A good in vitro model should approximate an in vivo-like behavior as closely as possible in order to reflect most likely the in vivo situation . Regarding renal physiology of different species, humans are more closely related to pigs than to rodents [5, 13]. Therefore porcine renal cells could prove to be a valuable tool for renal in vitro toxicology. Such a renal in vitro model should display specific characteristics (e.g. structural organization) and functions (e.g. transepithelial transport) of the respective part of the kidney. This study thus aimed to provide data in order to enable an informed assessment of the suitability of primary porcine renal cells as a model for in vitro toxicological testing.
The kidney plays an important role in the elimination of harmful compounds from the body and the reabsorption of vital substances from the glomerular filtrate back into the body. Both types of compounds need to be transported across the membranes of the tubular cells which is realized by specifically localized transporters especially those of the ABC and SLC families [14–25]. A good in vitro kidney model must therefore functionally express these transporters in the correct location. In the present study, PKCs were characterized with respect to morphological features (monolayer with the typical epithelial appearance and dome formation). A basolateral-apical polarity can be assumed due to the transporting capabilities observed (see below). This is in accordance with the data from other researchers who additionally tested similarly prepared and cultured PKCs for membrane-specific localization of pMdr1 (abcb1, apical) and pNbc1 (slc4a4, basolateral) using confocal microscopy . These data corroborate earlier findings in PKC, including stable cytokeratin expression patterns up to passage 2 and correspondingly very low but stable vimentin expression as markers of epithelial and mesenchymal cell types, respectively [8, 27]. Growth patterns (cellular doubling time, total cell number, monolayer formation) of PKC were found to be similar over three passages although signs of senescence (and/or degeneration) were increasing with time in culture and were apparent especially for cells of passage three. Enlarged cell bodies and cytoskeletal changes are typical phenomena known for senescent cells. Cellular degeneration, especially with respect to change of phenotype to a more fibroblastoid type would possibly be reversible and would include for example a change in marker protein expression. The latter was not observed with cytokeratin and vimentin in earlier studies [8, 27].
The data presented here demonstrate that PKC express various transporters at the mRNA level including pMrp1 (abcc1), pMrp2 (abcc2), pOat1 (slc22a6) and pOat3 (slc22a8), whereas pMdr1 (abcb1) and pOatp1a2 (slco1a2) mRNA could not be detected in either the PKCs or in the porcine cortical tissue. The latter might suggest the absence or very low abundance of these transporters in porcine kidney and PKC, a result of the detection technique employed, or a combination of both, i.e. low abundance and limited detection. Results published by Schlatter et al. confirm the latter interpretation as they demonstrated pOatp1a2 mRNA in their PKC, but only in freshly isolated cells and with a very faint band after a very high number of PCR cycles . Moreover, they were able to detect pMdr1 mRNA in cultures from different cell preparations using quantitative PCR .
To confirm the presence of various transporters in PKC detected at the mRNA level Oat protein expression was assessed also via Western blotting analysis. Due to the fact that transporters are low abundant proteins relative to total cellular protein , it was necessary to use higher film exposure times, thereby resulting in relatively high backgrounds. These experiments however confirmed the mRNA expression data for pOat1 and pOat3 in PKC, although the protein bands were different from those obtained with the renal tissue samples from rat and porcine origin. Specific immunoreactive proteins at approximately 75, 45 and 40 kDa were detected. It is suggested, that the large protein (~75 kDa) represents the mature, fully glycosylated form. The smaller proteins, which predominate in vitro, may be deglycosylated forms of splice variants  and/or forms that are more affected by the experimental procedure, as previously been reported for human OAT1 . However despite that Schlatter et al.  demonstrated the expression of pMdr1 (abcb1), pMrp1 (abcc1) and pMrp2 (abcc2) in PKC by immunocytochemistry using human protein derived antibodies, Mrp proteins were not detectable in porcine cortical tissue and cells in the experimental set-up reported here. The latter lack of detection, which stands in contrast to the mRNA data, most likely is due to the use of the Mrp antibodies employed. Here, polyclonal anti-human MRP1/2 antibodies were used, that had been predicted to cross-react with the porcine homologues. Obviously this was not the case as these transporters were not detectable in PKC as well as in porcine cortical tissue, although the general function of the antibodies was confirmed by earlier experiments using human tissue (data not shown). These negative results stand also in contrast to data from other investigators showing that Mrps are highly expressed in the kidney of several other species, as reviewed by Klaassen and Aleksunes .
In contrast to PKC, the continuous cell lines LLC-PK1 and NRK-52E were negative for all transporters tested at mRNA and protein level. Whereas other investigators described low but detectable functional expression of these transporters in NRK-52E cells [31–34], LLC-PK1 cells were extensively used as models for transporter over-expression by transfection due to their lack of endogenous transporter expression [35–41]. The negative results for NRK-52E cells here are most likely a result of using different strains of NRK-52E cells (ATCC vs. DSMZ), but different cell culture or detection conditions might be causative as well.
Transporter functionality was assessed using radiolabelled PAH and MTX as model transport substrates. PAH uptake occurred across the BLM and BBM, but was found to be different in capacity and affinity, with BBM uptake being about 3 times higher compared to BLM uptake. PAH uptake was time-dependent with saturation reached within approximately 5-10 minutes. MTX, OTA and BSP reduced the PAH BBM uptake by up to 30%, whereas CIM, FA and OTB showed similar inhibitory effects on the basolateral PAH uptake. All other substances tested did not inhibit PAH uptake from either the basolateral or luminal direction.
The observed PAH transport kinetics in PKCs as determined appeared typical for the expected specific transport and thus was considered, at least in part, as attributable to specific expression of pOat1 and pOat3. However, due to the lack of corresponding published data for porcine cells the latter findings could not be corroborated. Indeed, Schlatter and co-workers who also investigated transport in PKC did not report Km values or other parameters that might be suitable for comparison . On the one hand, the calculated Km values for PAH uptake across BBM and BLM were comparable to those observed in primary human proximal tubular cells (BLM Km of 57.5 ± 4.1 μM; BBM Km = 36.6 ± 1.2 μM , which is consistent with the high protein homology between porcine and human OATs. The amino acid sequence of pOat1 and pOat3 showed 89% and 81% homology to the human counterparts, respectively [29, 42]. Hagos and co-workers investigated two slice variants of pOat1 and one variant of pOat3 expressed in Xenopus laevis oocytes [29, 42]. While one pOat1 variant did not show any affinity for [3H]PAH, the other presented with an apparent Km for [3H]PAH of 3.75 ± 1.6 μM. The latter uptake could be inhibited by 0.5 mM glutarate or 1 mM probenecid [25, 29], which is comparable to that observed with hOAT1 (Km = 3.9-22 μM) . The reported Km values refer to single transporter proteins overexpressed in oocytes, whereas the presented Km values in our study cannot be assigned to a single transporter but rather to mixed action of at least two different pOats. Nevertheless, based on expression data the observed uptake of PAH is most likely related to pOat1 and to a lesser extent to pOat3. ES, a model substrate for pOat3, was not able to inhibit PAH uptake in our study, which further suggests a predominantly pOat1-mediated uptake. Functional investigations of pOat3 expressed in X. laevis oocytes resulted in a high affinity [3H]ES transport with an apparent Km of 7.8 ± 1.3 μM with an inhibitory effect of glutarate, DHEAS and probenecid , which is similar to hOAT3 (Km = 3.1-9.5 μM) . In our study, probenecid did not show any inhibitory effect on PAH uptake from either the basolateral or the luminal side suggesting that either an additional transporter may be involved in PAH uptake which is not sensitive to probenecid or that the detected transporters pOat1 and pOat3 are indeed slightly different (glycolysation-) variants when compared to their in vivo counterparts.
MTX has been shown to interact with several human transporters including hOAT1 (Km = 554-724 μM), hOAT3 (Km = 10.9-21.1 μM), hOAT4 (Km = 17.8 μM), hMRP2 (Km = 250-480; 2,500-3,000 μM) and hMRP4 (Km = 220-1,300 μM) . Cellular MTX uptake was shown to be mediated by hOAT1/3, whereas hOAT4 and hMRPs seem to be involved in MTX efflux . In human kidney cells (HKC), comparable to the PKC cells employed here, hOAT1-associated MTX transport was reported across the BLM with Km = 28.5 ± 1.1 and across the BBM with Km = 80.4 ± 3.4 μM . Based on human data, it can be assumed that MTX uptake is of high affinity/ low capacity, whereas the opposite seems to be true for MTX efflux. If the same holds true for the renal porcine transporters tested here, the results obtained with PKC could be explained by a highly efficient MTX excretion surmounting the concurrent uptake and thus resulting in the low intracellular concentrations determined. Indeed, MTX has been used by others in PKC as a model substrate for Mrp1-6 (abcc1-6), and hypothesized to pump MTX out of the cells . Beyond the latter, MTX may also not be a substrate or a low affinity/low capacity substrate for the transporter variants expressed in PKC. Due to the low intracellular concentrations determined, the question may also be raised whether the transporter variants expressed are truly functional.
mRNA expression levels of pOat 1 and pOat3 in PKCs of passage 0 at different states (cell density and time in culture) was examined and showed that the expression levels for both transporters decreased over time. PKCs of passage 1 did not show any mRNA expression of pOat1 and pOat3 at all. This confirms earlier results from Schlatter and co-workers who observed a down-regulation of pOat1 in culture and detected pOat3 in freshly isolated cells only . The latter observations were also true for other transporters such as pOct1 (slc22a1), which was detectable in freshly isolated cells only, pMrp1 (abcc1) was up-regulated and pMrp2 (abcc2) mRNA expression was down-regulated in culture . To provide a more comprehensive picture of the situation, pOat1/3 expression in PKC was also investigated at the protein level. In contrast to the mRNA data, pOat1 was detectable in PKC up to passage three, whereas pOat3 was present only up to passage one and then at a much lower amount than pOat1. The latter observations suggest that in PKC protein stabilization and/or degradation mechanisms could be altered. This effect and the observed alterations in mRNA expression further confirm the assumption of increasing senescence in the PKC cell strain.
Differences in transporter function amongst individuals may be the result of an up-or down-regulation of OAT expression subsequent to prior exposure to drugs and xenobiotics, or to differences in hormonal status . In addition to the initial expression, post-translational mechanisms including glycosylation, phosphorylation, ubiquitination and SUMOylation seem to be involved in regulating the level and functionality of the expressed transporters [24, 25]. While glycosylation appears to be critical in membrane targeting/ trafficking, protein folding and possibly regulation of OAT function, data suggest that phosphorylation and interactions of the OATs with protein partners may change OAT function [24, 25]. Indeed T3, DHEAS, TEST and DEX were proposed by several authors [26, 44, 45] as potential inducers of transporter expression. Accordingly PKC were treated with T3, DHEAS, TEST, BSP, PAH, TCA and DEX and tested whether this would result in a more stable expression of transporters. However, none of the tested substances showed any stimulatory effect on transporter mRNA expression in the tested concentration and time frame. Whether this is due to species differences or to the lacking tissue structure compared to renal slices or tubules used in other studies [44–46] remains to be elucidated.