Improved Cell Line IPEC-J2, Characterized as a Model for Porcine Jejunal Epithelium

Cell lines matching the source epithelium are indispensable for investigating porcine intestinal transport and barrier properties on a subcellular or molecular level and furthermore help to reduce animal usage. The porcine jejunal cell line IPEC-J2 is established as an in vitro model for porcine infection studies but exhibits atypically high transepithelial resistances (TER) and only low active transport rates so that the effect of nutritional factors cannot be reliably investigated. This study aimed to properly remodel IPEC-J2 and then to re-characterize these cells regarding epithelial architecture, expression of barrier-relevant tight junction (TJ) proteins, adequate TER and transport function, and reaction to secretagogues. For this, IPEC-J2 monolayers were cultured on permeable supports, either under conventional (fetal bovine serum, FBS) or species-specific (porcine serum, PS) conditions. Porcine jejunal mucosa was analyzed for comparison. Main results were that under PS conditions (IPEC-J2/PS), compared to conventional FBS culture (IPEC-J2/FBS), the cell height increased 6-fold while the cell diameter was reduced by 50%. The apical cell membrane of IPEC-J2/PS exhibited typical microvilli. Most importantly, PS caused a one order of magnitude reduction of TER and of trans- and paracellular resistance, and a 2-fold increase in secretory response to forskolin when compared to FBS condition. TJ ultrastructure and appearance of TJ proteins changed dramatically in IPEC-J2/PS. Most parameters measured under PS conditions were much closer to those of typical pig jejunocytes than ever reported since the cell line’s initial establishment in 1989. In conclusion, IPEC-J2, if cultured under defined species-specific conditions, forms a suitable model for investigating porcine paracellular intestinal barrier function.


Introduction
In intensive pig farming, a significant fraction of piglets die after weaning, in many cases due to infectious diarrhea [1]. Intense research efforts are therefore made to reduce mortality in animal breeding. For molecular studies on mechanisms and signaling pathways between germ exposure and diarrheal effect, porcine cell cultures are highly desirable. However, these cultures are only suitable if they closely match the properties of pig small intestinal epithelium. Thus, for research on intestinal barrier function, cell models have to meet specific physiological requirements: reflecting epithelial architecture, displaying adequate transepithelial resistance (TER) and transport properties, reacting to secretagogues, and expressing bowel-relevant tight junction (TJ) proteins. If these prerequisites have been achieved, the model system will be potentially suitable for studying effects of e.g. nutritional factors.
In contrast to cultures of rodent cells, a unique side aspect of porcine cell culture models is the potential application for human purposes because the pig gastrointestinal tract physiology is highly comparable to that of humans [6]. It immediately stands out, compared to other commonly used intestinal cell lines (CMT-93, TER: 400 V?cm 2 [7]; HT-29/B6, TER: 500 V?cm 2 [8]) and pig bowel mucosa (R epi : 55 V?cm 2 , [9]), that all porcine cell lines mentioned above exhibit extraordinarily high TER values (1 to 15 kV?cm 2 ) when believed to be fully differentiated by the respective author [5,[10][11][12]. TER is a key parameter of epithelial tightness and is determined by para-and by transcellular processes [13]. The paracellular pathway between enterocytes is limited by the TJ which is formed by opposing transmembrane TJ proteins and mediates different degrees of tightness. The TJ is of central interest as it forms a barrier against uptake of putatively immunogenic macromolecules and an excessive passage of water, small ions, and other solutes [14]. The transcellular pathway through enterocytes is defined by tissue-specific channels and carriers, passive diffusion of lipophilic solutes, and complex transcytosis of large molecules. The jejunal layer is a leaky epithelium which is defined by a ratio of para-and transcellular resistances as R para /R trans ,1 [15]. With respect to the observed high TER values it is questionable whether IPEC-1, IPEC-J2, and PSI could serve as appropriate models reflecting porcine small intestinal epithelium, however, they are often employed as such.
Before using an in vitro cell culture model as an in vivo substitute, it has to be characterized functionally, morphologically, and on a molecular level. So far, most work has been carried out on IPEC-J2. Generated in 1989 by Berschneider, IPEC-J2 were judged as a usable model for research on jejunocyte differentiation and ion transport. This result was based on confluent monolayers of cuboidal to columnar-shaped jejunocytes, the presence of typical cell-cell contacts and marker enzymes, inducible Cl 2 secretion, and adequate TER (549639 V?cm 2 ). In the following 17 years little research was done on IPEC-J2. However, during that interval, IPEC-J2 electrophysiology appears to have changed, as in 2006 Schierack et al. re-characterized the cell line and found a strongly increased TER (1.2 to 6.5 kV?cm 2 , depending on membrane support material) [16]. Nonetheless, a lot of further studies, mainly on microbe-associated adhesion/invasion, were performed, based on the sustained suitability for this research field because relevant parameters had remained unaltered [16][17][18][19][20]. However, IPEC-J2 are also used for virus, nutrition, and toxicity research and when transport function and/or barrier properties are relevant, it has to be guaranteed that trans-and paracellular routes represent the in vivo situation as closely as possible.
As conventionally cultured IPEC-J2 monolayers differ considerably from porcine intestine, we aimed to establish a porcine jejunocyte cell culture model which closely matches (electro-)physiological pig jejunal properties and epithelial architecture in order to apply it in swine research on transepithelial transport and paracellular intestinal barrier function.
Cells were grown at 37 uC, 5% CO 2 , and 95% relative humidity. They were fed every other day.

Ethic statement
Experiments were conducted on intestinal tissue of weaned piglets in strict accordance with the German law for the care and use of experimental animals. All procedures involving animal handling were approved by the local state office of occupational health and technical safety (Landesamt für Gesundheit und Soziales Berlin, Permit Number: G 0347/09).

Jejunal tissue preparation
Piglets were housed and fed control diets. At age 5461 days, piglets were sedated with 20 mg/kg BW of ketamine hydrochloride (Ursotamin, Serumwerk Bernburg) and 2 mg/kg BW of azaperone (Stresnil, Jansen-Cilag) and killed by intracardial injection of 10 mg/kg BW of tetracaine hydrochloride, mebezonium iodide, and embutramide (T61, Intervet) and exsanguination (see also [22]). The mid jejunum was removed, cut open, and rinsed with and transported in cooled saline solution (0.9% NaCl, 1 mM CaCl 2 ). Jejunal tissue was stripped off the muscle layer and explants either were mounted in Ussing chambers, fixed in 2% paraformaldehyde (PFA) or frozen in liquid nitrogen.
In order to block apical Na + and K + channels in IPEC-J2, amiloride (Sigma, final concentration 10 mM), tetraethylammonium (Sigma, 5 mM), and barium chloride (Merck, Darmstadt, Germany, 5 mM) were added apically prior to basolateral administration of forskolin (Calbiochem, Merck, 10 mM) or apical addition of phlorizin (Sigma, 0.5 mM) in plain Ringer bath solution. Apical administration of glucose (Roth, Karlsruhe, Germany, 10 mM) followed cation channel block in glucose-free standard solution. In jejunal tissue studies forskolin, phlorizin, and glucose were applied without previous cation channel block in standard Ringer solution supplemented with substrates and antibiotics or glucose-free standard solution, respectively.

Electrophysiological measurements
TER progression of IPEC-J2 monolayers was monitored over weeks using a chopstick electrode with automatic height control [23]. TER measurements were corrected for membrane and culture medium resistance and multiplied by the effective monolayer area.
For Ussing chamber experiments, IPEC-J2 monolayers (0.6 cm 2 effective area) grown on membrane supports were directly mounted [24], whereas stripped jejunal explants were glued to Plexiglas TM rings (0.28 cm 2 effective area) before being mounted [25]. Ussing chambers and water-jacketed gas lifts were kept at 37uC. Preparations were allowed to equilibrate for 30 (monolayer) or 45 min (tissue), respectively. Resistance of bath solutions alone and electrode offsets were recorded prior to each experiment and subtracted from experimental data. Data produced via chopsticks and in the Ussing chamber were fully compatible.
In order to determine Na + /Cl 2 permeability ratios of IPEC-J2 and jejunal tissue, dilution potential measurements were conducted in the Ussing chamber. After equilibrating monolayers/tissue in phosphate-free Ringer solution, dilution potentials were evoked by iso-osmotically partially replacing NaCl with mannitol. Permeability ratios to Na + /Cl 2 were calculated using the Goldman-Hodgkin-Katz equation as reported before [26]. Dilution potential measurements were performed at 37 uC, except for IPEC-J2/FBS, where part of the experiments were carried out at 25 uC in order to minimize interfering potentials from active, transcellular transport. Nevertheless, values of dilution potentials at both temperatures were not significantly different.
Permeability to fluorescein was determined in the Ussing chamber under voltage clamp conditions. After equilibrating cell layers in standard Ringer solution, fluorescein (Sigma) was added apically (final concentration 100 mM). After 0, 10, 20, 30, 40, and 50 min post administration basolateral samples were taken and replaced with Ringer solution. Fluorescein concentrations were determined at 525 nm fluorescence emission (Infinite M200, Tecan, Crailsheim, Germany) and permeabilities were calculated.

Impedance spectroscopy
In order to discriminate between epithelial (R epi ) and subepithelial (R sub ) resistances, one-path impedance spectroscopy was performed as described previously [25]. Briefly, membrane supports or jejunal explants were mounted in Ussing chambers modified for impedance measurements [27]. A total of 48 frequencies of alternating current (1.3 Hz to 65 kHz) were applied and resulting voltages analyzed through a programmable frequency generator/response analyzer in combination with an electrochemical interface (1250 and 1286, Solartron, Schlumberger, Farnborough, UK). Complex impedance values were plotted as Nyquist diagrams and fitted by circular arcs using least square analysis. The arc intercept with the x-axis at low frequency represents TER and that at high frequencies equals R sub . TER minus R sub represents the true epithelial resistance, R epi . The frequency at which the semicircle has its minimum is used to calculate the epithelial capacitance (C epi ) [8]. For splitting R epi in trans (R trans )-and paracellular (R para ) resistances, two-path impedance spectroscopy was conducted as described by Krug et al. [8]. In brief, this technique combines one-path impedance spectroscopy and flux measurements of a paracellular marker (i.e. fluorescein) during a Ca 2+ switch experiment.

Live cell imaging
Membrane supports covered with IPEC-J2 monolayers were cut out, positioned on a cover slip (apical side down), and covered by a small volume of bicarbonate-free Ringer solution. 20 ml of 4 kDa FITC-dextran (TdB, Upsala, Sweden, 25 mM) were added from the basolateral (upper) side to let the dye flood the paracellular space and surroundings. An inverted confocal laser-scanning microscope (LSM 510 Meta, Zeiss, Jena, Germany) was used to take images in xy and xz plane, which were processed using fiji imaging software [28]. Cell heights of three independent seedings were measured applying Zeiss LSM 510 META software.
Jejunal tissue. Slices of PFA-fixed tissue were boiled in ethylenediaminetetraacetic acid buffer (1 mM EDTA, pH 8.0 adjusted with NaOH; Merck) for 15 min and then washed in PBS + , whereas slices from tissue directly frozen in liquid nitrogen were fixed in methanol at -20uC for 10 min and subsequently washed in PBS + . Tissue slices were then permeabilized (5 min, 0.5% Triton X-100 in PBS + , RT) and blocked in 5% goat (PAA) or donkey serum (Sigma) plus 1% bovine serum albumin (Roth) for 1 h. Tissue sections were incubated with primary antibodies in blocking solution for 1.5 h at RT, washed, and incubated with secondary antibodies for 1.5 h at RT (Table S1). After washing with PBS + and doubly distilled water, slices were embedded using ProTaqs Mount Fluor.
Images were taken with a confocal laser-scanning microscope and software as introduced above. A defined image area ($400 mm 2 ) of IPEC-J2 grown on three independent membrane supports was used to determine cell width assuming an ideally hexagonally shaped cell corpus. Pig jejunal epithelium morphometry was assessed correspondingly using IF staining images.
To determine the epithelial enlargement factor provided by the villus and crypt surface with reference to the lamina muscularis mucosae, jejunal tissue of 13 piglets was fixed in triplicate for 26 h in Zamboni's fixation solution, dehydrated in graded series of ethanol, embedded in paraffin, cut to 5 mm thin sections, deparaffinized, and rehydrated. Three slides were prepared for each sample and stained according to H&E standard staining protocols [30]. Five section areas, where at least four villi were cut completely from top to bottom and crypts were cut vertically were analyzed at 50-fold magnification. Enlargement factors were measured according to Wiese et al. [31] using the image analysis program NIS-Elements (Nikon, Düsseldorf, Germany).

Western blot (WB) analyses
WB analyses were performed using standard techniques on IPEC-J2 of four consecutive passages or on stripped jejunal explants of four animals. Primary and secondary antibodies are given in Table S1.
The Lumi-Light PLUS Western Blotting Kit (Roche) was used to detect relevant protein bands via the Fusion FX 7 image acquisition system (Vilber Lourmat, Eberhardzell, Germany). Densitometric signal analysis was performed using AIDA software (Raytest, Berlin, Germany).

Proliferation assay
Cell proliferation was measured using the Cell Proliferation Reagent WST-1 (Roche) by quantifying the number of metabolically active cells. IPEC-J2 of four different passages were seeded at a density of 10 4 cells/well in 96-well microplates and maintained in a cell incubator for either 4 or 72 h. After each growing period, WST-1 Reagent was added and cells were incubated for 1 hour before the absorbance was measured at 450/630 nm using an ELISA reader (Bio-Rad, Munich, Germany). The absorbance at time point 4 h was set as 100%.

Statistical analysis
Data are expressed as means 6 standard error of the mean (SEM). Statistical analyses were carried out using either a twotailed, unpaired Student's t-test or a one-way ANOVA with Tukey HSD post hoc test (SPSS, version 20, Chicago, Illinois). Significances are depicted as: *, p,0.05; **, p,0.01; ***, p,0.001.

PS culture reduces transepithelial resistance
IPEC-J2 cultured in medium supplemented with FBS (IPEC-J2/FBS) and grown on permeable membrane supports exhibit huge TER values in the range of 1-15 kV?cm 2 either in a peak (e.g. [19]) or as plateau value (e.g. [32]; Fig. 1A, black and grey squares). In order to compare TER values of IPEC-J2 with values from pig jejunal epithelia, the jejunal surface enlargement by villi and crypts has to be taken into account. Morphometric studies revealed a surface enlargement factor (11.360.5, n = 13; range 8.7 to 14.7) which was set 10 for practicability reasons. Having corrected cell culture values (TER corr ) for the effective jejunal epithelial surface area, values measured in vitro and ex vivo are directly comparable. As shown in Fig. 1B, TER corr of IPEC-J2/ FBS (384612 V?cm 2 , n = 20) still differs by one order of magnitude from TER of pig jejunal epithelium (2865 V?cm 2 , n = 15).
While improving the cell culture protocol to obtain more physiological cell characteristics, fetal bovine serum was replaced by serum from adult pigs. As a striking result, 10% PS instead of FBS prevented cells from developing extreme TER over a time period of at least three weeks (Fig. 1A, grey triangles, black squares). However, culturing IPEC-J2 with culture medium containing 5% PS, ITS, and EGF (Fig. 1A, black triangles) resulted in a more regular cell layer (data not shown). Notably, under PS condition TER developed a maximum around day 7 post seeding and reached a lower plateau level (200-400 V?cm 2 ) at around day 14 post seeding which was maintained for several days.
To determine whether the effect of lower TER was serum species-specific, brought about by age of the blood donor animal or was caused by medium supplementation with ITS and EGF, appropriate medium compositions were tested (Fig. 1A). Neither medium supplemented with 5% FBS, ITS and EGF nor the use of adult bovine or adult goat serum considerably affected TER values and course compared to the 10% FBS condition. In addition, the use of porcine serum obtained from another company (Biochrom, Berlin, Germany) corroborated the effect of lower TER (data not shown). Resulting from this, 10% FBS and 5% PS with ITS and EGF (conventional and species-specific condition, respectively) were chosen for further comparative experiments. Pig jejunal TER (2865 V?cm 2 , n = 15, p,0.001) and TER corr of IPEC-J2/PS (2761 V?cm 2 , n = 20) did not differ significantly (Fig. 1B). When FBS condition was replaced by PS condition (and vice versa) two weeks post seeding, IPEC-J2 immediately started to develop respective serum-typical TER values (Fig. 1C), whereas switches induced by PS were much faster than those induced by FBS.

PS improves transport properties
Typical features of jejunum in general are ion transport induced by secretagogues as well as glucose absorption via sodium-glucose transporter 1 (SGLT1). IPEC-J2/PS had a significantly stronger secretory response to forskolin (nI SC corr , 4462 mA/cm 2 , n = 11, p,0.01; cell culture values corrected for jejunal surface enlargement) than IPEC-J2/FBS (1963 mA/cm 2 , n = 10) but comparable to that of porcine tissue (4868 mA/cm 2 , n = 9) ( Fig. 2A). Glucose absorption was either evaluated directly as glucose-stimulated nI SC or indirectly as nI SC observed upon SGLT1 inhibition by phlorizin. However, both alternatives did not demonstrate any change in glucose-dependent I SC between PS and FBS condition (data not shown).

PS alters cell morphometry and substructure
TER values depend on TJ composition as well as TJ length and cell membrane area per area of the membrane support, whereas I SC depends on the latter only. Total membrane area follows from cell morphology and is reflected by epithelial capacity (C epi ). C epi was increased by a factor of four in IPEC-J2/PS (4.0660.25 mF/ cm 2 , n = 15, p,0.001), compared to IPEC-J2/FBS  To gain further insight into these morphological changes, the influence of both culture conditions was analyzed ultrastructurally. IPEC-J2/PS exhibited a dome-shaped apical membrane with numerous and long microvilli compared to IPEC-J2/FBS (density, 7.061.1 vs. 2.760.4 microvilli/mm, n = 3, p,0.05; length, 450630 nm, vs. 110610 nm, n = 3, p,0.01; n, number of cells, analysis of .2 mm apical cell membrane each), (Figs. S2A, B). Tight junctional structures were located at the apical-most part of cell-cell contacts, but due to the dome shape of the apical cell membrane, IPEC-J2/PS TJs did not represent the maximum cell height (Figs. S2A, B). The amorphous substance (Fig. S2A, labeled with M) which is mostly present in IPEC-J2/PS could be identified as neutral mucopolysaccharides (Fig. S2C).  (Figs. 4A, B, C). In contrast, IPEC-J2/FBS differed considerably (meshwork depth, 577670 nm, p,0.01; horizontal strands, 7.0060.53, p,0.05; both n = 20) from jejunal tissue. The occurrence of continuous type TJ strands was lowest in IPEC-J2/FBS, medium in IPEC-J2/ PS and high in pig jejunal epithelium, whereas an inverse incidence of particle type structures was observed (Fig. 4D). Thus, there was no correlation between the presence of continuous TJ strands and high TER values.
In terms of epithelial polarization markers, the presence of ezrin and SGLT1 within the porcine apical membrane as well as GLUT2 and Na/K-ATPase within the basolateral membrane could be confirmed (Fig. 7A) and the absence of vimentin, a mesenchymal marker, within epithelial cells was verified (Fig. 7A). Under both cell culture conditions, the expression of GLUT2 and Na/K-ATPase within the basolateral membrane could be proved (Fig. 7A), whereas the expression of ezrin and SGLT1 within the apical membrane could be verified in IPEC-J2/PS only, since in IPEC-J2/FBS imaging apical and basal membranes spatially separately was difficult due to the low cell height (Fig. 7A). However, against expectations, IPEC-J2 were positive for vimentin (Fig. 7A). Protein quantities of ezrin and GLUT2 (p,0.01), SGLT1 (p,0.05), and the mesenchymal marker vimentin (p,0.01) were increased in IPEC-J2/PS, when compared to IPEC-J2/FBS (Fig. 7B).   Considering that the jejunal lining is enlarged by villi and crypts, cell culture values were corrected using an enlargement factor of 10 which had been obtained from cross-sections of porcine jejunum. The resulting TER discrepancy between IPEC-J2 and jejunal tissue still amounts to one order of magnitude. In the current study we demonstrated that using a species-matched serum caused IPEC-J2 to develop an improved morphology accompanied by dramatically lower TER, resembling pig jejunocyte features.

Relation between morphometry and TER
TER reflects two resistors in parallel, R trans and R para . Assuming a constant R para per unit TJ length and a constant number of ion channels per unit membrane, then cell layers consisting of small, tall cells will have lower TER values than cell layers consisting of large, flat cells. Microvilli and membrane invaginations which enhance the membrane area of the cells will further decrease TER [13]. Thus, it has been shown in a study which focused on a porcine colonic cell culture model [38], that cell monolayers possess adequate TER, which meet pig colonic tissue dimensions when corrected for the respective epithelial enlargement by crypts. Concomitantly, cells were found to exhibit appropriate cell dimensions, abundant and well-shaped microvilli as well as tight junctional and further intercellular contacts. A similar relationship between morphometry and TER has been observed in differently cultured human bronchial epithelial cells [39] as highly columnar shaped Calu-3 cells exhibit lower TER than cuboidal ones. In conventionally cultured IPEC-J2 monolayers, high TER values which do not correspond to the ex vivo counterpart are mainly a consequence of the enormous cell size. In addition, microvilli were small and sparse. In contrast, species-specific cultivation remodeled large and flat IPEC-J2 into smaller and taller ones with numerous long microvilli and this change in morphology was paralleled by a dramatic change in TER.
Previous ultrastructural studies demonstrated a correlation between the number of horizontal TJ strands and TER (or, more exactly, R para ) [40,41] even though other factors, such as TJ protein composition, evidently interferes with this correlation [42]. Thus, although in the current study total TJ length appeared to be the major determinant of TER, TJ ultrastructure may also have affected TER, as number of TJ strands and network depth were higher in IPEC-J2/FBS.
Besides TJ length and ultrastructure, the change in membrane area is likely to contribute to the drastic differences in TER between both culture regimes. These differences result from increased cell height, microvilli length and number, apical invaginations, and lateral infoldings in IPEC-J2/PS. The increase in membrane area can be directly deduced from the increase in epithelial capacitance observed during impedance measurements.
In IPEC-J2 of both kind, the ratio R para /R trans is smaller than 1. Despite the high absolute resistances, this indicates a "leaky" epithelium [15] as it is typical for the jejunal lining.
Generally, the ion permeability of tight junctions is strongly augmented by the paracellular cation channels claudin-2 and claudin-15. Both claudins were present in porcine jejunum but were not detectable in IPEC-J2 of either kind. This may explain the lower cation selectivity of IPEC-J2 compared to that of jejunum.
Furthermore, forskolin-stimulated chloride secretion in IPEC-J2/PS was enhanced compared to IPEC-J2/FBS, indicating enhanced currents from the enlarged apical membrane area (dome-shaped, microvilli). Potential effects of higher channel density per unit membrane area were not investigated but may add to the observed current.

Comparison of cell culture conditions inducing morphometry changes
Cells grown on membrane supports are commonly fed by medium from both the apical and the basolateral compartment. Due to poor IPEC morphology, Nossol and co-workers [32] suggested that the monolayer's oxygen supply associated with ATP-dependent cellular biochemical and transport processes may be insufficient. In order to augment differentiation and functionality of IPEC monolayers, Nossol et al. performed ALI cultivation which optimized the cell height/diameter relation of IPEC-1 but did not improve these parameters in IPEC-J2. Inversely, TER of IPEC-1 did not show relevant changes two weeks after switching to ALI mode, whereas TER of IPEC-J2 declined by almost 50%, nevertheless retaining values at kV?cm 2 level. The fact that morphometry and TER act contrarily to the TER/morphometry hypothesis postulated above indicates that TJ structure and/or membrane channel density must have changed greatly during ALI culture. As yet, the mechanisms responsible for changes in cell morphology under ALI conditions remain unclear.
Alternatively, it is not unlikely that morphological changes observed under ALI conditions and in the presence of PS are two sides of the same coin. Changes observed under ALI conditions may be caused by an accumulation of apically secreted substances within the very thin residual apical liquid layer. Thus, it can be hypothesized that IPEC-1 are able to synthesize components affecting cell morphology, whereas IPEC-J2 are not or only to a minor degree, so that IPEC-1 effectively accumulate signaling agents only when the apical extracellular volume is restricted. In porcine serum, whose proteome is not yet fully analyzed [43], all required signaling agents associated with differentiation/maturation presumably are included, so that IPEC-J2 morphometry and functionality was altered by replacing FBS with PS in culture media.
In all, FBS seems to be insufficient as medium supplement for IPEC-J2 culture although adequate cell morphometry and electrophysiology had initially been shown by Berschneider [4]. Porcine serum by contrast, apparently has the ability to compensate for a substance whose secretion by IPEC-J2 has been lost with increased number of passages. It needs to be elucidated, whether the saving serum agent is solely included in porcine serum and thus does not exist in bovine or goat sera of any age or whether its sequence differs between these species, leading to a loss of function of the potential bovine/goat analog in porcine cell culture.

Differentiation and maturation
Once arisen from pluripotent stem cells located in the deep crypt epithelium, proliferative jejunocyte progenitor cells functionally and structurally start to differentiate during their migration along the crypt-villus axis. When the villous epithelium is reached, enterocyte differentiation is completed and jejunocyte maturation begins. The terms differentiation and maturation are often used interchangeably. However, differentiation describes quality changes, whereas maturation is the process of quantity changes in cell phenotype [44]. In order to judge the differentiation state of cultured epithelial cells, a variety of different markers could be used. Markers comprise cell type specific proteins which progressively complement the cell phenotype.
We tested protein expression of exemplarily chosen marker candidates, which are specific to polarized, barrier forming, and transport-active jejunal epithelial cells.
It was evident that all epithelial marker molecules tested, except claudin-2, -12, and -15, were qualitatively available under either growing condition. This suggests that IPEC-J2/FBS and IPEC-J2/ PS display the same differentiation status. However, differences existed in terms of marker quantity. Compared to FBS condition, apical and basolateral marker proteins (ezrin, SGLT1, GLUT2) were significantly higher expressed, suggesting an advanced maturation state under PS culture. Having shown that IPEC-J2 were differentiated, the mesenchymal marker vimentin, against expectations, was excessively expressed under both culture regimes. This phenomenon also appears e.g. in differentiated bovine primary culture jejunocytes, where it was attributed to a suppression of post transcriptional inhibition of the vimentin synthesis [45]. The appearance of vimentin may suggest an onset of epithelial-mesenchymal transition (EMT). EMT is triggered by several signaling pathways which induce the transcription factor snail. This, in turn, transcriptionally represses E-cadherin and TJ proteins [46] and in parallel increases vimentin expression [47]. Since snail expression was not elevated and junctional proteins were not down-regulated, a dedifferentiation of IPEC-J2 can be excluded. Incidentally, many established epithelial cell lines (MDCK-C7 [48], MDBK [49]) as well as epithelial cells in vivo [50] also express vimentin.
Together with respective protein qualities and quantities, the time course of TER reflects the differentiation/maturation state of either culture conditions. TER of IPEC-J2/FBS and IPEC-J2/PS reached highest values about one week post seeding which could indicate complete differentiation. Under FBS condition, TER values then remained at a more or less constant level, whereas under PS condition TER declined and reached lower plateau values arguing for a maturational process with leveled protein composition. In the absence of evidence to the contrary we can assume that the loss of internal maturational competence in IPEC-J2 is one disadvantage of non-immortalized continuous cell lines. In order to compensate such loss of function events, speciesspecific cultivation represents possibly the most effective proceeding.

Conclusion
IPEC-J2 represent a unique tool for investigating porcine jejunal barrier function ex vivo. We describe an improved speciesspecifically cultured model, which exhibits morphology and barrier parameters close to the source tissue. This was urgently needed in order to replace rodent, tumor-derived, and nonphysiologically behaving swine cell models in the field of pig small intestine and digestive research.