycin™).

Endocytosis of albumin by podocyte-like cells
The ability of endocytic uptake of FITC-labeled albumin was investigated by fluorescence microscopy. Serum-starved, fully differentiated podocyte-like cells were incubated with 1 mg/ ml FITC-conjugated bovine serum albumin for 1 h at 4˚C and 37˚C, respectively, and for 24 h at 37˚C. As shown in Fig 7, fluorescence microscope images show a clear temperature-dependent endocytosis of albumin after 1 h and 24 h of incubation at 37˚C. No endocytic uptake of labeled albumin was observed at 4˚C for 1 h. Podocytes are a specific target for doxorubicin (Adriamycin) Doxorubicin (DOX), a nephrotoxic anti-cancer agent is, aside its cardiotoxicity, known to be specifically toxic for glomerular podocytes. DOX leads to podocyte foot process effacement and increased glomerular permeability causing proteinuria. DOX-associated nephropathy is caused by interfering with actin reorganization of podocytes [40] and mitochondrial dysfunction [41]. Therefore, the impact of DOX on differentiated podocyte-like cells was used as another functional endpoint. Cultures of iPSC-derived differentiated podocyte-like cells and of undifferentiated and differentiated human immortalized podocytes were incubated with DOX for 48 h at a concentration of 10 nM and 1 μM (Fig 8). While 10 nM DOX showed no effect on either cell type, 1 μM DOX caused a drastic decrease in cell proliferation rates in differentiated podocyte-like cells, whereas undifferentiated human podocytes were unaffected, indicative for the susceptibility of differentiated, mature podocytes to DOX intoxication.

Discussion
Podocytes are key cells in the formation of the glomerular filtration barrier. In the complex filtration barrier network of the glomerulus, podocytes are the most vulnerable components. Most glomerulopathies are associated with podocyte injury causing various types of diseases, that may account for approx. 90% of end-stage kidney disease [42]. When podocytes are injured, their main function for sieving macromolecules cannot be longer maintained. Disruption of the complex podocyte structure or reduction in the number of podocytes results in proteinuria [43]. Podocyte disruption or loss are involved in renal diseases like focal segmental glomerulosclerosis (FSGS) or diabetic nephropathy [5,6].
Thus, the central role of podocytes in glomerular dysfunctions necessitates reliable model systems to study normal and impaired podocyte functions at the cellular and subcellular level. In this respect, in vitro cultures of human podocytes gained specific importance. However, there are intrinsic difficulties of growing differentiated podocytes in in vitro culture, that have been discussed extensively [11][12][13]. Mature podocytes are terminally differentiated cells. This is a profound limitation in cell culture, since growth arrested, differentiated cells do not proliferate and obtaining cells in sufficient quantities for basic research is extremely difficult. To circumvent this problem, undifferentiated podocyte "precursor" cultures should be expanded to a certain cell density, and then be differentiated by specific stimuli. Two strategies were pursued: 1. conditionally immortalized podocytes of human origin and 2. induced pluripotent human stem cells (iPSCs). The first strategy was applied using a conditionally-immortalized human podocyte cell line carrying a temperature-sensitive mutant of the SV40 large-T antigen [15]. Thereby the cells can be expanded at the permissive temperature of 33˚C to an appropriate cell density. Thereafter, when switched to the non-permissive temperature of 37˚C, the cells enter growth arrest and start to differentiate into mature podocytes, exhibiting foot processes and expressing synaptopodin, nephrin and podocin (Fig 4). However, although conditionally immortalized, these cells were only usable for approx. 20 passages, when new human glomerular primary cultures had to be initiated, expanded and transfected/immortalized for further propagation [13,15].

iPSC-derived podocytes
The second strategy is the employment of human iPSC. In 2007 Takahashi and Yamanaka discovered a technique to reprogram adult human cells into pluripotent stem cells [21], which enables a non-controversial, ethical acceptable method to differentiate pluripotent cells into any human somatic cell type. iPSCs are, by definition, immortal and can be expanded indefinitely. This paved new ways for investigating diseases ("disease in a dish") [23,24,27,28,44], for in vitro toxicity testing and drug safety assessment, and for tissue engineering and cell therapy applications [25,29,45].
Furthermore, human iPSC opened the possibility of directed differentiation into all cell types of the nephron [30]. Recent strategies follow the in vivo stages of embryonic kidney development, with differentiation of iPSC into intermediate mesoderm and metanephric mesenchyme, and the induction of nephron progenitor cells, thereby recapitulating embryonal developmental stages of the kidney in vitro [31,46,47].
Here we present a method to directly differentiate human podocytes from human induced pluripotent stem cells (iPSCs) in a one-step protocol, using a cocktail of retinoic acid, activin A, and BMP7. Already published protocols [34,35] were modified for adherent cell culture systems. As mentioned above, for the present protocol the culture medium content of FBS was lowered to a minimum of 1.25% (v/v) in differentiation medium M1 and 2,5% (v/v) in maintenance medium M2, respectively, and therefore β-ME could be omitted.
Characteristic podocyte morphology, such as long spindle-like projections, fine processes, short rounded projections, and interdigitations between cells could be observed on day 10 of differentiation or earlier (Fig 3). The immunofluorescence expression of podocyte markers synaptopodin, an actin-associated protein, podocin and nephrin, components of the slit diaphragm, and of the transcription factor WT1 could be shown (Fig 5, S1 and S2 Figs). Podocin, nephrin and WT1 could be further verified by Western blot analysis (Fig 6 and S3 Fig) and synaptopodin by qPCR (Fig 2) at different time points.
qPCR results showed a gradual decrease in mRNA levels of pluripotency markers Oct4 and Sox-2, paralleled by an increase in podocyte marker synaptopodin mRNA during iPSC differentiation (Fig 2). Oct4 and Sox-2 are specific markers for the pluripotent state of embryonic stem cells. The essential role of Oct3/4 and Sox-2 in the maintenance of the pluripotent state is further substantiated by the crucial role of the transcription factors in the reprogramming of somatic cells into iPSCs [21]. Podocin and nephrin also showed a differentiation induced increase in expression at a protein level. WT1 however, is also expressed in undifferentiated iPSC and its expression was maintained during differentiation.
Variations between iPSC lines are well documented. Regardless of the source of somatic cells or the method chosen, the reprogramming process itself is influenced by a number of variables that affect the reproducibility and quality of the resulting iPSCs. Each iPSC line can be considered to be a distinct clone [48]. The reproducibility and robustness of the present protocol was proven on four different iPSC lines developed during the course of the IMI-funded StemBANCC project (http://stembancc.org) [23]. iPSCs from reprogrammed fibroblasts of adult (SBAD2 and SBAD3) and neonatal origin (SBNEO), as well as of a migraine patient (SFC018), respectively, were differentiated resulting in well comparable podocyte-like cells, in terms of morphology (Fig 3) and expression of podocyte-specific markers (Figs 5 and 6). Since iPSCs can be obtained from any individual [29], the generation of iPSC from patients with e.g. genetic diseases [45] offers potential opportunities for human disease modelling. Patient-specific iPSCs enable in the future to study the pathophysiology of the disease as well as the response to drugs in vitro at the cell level [24,27].
Another major finding of the present study was that the plating density of the undifferentiated iPSC stocks, as well as the seeding density for differentiation, respectively, are absolutely crucial for a sufficient and proper differentiation efficiency. Elaborated plating protocols were mandatory for differentiating evenly distributed iPSC into consistently spread podocytes at well-defined cell densities. An initial plating density of 50-100,000 cells/cm 2 was optimal for pre-cultures of undifferentiated iPSC. High cell numbers in dense cultures of iPSC result in rapid acidification of the mTeSR1 medium. We have recently shown that high cell densities can decrease the culture medium pH to 6.8 within 24 h. At this low pH, undifferentiated iPSC enter a transient growth arrest in G0/G1-phase [49]. This is likely to have an impact on subsequent differentiation and should therefore be avoided with low initial seeding densities. A subsequent seeding of 9,000 cell/cm 2 were proven to be most efficient for terminal differentiation (Fig 1). Cells need space for proper differentiation. When iPSC differentiate into podocyte-like cells, they substantially increase in size (Fig 3) and exhibit a more complex morphology with slit diaphragm-like cell contacts (Fig 4). Overcrowding of undifferentiated cells is one of the major pitfalls in podocyte culture [11]. For growth-restrictive conditions of immortalized podocytes, the authors also recommend plating densities of 5,000 to 10,000 cells/cm 2 .
The differentiation protocol applied in the present study involved an inducer cocktail, containing retinoic acid, BMP7, and activin A. Retinoic acid is known to have an anti-proliferative and pro-differentiation effect in many tissues, including the kidney and especially the podocyte [31,50], and has a clear impact on the podocyte phenotype in vitro and in vivo [51]. It is crucial for differentiation towards intermediate mesoderm and regulates the expression of the transcription factor WT-1, which is required for the expression of several podocyte genes [52]. WT-1 is only expressed in podocytes of the mature kidney [7,53]. BMP7 acts on podocytes in an autocrine, paracrine and endocrine manner, it phosphorylates Smad5, which is involved in podocyte differentiation and survival. It is expressed in many tissues during development including podocyte precursor cells [54]. BMP7 is a potent inducer of OSR1 expression, a specific intermediate mesoderm marker [55]. Activin A is needed for induction of intermediate mesoderm [56].
The ability of endocytotic uptake of albumin is a well described feature of differentiated podocytes [57,58]. Albumin endocytosis by differentiated podocytes was investigated by incubation of cells with FITC-labelled albumin for 24h at 37˚C and for 1h at 37˚C (uptake) and 4 C (no uptake). After 1h and 24h of incubation at 37˚C the specific, temperature-dependent endocytic uptake of albumin could be detected via fluorescence (Fig 7).
In functional assays, doxorubicin, an anti-cancer chemotherapeutic drug, showed a toxic effect on differentiated iPSC-derived podocytes. Doxorubicin at 1 μM resulted in a 50% reduction in cell viability of iPSC-derived podocyte-like cells after 48h of exposure. This was comparable to immortalized podocytes cultured under non-permissive, but not permissive conditions (Fig 8). Recently, doxorubicin-induced podocyte injury was also demonstrated in a glomerulus-on-a-chip model using iPSC-derived podocytes [59].
To summarize, the present investigation demonstrates a direct method to generate podocyte-like cells from human iPSCs. This represents a minimal invasive and ethically accepted way to obtain human podocytes in appropriate quantities for basic research and offers new possibilities for in vitro toxicity studies and drug safety assessment. Control and patient iPSCderived mature human podocytes in monolayer culture may also facilitate new avenues for drug development and development of personalized therapeutic strategies.

Cell culture
Human iPSCs used in the present study were obtained during the course of the IMI-funded StemBANCC project (http://stembancc.org) [23]. Human fibroblasts of adult (SBAD3 clone 1 and SBAD2 clone 1) and neonatal (SBNEO clone 1) origin (Lonza) and primary fibroblasts of a migraine patient (SFC018 clone 03-01) were reprogrammed by non-integrating Sendai virus transfection [36,37,60] according to project-specific SOPs. Reprogramming efficiency and cell status was periodically verified by the iPSC producing laboratory. Culture and propagation of iPSC adhered to specific SOPs that have been elaborated within the StemBANCC consortium. iPSC were routinely cultured on matrigel-coated plates (Geltrex, GIBCO) in mTeSR™1 medium (StemCell Technologies). Cultures were passaged twice a week with Versene (Lonza) at a split ratio of 1:4 to 1:12.
Periodical cell authentication of the four iPSC lines used in the present study was performed by short tandem repeat (STR) profiling [61] as described recently [49]. In brief, cellular DNA was isolated (Qiagen Puregene Cell Kit, Cat.No. 158745) and STR profiles were analyzed with the Promega PowerPlex 16 HS System in an Applied Biosystems Genetic Analyzer and GeneMapper ID-X 1.2 software.

Differentiation of iPSC into functional podocytes
The critical role of initial cell density. During a series of experiments it was observed, that the initial cell densities of stock iPSCs as well as the seeding densities of the differentiation cultures play a critical role in order to obtain reproducible results in the differentiation of iPSC into podocytes. To this end, pre-cultures of undifferentiated iPSC were plated at densities of 50-100,000, 200,000 or 400,000 cells/cm 2 . After 4 days, cultures were passaged with Accutase (GIBCO) to obtain single cell suspensions, and iPSCs were seeded on plates coated with Geltrex at a density of 9,000, 15,000 and 40,000 cells/cm 2 for differentiation. In order to avoid uneven concentric distribution of cells in multiwell plates, and thus non-optimal conditions for a proper development of foot processes and interdigitations between adjacent/neighbouring cells, cells were seeded with half of the recommended volume of media per well. Back and forward and left and right gentle shaking movements were performed instead of gentle rotation before placing the plates into the incubator. The residual media was added after 5h, when cells were attached.

Quantitative real-time PCR (qPCR)
Total RNA was isolated with QIAzol (QIAGEN, Cat. No. 79306) from undifferentiated iPSC cultures, and from podocyte-like cell cultures at day 1, 3, 6, 9, 13 and 17 in differentiation media. 50 ng of total RNA were used for first-strand cDNA synthesis with Superscript II (Life Technologies), following the manufacturer's instructions.

Albumin uptake assay
Differentiated podocyte cultures were serum-starved overnight. After rinsing the monolayers with PBS, cultures were incubated with 1 mg/ml FITC-conjugated bovine serum albumin for 1 h at 4˚C (albumin binding only) and 37˚C (binding and endocytosis), respectively, and for 24 h at 37˚C. Albumin uptake was determined by fluorescence microscopy, using a ZEISS Axiophot Fluorescence Microscope.

Doxorubicin toxicity and cell viability assays
In doxorubicin experiments, adult, neonatal, and migraine patient-derived differentiated podocyte-like cells, and undifferentiated and differentiated human immortalized podocytes were incubated for 48 h at concentrations of 10 nM and 1 μM. The resazurin assay was used to assess metabolic activity of the cultures [63]. Resazurin is reduced to resorufin by viable cells. Thus, resazurin reduction is directly proportional to the number of viable cells. Resazurin solution was prepared as a 20-fold stock in PBS from resazurin salt (SIGMA, R7017) and further diluted in culture media. Cultures were incubated for 2 h. The supernatants were transferred into 96 wells for measurement at an excitation wavelength of 540 nm and an emission at 590 nm in a Tecan GENios plus microplate reader.

Statistical analysis
All experiments were conducted at least three times. Statistical significance was tested using the unpaired Student's t test. P values <0.05 were considered statistically significant. See legends for more details.