The state-of-the-art cultured podocyte is conditionally immortalized by expression of a temperature-sensitive mutant of the SV40 large-T antigen. These cultures proliferate at 33°C and differentiate at 37°C into arborized cells that more closely resemble in vivo podocytes. However, the degree of resemblance remains controversial. In this study, several parameters were measured in podocyte cell lines derived from mouse (JR, KE), human (MS), and rat (HK). In all lines, the quantities of NEPH1 and podocin proteins and NEPH1 and SYNPO mRNAs were comparable to glomeruli, while synaptopodin and nephrin proteins and NPHS1 and NPHS2 mRNAs were <5% of glomerular levels. Expression of Wilms' tumor-1 (WT1) mRNA in mouse lines was comparable to glomeruli, but rat and human lines expressed little WT1. Undifferentiated human and mouse lines had similar proliferation rates that decreased after differentiation, while the rate in rat cells remained constant. The motility of different lines varied as measured by both general motility and wound-healing assays. The toxicity of puromycin aminonucleoside was MS ∼ JR >> KE, and of doxorubicin was JR ∼ KE > MS, while HK cells were almost unaffected. Process formation was largely a result of contractile action after formation of lamellipodia. These findings demonstrate dramatic differences in marker expression, response to toxins, and motility between lines of podocytes from different species and even between similarly-derived mouse lines.
- nephrotic syndrome
- puromycin aminonucleoside
in the kidney glomerulus, the visceral epithelial cells (podocytes) are attached to the basement membrane that surrounds the glomerular capillary by processes that ramify to form a fine network of tertiary (“foot”) processes. The nature of this attachment precludes direct isolation of in vivo podocytes, so biochemical and cell biological studies of podocytes depend upon cell cultures. Primary cultures can be derived from cells growing from isolated glomeruli or cultured from urine, but most current studies with cultured podocytes use conditionally immortalized mouse or human podocyte cell lines that express temperature-sensitive variants of the simian virus 40 large T-antigen (tsTAg), allowing investigators to proliferate cells at the 33°C permissive temperature and to obtain differentiated cells by culture at the 37–39°C nonpermissive temperature (reviewed in Ref. 22).
Despite the popularity of conditionally immortalized cultured podocytes, the degree to which they resemble in vivo podocytes remains controversial. The quantity of specific podocyte marker proteins such as nephrin expressed by cultured podocytes has been questioned, to the extent that one group determined culture conditions necessary to recover nephrin expression in murine cultures (41). Whether the cultured podocytes derived from cobblestone-like outgrowths of glomeruli are of visceral (podocyte) or parietal (Bowman's capsule) epithelial origin has also been questioned (50), a caveat particularly interesting in light of the elegant study of Appel et al. (1) demonstrating that parietal epithelial cells can become podocytes in vivo.
These caveats have been extensively rebutted in a recent review (39) that contained a table, labeled in part, “. . . if they look, smell, and act like a podocyte, they are probably podocytes.” Three of the criteria cited in support of cultured podocytes were the presence of slit diaphragm-like structures, stable expression of the majority of podocyte-specific proteins, and the quiescent state of differentiated cultures. Since few of the studies cited made quantitative measurements of these parameters, and no studies directly compared different lines of cultured podocytes, this study was performed to compare four different conditionally immortalized cultures from three different species. Two murine podocyte lines produced from the Immortomouse (16) by Mundel et al. (28) in 1996 (referred to as “JR” in this manuscript) and Schiwek et al. (38) in 2004 (“KE”), a human podocyte line produced through retroviral insertion of a sequences encoding a tsTAg and telomerase by Saleem et al. (37) in 2002 (“MS”), and clone C7 of a series of rat cell lines isolated from a tsTAg (tsA58) transgenic rat by Kurihara (9), with results from clone D7 later published in 2007 (“HK”) were examined. We determined specific culture conditions to standardize cultures, then measured mRNA and protein expression of podocyte markers, rates of cell division and motility, and the response of cultures to puromycin aminonucleoside (PAN) and doxorubicin (DOX), chemicals commonly used to induce animal models of experimental nephrotic syndrome. These results provide direct comparisons between different podocyte cultures and the expression of podocyte markers in vivo and in vitro, and in addition define the genesis of the major cell processes of cultured podocytes.
MATERIALS AND METHODS
PAN, DOX, and thiazolyl blue tetrazolium bromide (MTT) were from Sigma-Aldrich (St. Louis, MO). An LDH Cytotoxicity Assay Kit was from Takara Bio (Otsu, Shiga, Japan, distributed by Thermo Fisher Scientific, Waltham, MA). All tissue culture plastic ware (Cellstar) was from Greiner Bio-One (Kremsmuenster, Austria, distributed by Thermo Fisher Scientific). Cell culture media, antibiotics, fetal bovine serum, and TRIzol were from Invitrogen (Carlsbad, CA). An RNeasy mini-kit was from Qiagen Sciences (Valencia, CA). γ-Interferon was from R&D Systems (Minneapolis, MN). Interferon-transferrin-selenium solution was from Mediatech (Manassas, VA). Antibody against nephrin (guinea pig polyclonal) was from Acris Antibodies (Hereford, Germany, distributed by Novus Biologicals, Littleton, CO), antibody against fascin was from EMD Millipore (Temecula, CA), and antibodies against podocin (H-130), synaptopodin (P-19), and NEPH1 (H-150) were from Santa Cruz Biotechnology (Santa Cruz, CA). Secondary antibodies conjugated to horseradish peroxidase were from JacksonImmunoresearch (West Grove, PA). Bovine fibronectin, murine laminin I and collagen IV, and Engelbreth-Holm-Swarm basement membrane extract were from R&D Systems. Formaldehyde was from Electron Microscopy Sciences (Hatfield, PA), and Texas Red-X phalloidin was from Invitrogen.
General culture conditions.
Collagen (type I) was isolated and purified from rat tail tendons. Ethanol-sterilized collagen was collected, lyophilized, dissolved in 0.15 M acetic acid at 1 mg/ml, and stored frozen at −80°C until use. Tissue culture plastics were coated with collagen by incubation for 2 h at room temperature with solutions of collagen or other extracellular matrix components prepared in 1× PBS immediately before use to prevent premature collagen fibrillogenesis. Tissue culture plastics were coated with rat tail collagen I, murine laminin I, bovine fibronectin, or murine collagen IV at 1 μg extracellular matrix component/cm2 of culture surface area, while EHS basement membrane extract was used at 5 μg/cm2.
All podocyte lines were cultured in a similar fashion with some exceptions noted in specific sections below. Proliferative growth was at 33°C in 75-cm2 flasks coated with rat tail collagen I. Culture medium was changed every 2 days during proliferation, and cells were never grown to >85% confluence before passage or seeding for differentiation. To passage cells, culture medium was removed and the flasks were rinsed with 2 ml of a trypsin-EDTA solution. Cells were detached by incubation with 5 ml of trypsin-EDTA solution for <5 min at 37°C. The HK rat podocyte cells required a longer incubation in trypsin-EDTA to detach.
Cells were seeded into culture vessels coated with rat tail collagen for differentiation at specific cell densities that were determined empirically (see below). Culture medium was changed at 2- to 4-day intervals (see results). All cells were differentiated for 14 days except HK cells, which were used 6 days after seeding. All tissue culture incubators used in this study were carefully calibrated before use, and the temperatures were monitored on a day-to-day basis with a thermometer accurate to within 0.05°C. All cell lines tested negative for the presence of mycoplasma (MycoProbe Mycoplasma Detection Kit, R&D Systems).
Cells were treated with PAN or DOX at various concentrations by addition of 1,000× stock solutions (in phosphate-buffered saline) to prewarmed medium, and then adding the PAN- or DOX-containing medium to cells. Cultures were incubated for 3 days after addition of PAN or DOX, and viability was measured by LDH and MTT assays.
JR and KE murine podocytes.
The cultured murine podocytes we have referred to in this manuscript as “JR” cells, derived from the Immortomouse in 1996 by Reiser et al. (34), were the kind gift of Jochen Reiser. These cells were obtained in 2008 and were designated as passage 1 when received. All studies were performed on cells in passages 4–8. The cultured murine podocytes we have referred to in this manuscript as “KE” cells, derived from the Immortomouse in 2004 by Schiwek et al. (38), were the kind gift of Karlhans Endlich. These cells were obtained in 2006 and were labeled passage 4. All studies were performed on cells in passages 8–12.
All JR and KE cells were cultivated under the same conditions except that the initial seeding densities of cultures in six-well plates were 1.0 vs.. 1.2 × 104 cells/well for JR vs. KE cells, while both were seeded at 1.5 × 105 cells/75-cm2 flask. The protocol was largely the same as previously described (28), although specific details were from a recent review by Shankland et al. (39).
MS human podocytes.
The cultured human podocytes we have referred to in this manuscript as “MS” cells, derived from human glomeruli by transfection with a construct encoding a temperature-sensitive variant of the SV40 T-antigen and telomerase in 2002 by Saleem et al. (37), were the kind gift of Moin Saleem. These cells were obtained in 2006 and were labeled passage 6. All studies were performed on cells in passages 10–14. Cells were cultivated essentially as previously described (37), with cells seeded at 6.0 × 104 cells/well in six-well plates and 5.0 × 105 cells/75-cm2 flask.
HK rat podocytes.
The cultured rat podocytes we have referred to in this manuscript as “HK” cells, derived from decapsulated glomeruli isolated from a temperature-sensitive SV40 T-antigen (tsA58) transgenic rat, were the kind gift of Hidetake Kurihara. These cells are clone C7, were obtained in 2006, and were designated as passage 1 when received. All studies were performed on cells in passages 4–8. Cell culture conditions were essentially as previously described (9), with cells seeded at 2.0 × 104 cells/well in six-well plates and 1.5 × 105 cells/75-cm2 flask.
Isolation of Rodent Glomeruli
Rat glomeruli were isolated from male Sprague-Dawley rats (Charles River Laboratories, Wilmington, MA) using graded sieves as previously described (21). Mouse glomeruli were isolated from 129/SvJ mice (Jackson Laboratories, Bar Harbor, ME) using magnetic beads as previously described (42). Human kidney cortex tissue was obtained from The National Disease Research Interchange (Philadelphia, PA), with exemption granted from the Nationwide Children's Hospital IRB.
Phase-contrast microscopy was performed on a DMI6000B inverted microscope with fully motorized stage and optics and TIRF (all black) environmental enclosure with full temperature, humidity, and CO2 control (Leica Microsystems, Bannockburn, IL). Digital micrographs were captured using a Retiga SRV 14-bit grayscale CCD camera (QImaging, Surrey, BC). Image capture and microscope control was done using Openlab 5.5.0 (Improvision, Waltham, MA) running on a Macintosh MacPro running OSX 10.5.4 (Apple, Cupertino, CA). Image postprocessing was limited to resizing and cropping images and contrast and brightness adjustments and was performed using Photoshop CS3 (Adobe, San Jose, CA). Time-lapse images were analyzed using ImageJ (National Institutes of Health, Bethesda, MD).
For fixed-cell imaging, cells were fixed in 4% formaldehyde in PBS for 30 min, permeabilized in 0.1% Triton X-100 in PBS for 5 min, and incubated for 15 min with 1 U of Texas Red-X phalloidin in 0.05% Tween 20 in PBS, followed by two washes in the same buffer. Images were captured under phase contrast and fluorescence (Leica TX2 cube, excitation: 560/40, emission: 645/75 nm) illumination.
Live cell imaging was performed in a chamber with full temperature, CO2, and humidity control. Images were captured under the control of a specific Openlab macro program that controlled stage position and limited illumination to that required for image capture. Cell motility was measure by manually determining the position of a nucleolus in nuclei of cells in images captured at 2-min intervals. The difference between the positions of nucleoli in serial images allowed the calculation of a “nuclear velocity” that reflected the change in the position of the cell's center of mass over time. Results were expressed as the percentage of cells (n = >20 cells/field, 6 fields/measurement) with nuclear velocity >1 cell diameter/h. This cell motility was measured in cultures that were ∼85% confluent. Proliferation was measured in the same images by direct observation of cells undergoing cell division and was expressed as the number of cells undergoing complete division·100 cells−1·h−1. Cell migration was measured by a “wound healing” assay, where a pipettte tip was used to scrape cells from an ∼1-mm-wide line across cultures differentiated in six-well dishes. Images were captured of four low-magnification fields in each of three immediately after wounding and 24 h later. The regions occupied by cells migrating into the wound were obtaining by hand-tracing, and a custom macro created for ImageJ was used to measure the mean distance migrated (in pixels, then converted to μm).
The viability of cells was measured by two assays, LDH release (a measure of the permeability of the cell membranes) and MTT reduction (a measure of cellular metabolic activity). Cells were plated in 96-well plates, and all treatments were performed with n = 8 wells. Medium was removed from the wells after 3-day treatments with DOX or PAN, and LDH activity was measured in the clarified medium following the manufacturer's instructions. Fresh medium containing 0.5 mg/ml thiazolyl blue tetrazolium bromide was added to wells, and cells were cultured for 3 h. Medium was removed, cells were washed with PBS, and the formazan crystals in cells were dissolved by agitation in 0.04 M HCl in isopropanol for 30 min. Absorbance was measured at 590 nm.
Protein was extracted from podocyte cultures in 75-cm2 flasks into 0.5 ml of buffer (63 mM Tris·Cl, pH 6.8, 10% glycerol, 2% SDS, 2 M urea). Urea was included in the standard SDS-PAGE sample buffer to improve dissolution of integral membrane proteins. After incubation for 3 min at room temperature, the extract was collected using a rubber policeman and sonicated on ice for 10 s at 100% displacement (Vibra-Cell, Sonics, Newtown, CT). Protein was extracted from isolated rat glomeruli with a teflon-glass Dounce homogenizer into 10 volumes of the same buffer. Extracts were centrifuged at 18,000 g for 15 min at 4°C, an aliquot was removed for protein assay, and the supernatant was brought to 5% with 2-mercaptoethanol and stored frozen at −80°C. Protein concentration was determined by the BCA assay (Pierce, Rockford, IL).
Proteins were separated by Tris·Cl SDS-PAGE on 10.5–14% Criterion precast gradient gels (Bio-Rad) according to the manufacturer's protocol with the exception that the gels were run under constant voltage at 150 V (rather than 200 V). Proteins were transferred to nitrocellulose membranes (Bio-Rad) using the procedure of Otter et al.(31), a protocol designed to improve transfer of high-molecular-weight and integral membrane proteins.
Nonspecific protein binding sites on dried transfer membranes were blocked by incubation in 5% nonfat dry milk in T-PBS (0.05% Tween 20 in PBS) for 1 h at room temperature. After three brief washes in T-PBS, primary antibodies diluted in 2% BSA in T-PBS were applied for overnight incubation at 4°C. After three 15-min washes in T-PBS, secondary antibodies diluted in 2% nonfat dry milk in T-PBS were applied for 2-h incubation at room temperature. After a final three 15-min washes, secondary antibody binding was visualized using a chemiluminescence reagent (SuperSignal, West Pico, Pierce) as captured and measured using the ChemiDoc SRS system and Quantity One software (Bio-Rad). The antibodies against podocin, NEPH1, and synaptopodin were used at 1:200 dilution, and the antibody against nephrin was used at 1:1,000 dilution. All secondary antibodies were used at 1:10,000 dilution.
RT- and Quantitative RT-PCR
Podocyte cultures in 75-cm2 flasks and isolated rat glomeruli were extracted into TRIzol (Invitrogen) and processed to isolate RNA according to the manufacturer's instructions. After the final ethanol wash and centrifugation, the RNA in the pellets was cleaned using the RNeasy mini-kit (Qiagen) according to the RNA Cleanup protocol in the manufacturer's instructions. The RNA concentration in the solutions was measured by absorbance at 260 nm, and cDNA was synthesized from RNA using a High Capacity cDNA Reverse Transcriptase Kit (Applied Biosystems, Foster City, CA) according to the manufacturer's instructions.
Quantitative RT-PCR was performed on an iQ5 Real Time PCR Detection System (Bio-Rad) using an IQ SYBR Green cocktail (Bio-Rad) according to the manufacturer's instructions. Primer pairs were identified using the online primer selection program AutoPrime (47) and are shown in Table 1. The PCR protocol consisted of 3-min denaturation at 95°C followed by 40 cycles of 95°C for 15 s and 55°C for 30 s. After amplification, products were annealed and a temperature gradient was applied to obtain melting temperatures. Single peaks in the resulting melt curves (indicating a single amplified species) were obtained for all reactions reported in Table 2 save those indicated by “ND,” where no specific product was detected, and the size of products was confirmed by separation by agarose gel electrophoresis.
The primary culture parameter found to be critical for uniform podocyte cultures after differentiation was the number of undifferentiated cells used to seed cultures. Protocols obtained from Drs. Moin Saleem, Jochen Reiser, and Hidetake Kurihara specified seeding densities based on percent confluence of the proliferating culture, but we found that the resulting differentiated cultures that varied unpredictably between sparse (≤40% confluence) and highly confluent with wide variability in cell morphology.
When seeded at low densities, the resulting differentiated cultures attained ∼60% confluence and primarily consisted of large, arborized cells in JR, KE, and HK cultures and cells with active marginal lamellipodium in MS cultures (Fig. 1, A–D, low). The most uniform differentiated cultures were obtained with JR, KE, MS, and HK cells seeded at 12, 10, 60, or 20 × 103 cells, respectively (Fig. 1, A–D, med) and consisted of closely spaced, roughly circular cells with abundant cell processes (JR, KE, HK) or cell-cell contacts resembling tight junctions (MS). Phalloidin labeling revealed prominent stress fibers in all cells, and JR, KE, and MS cells also displayed subcortical rings of filamentous actin (Fig. 1, A–D, phalloidin). Some JR and HK cells contained weblike clusters of actin filaments (white arrows). Finally, when seeded at about twofold higher densities, all podocyte lines formed highly confluent cultures with altered morphology (Fig. 1, A–D, high). Despite the supraconfluent densities of JR, KE, and HK cells, gaps between cells bridged by cell processes were always visible at higher magnifications (data not shown). Only MS cells formed long, continuous cell-cell contacts in culture.
Other culture parameters tested included the type of extracellular matrix protein used to coat culture dishes (bare culture plastic, collagen I, collagen IV, laminin, fibronectin, and Engelbreth-Holm-Swarm matrix), temperature of differentiation (37, 38, or 39°C), and frequency of medium change (2, 3, or 4 days). The only changes in the final differentiated cultures were higher cell densities at a given initial seeding density in cultures grown on any type of extracellular matrix vs. bare plastic, in cultures grown at 37 or 38°C vs. 39°C, and in cultures fed every 2 days vs. every 3 or 4 days (data not shown). No significant changes in protein (Fig. 2) or mRNA (data not shown) expression were observed in cultures grown under different conditions.
Expression of Cultured Podocyte Proteins
The relative expression of four podocyte proteins was determined in protein extracts of undifferentiated cultured podocytes and in cultures differentiated for 14 (JR, KE, MS) or 6 days (HK) at either 37 or 39°C (Fig. 2, A, C, E, and G). To compare specific protein expression in cultured podocytes with podocytes in vivo, duplicate Western blots of cultured podocyte proteins (40 μg/lane) were probed and exposed together with blots of various quantities (6, 12, 25, and 50 μg/lane) of proteins extracted from isolated rat glomeruli (Fig. 2, B, D, F, and H). Synaptopodin protein was detected in all cultures, with greater expression in mouse podocyte lines after differentiation (Fig. 2A). Smaller amounts of synaptopodin were found in extracts of human podocyte cultures, and expression did not change after differentiation, while rat cultures expressed an intermediate quantity of synaptopodin although the amount was similarly unchanged by differentiation. Interestingly, the amount of synaptopodin in cultured podocytes was dramatically less than found in extracts of rat glomeruli, even when separations of 40 μg of total protein from cultured podocytes were compared with 6 μg of rat glomerular protein (Fig. 2B).
Surprisingly, the amount of NEPH1 protein was greater in undifferentiated mouse and human cultured podocytes than in differentiated cultures (Fig. 2C). The amount of NEPH1 in these cultures was less than but comparable to the amount found in rat glomeruli (Fig. 2D). As with synaptopodin, in cultured rat cells the amount of NEPH1 protein was similar in undifferentiated and differentiated cells (Fig. 2C). Podocin was detected in differentiated JR mouse podocyte cultures, and in undifferentiated and differentiated KE mouse podocyte cultures (Fig. 2E), and the amount of podocin was similar to that found in rat glomeruli (Fig. 2F). Little or no podocin was detected in human and rat podocyte cultures (Fig. 2E). Finally, little or no nephrin was detected in any podocyte cell line (faint bands in KE, MS, and HK), despite very long exposures (∼1 h) of the chemiluminescent Western blot to sensitive film (Fig. 2G). For comparison, the similarly processed Western blots of cultured podocyte extracts compared with rat glomerular extracts were exposed for 1 min (Fig. 2H). To confirm these results, a fresh set of protein extracts was obtained from new undifferentiated and differentiated cultures, and separations of 80 μg of protein isolated from cultured podocytes were compared with only 3 μg of protein extracted from isolated rat or mouse glomeruli or 6 μg of total human kidney cortex protein in a Western blot (Fig. 2I). This blot confirmed that all podocyte cell lines expressed detectable amounts of nephrin protein, but the amounts were vastly less than comparable in vivo podocytes.
Interestingly, no differences in podocyte marker protein expression were noted in cultures differentiated at a higher (39°C) nonpermissive temperature compared with the cultures grown at 37°C (Fig. 2, compare 37 with 39°C), despite the fact that the temperature-sensitive SV40 T-antigen is only completely inactivated at the higher temperature. No differences were noted in podocyte marker expression when other culture parameters (extracellular matrix coating on culture dishes or frequency of medium change) were altered (data not shown).
Expression of Cultured Podocyte mRNAs
The expression of specific mRNAs was measured in total RNA by qRT-PCR, and the result was normalized to a constitutively expressed mRNA (GAPDH) and expressed as fold-difference in mRNA expression compared with the normalized expression in kidney cortex tissue of the same species (Table 2). Since only a fraction of cortex tissue is composed of glomeruli, the expression of podocyte-specific mRNAs [NEPH1, NPHS2, NPHS1, Wilms' tumor-1 (WT1), and SYNPO] was, as expected, higher (between 3.42- and 49.7-fold greater) in total RNA extracted from rat glomeruli than rat kidney cortex. Similarly, since podocytes comprise only a fraction of all cells in a glomerulus, expression of podocyte-specific mRNAs would be expected to be greater in cultured podocytes than in glomeruli. However, only the levels of expression of NEPH1 and NPHS1 mRNAs in differentiated murine podocyte lines were greater than those in rat glomeruli. In addition, only the expression of NEPH1 and SYNPO mRNAs in all podocyte cell lines was consistently comparable to the expression in kidney cortex. Interestingly, the expression of NPHS1 mRNA (encoding nephrin) was consistently very low or undetectable in all podocyte cell lines, and the expression of NPHS2 (encoding podocin) and WT1 mRNAs was low or undetectable in human and rat cultured podocytes.
Proliferation, Motility, and Migration of Cultured Podocytes
Phase-contrast micrographs of podocyte cultures were captured at 2-min intervals over several hours and were used to directly measure the number of cell divisions (number of cell divisions·100 cells−1·h−1), cell motility in near-confluent cultures (percentage of cells that moving at least 1 cell diameter/h), and cell migration (mean distance migrated into ∼1-mm-wide wound in 24 h). As shown in Table 3, undifferentiated murine (JR Un, KE Un) and human podocytes proliferated at the rate of about three to five cell divisions·100 cells−1·h−1, while the rate of rat podocyte division was about one cell division·100 cells−1·h−1. Interestingly, there was little or no proliferation in differentiated murine and human podocyte cultures, while differentiated rat cultures proliferated at the same rate as undifferentiated cells.
Similarly, the motility of murine and human podocytes was greater in undifferentiated than undifferentiated cultures, although human cultured podocytes remained highly motile (54 ± 8% of cells moving more than 1 cell diameter/h) even after differentiation (Table 3). Interestingly, the motility of differentiated human cultured podocytes was high even in confluent cultures (Supplemental Fig. S1; all supplementary material for this article is available on the journal website), in marked contrast to murine podocytes. Undifferentiated murine and human podocytes were motile regardless of culture density. In contrast, neither undifferentiated nor differentiated rat podocytes were motile, although some observations were made for as long as 14 h.
Similar results were obtained when cell migration (by “wound-healing” assay) was measured in differentiated cultures. The JR murine podocytes migrated into a “wound” at approximately the same rate as HK rat podocytes but less rapidly than JR murine podocytes, while MS human podocytes migrated at more than twice the rate of other cultures (Table 3).
Cultured Podocyte Processes are Formed by Retraction of Lamellipodia
Time-lapse micrographs of cultured podocytes also revealed that long cultured podocyte cell processes consisted of “spikes” resulting from lamellipodium that extended, attached to the substrate or to adjacent cells, and finally formed spike-like processes as a result of subsequent retraction. This process was particularly evident in movies made of time-lapse images of JR murine podocytes (Supplemental Fig. S2), where lamellipodia formed, extended, and then retracted. This process results in extended spike processes formed between the cell body and points of cell-cell or cell-substrate attachment that do not detach as the lamellipodium retracts. The before-and-after results of this process are shown in Fig. 3, with white arrows indicating the position of lamellipodium that in the subsequent images leave behind spike processes indicated by a black arrow (Fig. 3A). Filopodia were also observed in JR cultures (Supplemental Fig. S2, white circles) but were relatively infrequent, transient, and short compared with the retraction spikes. The KE murine podocyte cultures behaved similarly (Fig. 3B, Supplemental Fig. S3), although the majority of lamellipodium formed by KE cells were smaller than in JR cultures. Filopodia were observed more frequently in cultures of KE cells, and these filopodia were longer and less transient (Supplemental Fig. S3, white circles), but the majority of the long cell processes in KE cultures were formed by retraction after extension of lamellipodia. The processes indicated by arrows mark the lamellipodium and spikes resulting from cell-cell contacts between MS cells (Fig. 3C). Interestingly, unlike the spikes formed by murine podocytes, the spikes formed in MS cultures were almost exclusively a result of cell-cell attachments, although the spikes visible in the initial frames of the time lapse were a rare example of spikes formed as a result of MS cell-substrate attachments (Supplemental Fig. S4). The spikes observed between MS cells were often extremely thin and sometimes extended >1 cell diameter. No long filopodia were observed in MS cultures. The spikes formed by HK rat cells were also typically long, though never as thin as those formed by MS cells (Fig. 3D). As in JR cultures, filopodia were infrequent, short, and transient in HK cells (Supplemental Fig. S5, white circles).
To confirm that the majority of cultured podocyte processes are formed by retraction, cells were fixed and labeled with fluorescent phalloidin (to label filamentous actin) and antibody directed against fascin (to label lamellipodia and filopodia). As shown in Fig. 4, spikelike cell processes on JR, KE, and HK cells as well as the very thin processes on MS cells contained actin filaments but were not labeled strongly by the anti-fascin antibody (white arrows). Only occasionally were cell processes, typically in association with lamellipodia, observed that were strongly labeled by anti-fascin antibody (characteristic of filopodia, yellow arrows).
Response of Cultured Podocytes to Podocyte-Selective Toxins
PAN and DOX are commonly used to induce experimental nephrotic syndrome in rodents and have been used in vitro to model podocyte injury. The toxicity of these agents was measured using two assays to measure both the “viability” of cells (their ability to reduce MTT) and membrane injury (leakage of LDH). Cells that were lysed by incubation with 0.1% Triton X-100 for 15 min before the assay were used as a positive control for the LDH assay (see Figs. 5 and 6; TL).
The toxicity of PAN varied enormously between different podocyte cultures. Cultured human podocytes (MS) were the most sensitive to PAN injury, as treatment with 5 μg/ml for 3 days resulted in a significant decrease in reduced MTT and treatment with 10 μg/ml caused a significant increase in LDH release (Fig. 5, E and F). Treatment with ≥25 μg/ml PAN killed all MS cells, as shown by the similar values of LDH release between these cells and lysed cells (TL). Results were similar when the JR murine podocytes were treated with PAN (Fig. 5, A and B), but the response of JR cells was very different from the response of the similarly derived KE murine podocytes, where only ≥25 μg/ml PAN treatment resulted in significant decreases in MTT and significant LDH release only after treatment with 100 μg/ml PAN (Fig. 5, C and D). Unlike MS and JR cells, KE cultures were not killed by 100 μg/ml PAN treatment. Interestingly, cultured rat podocytes were almost entirely resistant to injury by PAN, with no significant changes in MTT reduction after treatment with any PAN concentration tested, and significant release of LDH only after treatment with 100 μg/ml PAN (Fig. 5, G and H).
Interestingly, unlike the differential response to PAN, murine podocytes were similar in their sensitivity to injury by DOX. The ability to reduce MTT was decreased and the release of LDH was significantly increased in both cell lines by nearly the same concentrations of DOX (Fig. 6, A–D). The viability and integrity of human podocyte cultures were also affected by similar concentrations of DOX (Fig. 6, E and F). However, rat cultures were almost unaffected by DOX even at the highest concentrations tested (Fig. 6, G and H).
The objective of this study was to directly compare the behavior and characteristics of conditionally immortalized podocyte cell lines from different species with both the other podocyte lines and, to the extent possible, in vivo podocytes. Culture conditions were standardized, and the parameter most important to consistent differentiated cultures was found to be the initial seeding density. Other culture parameters such as temperature of differentiation, extracellular matrix coating of culture dishes, and frequency of medium change had no significant effect except on the final cell density of the differentiated culture. The expression of podocyte marker proteins and their corresponding mRNAs varied considerably between cell lines, and both synaptopodin and nephrin proteins were present in dramatically lower amounts in cultured podocytes than in podocytes in vivo. Cell motility, migration, and proliferation of differentiated cultures also varied between lines, with human podocytes retaining considerable motility and rat podocytes high proliferative capacity. Finally, the response of podocyte cultures to PAN and DOX varied dramatically between lines, particularly in rat podocytes that were almost entirely resistant to both chemicals. These results demonstrated that despite certain similarities, particularly between murine podocyte cultures derived in similar fashion, conditionally immortalized cultured podocytes varied considerably in their expression of podocyte markers, behavior, and response to agents used to induce experimental nephrotic syndrome in animals.
Podocytes were cultured using the specific protocols provided by the investigators who originally produced the lines (9, 28, 37, 38), and particular care was taken to culture the JR murine podocytes exactly as specified in a recent review article (39) that indicated that these cells are especially sensitive to many culture conditions. In addition, care was taken to use cells of the lowest possible passage number, and a fresh aliquot of JR podocytes was obtained from Jochen Reiser shortly before studies were performed. However, we found that varying a number of culture conditions during differentiation had little effect on cultures, and in particular the expression of podocyte marker proteins and mRNAs was unaffected by culture substrate, temperature of differentiation, and frequency of medium changes in differentiated cultures.
Of the four podocyte-marker proteins examined, only NEPH1 and podocin were present in podocyte cultures in quantities comparable to in vivo podocytes (in comparisons with glomerular extracts). Podocin has previously been detected in all of the cell lines tested except HK cells by Western blotting (7, 10, 11, 23, 25, 27, 36, 37), although only a single study (38) compared podocin expression in cultured podocytes (KE) to a standard (extract of whole kidney). Our results demonstrated that differentiated JR cells and both un- and differentiated KE cells expressed podocin protein at levels comparable to in vivo podocytes, consistent with the study by Schiwek et al. (38), while expression in HK and MS cells was much less. Podocin is expressed in both fetal and adult glomeruli in rodents and humans (18, 35), consistent with the differentiation-independent expression of podocin protein in KE. Surprisingly, NPHS2 mRNA expression (Table 2) was at least 30-fold less in all cultured podocytes than in rat glomeruli, although as expected mRNA expression was much greater in JR or KE cells than in MS or HK cells.
Only a single study (25) found NEPH1 protein in cultured podocytes (JR). Our finding that the quantity of NEPH1 protein is greater in undifferentiated cultured podocytes than in differentiated cells was unexpected, although no existing literature was found describing the expression of NEPH1 in the developing podocyte in vivo. In contrast, NEPH1 mRNA expression was greater in all differentiated cultures than undifferentiated and was comparable to expression in glomeruli. This disparity between NEPH1 mRNA and protein results may be a consequence of the transient nature of the cell-cell contacts between cultured podocytes and the lack of stable foot processes (see Supplemental Figs. S2–S5), as NEPH1 is found in vivo in the slit diaphragm between foot processes (3).
The quantities of nephrin and synaptopodin proteins in all cultured podocyte lines were much less than found in glomerular extracts. The quantity of synaptopodin protein was greater after differentiation only in cultured murine podocytes, but was detected in all cell lines. Synaptopodin has previously been detected in Western blots only in MS and JR cells (2, 7, 23), although it has been visualized by immunofluorescence microscopy in KE cells (38). Our results are consistent with the differentiation-dependent expression of synaptopodin in JR and KE cells, but not with the report of differentiation-dependent expression in MS cells (37), although synaptopodin mRNA expression was higher in MS cells after differentiation. Synaptopodin expression in human podocytes is differentiation specific in vivo (29). Although less synaptopodin protein was present in cultured podocytes, synaptopodin mRNA expression was comparable in differentiated podocytes and glomeruli. Both nephrin protein and mRNA were barely detectable in any podocyte cell line, and the amounts were vastly less than in glomeruli. Nephrin has been detected in Western blots of extracts of all podocyte cell lines in several studies (7, 9–11, 23, 25, 37, 44, 48, 49), and in KE cells was found to be comparable to levels in kidney extracts (38). The study by Eto et al. (9) that detected nephrin in cultured rat podocytes by Western blotting used a different clone (D7) of HK cells than used in this study (C7). These results demonstrate that, while detectable, the quantities of synaptopodin and nephrin in cultured podocytes are only a tiny fraction of that present in podocytes in vivo and suggest that conclusions about alterations in the expression of these proteins in cultured podocytes may not accurately reflect processes occurring in vivo.
The proliferation rate of podocyte lines was, as expected, greater in undifferentiated than differentiated JR, KE, and MS cells, findings in agreement with previously published analyses (28, 38). In contrast to a previous report (9), rat HK cells did not show evidence of growth arrest upon transfer to the nonpermissive temperature. The findings in the current report measured proliferation by direct observation of cell divisions, which may not be directly comparable to studies reporting increases in total cell number (9, 28, 38). The motility of cultured podocytes was largely lost upon differentiation in both KE and JR cells, and both un- and differentiated HK cells were nonmotile, but a majority of MS podocytes remained motile. MS cells were found to be motile regardless of culture density (Supplemental Fig. S1). In fact, the 54 ± 8% of MS cells found to be motile only measured the fastest moving cells, and prolonged observations suggest that all MS cells are motile. Migration of cultured JR, MS, and KE cells has been previously demonstrated by “wound-healing” assays that determined the number of cells migrating into a denuded area of a culture vessel (8, 15, 32). We also measured migration by this method, and found results (MS > KE > JR ∼ HK) similar to those obtained by motility assay, although HK cells proved capable of migration into denuded areas of the culture vessel but were not motile when in near-confluent culture. The motility (and higher rate of migration) of human but not rodent podocyte cultures could be explained by differences in podocyte behavior between species, by differences between cells derived from transgenic rodents vs. retroviral transduction of isolated glomeruli, or other factors including inclusion of telomerase in the immortalization construct used to create the human cell line. This result, however, demonstrates a fundamental disparity between existing lines of cultured human and rodent podocytes.
A number of studies have examined the processes produced by cultured podocytes. Electron microscopy revealed that cell-cell contacts were connected by tight junctions in JR cells (28), and found evidence of slit diaphragm-like cell-cell contacts in KE cells (38). Immunofluorescence microscopy was used to implicate a variety of molecules in process formation in cultured JR cells (19, 20, 33). However, the question of how podocyte processes form in culture cannot be answered by microscopic analyses of fixed cells. A switch from a morphology characteristic of filopodium to lamellipodium upon angiotensin II treatment was observed by time-lapse live-cell microscopy in KE cells (15), but the images presented were taken at 30-min intervals and so did not allow analysis of the formation of individual processes. In contrast, by taking images at 2-min intervals we determined that the vast majority of the long processes extending from cultured podocytes were generated by extension of lamellipodia followed by retraction. This finding is supported by our observation that only very rarely did processes in any cell line contain the characteristic filopodial marker, fascin. However, filopodia were infrequently found, typically in association with lamellipodia, and did contribute to some of the longer processes formed by KE cells. Filopodia can play many roles in cells, including promoting the establishment of cell-cell contacts and guidance of lamellipodia formation during migration (reviewed in Ref. 46). Lamellipodia and filopodia coexist at the protruding front of migrating cells, and our live-cell imaging results support the hypothesis that, at least in murine and rat podocyte cultures, filopodia form short extensions that are followed by broad lamellipodia. What is novel is the finding that the majority of the long processes (>10% of cell diameter) observed in cultured podocytes are a result of partial retraction of the lamellipodia, leaving behind thin processes where the leading edge of the lamellipodium remained attached to the substrate. Since the mean maximum diameter of podocyte foot processes is ∼150 nm (43), it is likely that the ∼5-μm-wide cell processes observed in cultured podocytes were formed by a different process, although MS cells formed <1-μm-diameter retraction spikes. These results lead to the conclusion that studies of process formation in cultured podocytes are analyses of lamellipodia formation, retraction, and the detachment of cell-substrate contacts.
The various cultured podocyte lines differed dramatically in their responses to PAN and DOX treatments. In particular, HK cells were highly resistant to changes in viability and membrane integrity induced by either PAN or DOX compared with other cell lines. Eto et al. (9), examined changes in actin filament morphology in a different clone (C7 vs. D7 used in this study) of HK cells in response to PAN but did not measure viability. Similarly, most studies of the effects of PAN on JR, KE, and MS cells examined morphological changes following treatment, typically using PAN concentrations of ∼100 μg/ml for periods of 24–48 h (13, 26, 36). Few studies have examined podocyte viability or cell death after treatment with PAN (6, 40), or have treated cultured podocytes with DOX (6, 24). It is therefore difficult to compare the current results with previous studies. However, it is intriguing to note that while rats are notably susceptible to developing nephrotic syndrome in response to subcutaneous, intravenous, or intraperitoneal PAN injection (12), and there is evidence that humans are also susceptible (30), most strains of mice are not (45). The only report of robust proteinuria in response to PAN in mice was after daily subcutaneous injections of PAN for 3 wk in the hyperprolinemic, prolinuric PRO/Re strain (17). Similarly, while intravenous injection of DOX induces nephrotic syndrome in rats (4) and mice (5), there are no reports of massive proteinuria in humans treated with DOX, despite its common use in chemotherapy. These in vivo susceptibilities to PAN (rat and human susceptible, mouse unaffected) and DOX (mouse and rat susceptible, human unaffected) contrast with the responses of cultured podocytes in vitro to PAN [mouse (JR) ∼ human >> mouse (KE) >>> rat] and DOX (mouse ∼ human >>> rat). Indeed, if the toxicities of PAN and DOX in cultured podocyte lines in vitro are an accurate reflection of the response of podocytes in vivo, these results indicate that the target of the putative “podocyte toxins” PAN and DOX in experimental nephrotic syndrome in rodents is not the glomerular podocyte, although the elegant transplant experiments performed by Hoyer et al. (14) demonstrated that the site of action of PAN is in the kidney.
This study provides the first direct comparisons between different conditionally immortalized cultured podocyte lines and found dramatic differences, even between similarly derived murine lines (Table 4). Most notable were the differences in cell motility and responses to the putative podocyte toxins PAN and DOX. While these findings do not deny the validity of cultured podocytes as a model of in vivo podocyte biology, they do define specific limitations to the model and raise intriguing questions concerning their behavior, response to injury, and the differences between podocyte cultures derived from different species.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grant RO1-DK075533.
No conflicts of interest, financial or otherwise, are declared by the authors.
The author gratefully acknowledges the generous contributions of cultured cells and suggestions by Jochen Reiser, Karlhans Endlich, Moin Saleem, and Hidetake Kurihara.
- Copyright © 2011 the American Physiological Society