Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Transport across the intestinal epithelium is often the rate-limiting step in the total process of absorption for orally administered hydrophilic macromolecular compounds. Passage of these compounds is most likely to occur extracellularly by passive diffusion through the paracellular pathway. However, tight junctions abolish passage of large molecules almost completely. Co-administration of an absorption enhancer is a potential way to increase bioavailability of these compounds, and considerable effort has been directed toward identifying agents able to loosen tight junctions, thus increasing paracellular permeability.

Although the exact mechanisms of action are not fully identified, in general two classes of absorption enhancers are distinguished: surfactants and calcium chelators (Hochman and Artursson, 1994). Surfactants (bile acids and salts, derivatives of fatty acids, Triton X-100 and so on) act by increasing the solubility of hydrophobic macromolecules in the aqueous boundary layer or by increasing the fluidity of the apical (and the basolateral) membrane. Calcium chelators (EGTA and EDTA) reduce the extracellular calcium concentration, which leads to disruption of cell-cell contacts. In general, the surfactants increase transcellular permeability, whereas the chelators increase paracellular permeability. Acylcarnitines seem to be an exception; it has been suggested that they enhance paracellular permeability in a calcium-independent way.

The absorption-enhancing properties of these fatty acid derivatives of L-carnitine have been studied extensively (Fix et al., 1986; LeCluyse et al., 1993; Sutton et al., 1993a + b). The C16 conjugate PCC was found to be very effective in increasing transepithelial transport of poorly absorbed drugs in vivo in rats and dogs (Fix et al. 1986; Sutton et al., 1993b) and in vitro in intestinal segments (Sutton et al., 1993a; LeCluyse et al., 1993) and Caco-2 cells (Raeissi and Borchardt, 1993; Hochman et al., 1994). Histological and ultrastructural examination of sections of rat jejunum and colon exposed to PCC showed slight alterations in the microvillus border but an intact cytoplasmic integrity and uncompromised junctional complexes (Fix et al., 1986; Sutton et al., 1993). To elucidate the mechanism by which PCC exerts its action, in vitro studies were performed using Caco-2 cells (Raeissi and Borchardt 1993; Hochman et al., 1994). These studies have shown an immediate drop in transepithelial electrical resistance (TER) and an increase in mannitol transport after apical exposure to submillimolar concentrations of PCC. The drop in TER proved to be transient and returned to approximately control values in 10 to 12 h. No visible damage to the epithelium and no alterations in the filamentous actin were seen, but the appearance of tight junctions in freeze-fracture replicas seemed to be altered; that is, junctional strands showed a beaded appearance, several discontinuities and diminished cross-bridging. These results might indicate that PCC is able to increase permeability reversibly without causing major morphological alterations in the intestinal epithelium. On the other hand, other studies have shown that PCC causes damage to erythrocytes (Cho and Proulx, 1971) and the vaginal epithelium (Richardson et al., 1992) at concentrations likely to be used to achieve absorption-enhancing effects. In addition, LeCluyse et al. (1993) reported that after PCC treatment, lucifer yellow transport could be enhanced without observed exfoliation of cells, however, an enhancement factor of 18 for lucifer yellow transport was accompanied by extensive cell exfoliation from rat colonic mucosa.

The exact mechanism by which PCC exerts its permeability-enhancing action has still not been elucidated. Moreover, the sequence of events (transport enhancement, membrane damage and tight junction disruption) is unknown. Therefore, we studied the relationship among absorption enhancement, cell viability and tight junction protein (ZO-1) distribution in the human colonic Caco-2 cell line (Hidalgo et al., 1989; Baily et al., 1996; Artursson et al., 1996) and the rat small intestinal cell line IEC-18 (Ma et al., 1992; Duizer et al., 1997). These two intestinal epithelial cell lines differ greatly in state of differentiation, TER and paracellular permeability. Both cells express ZO-1 and discriminate paracellular transport rates on the basis of molecular radius (Duizer et al., 1997). Filter-grown cells were exposed to 0 to 1 mM PCC for 30 min; TER and transport of mannitol and PEG-4000 were used as parameters for epithelial permeability to hydrophilic (macro)molecules. LDH leakage into the apical medium, PI and NR uptake were used to detect reduced integrity of the apical cell membrane and reduced cell viability. For Caco-2 cells, permeability enhancement, TER reduction and LDH leakage resulting from PCC exposure were compared with the effects of the surfactant Triton X-100 and the chelator EGTA. Additionally, the immunolocalization of the tight junctional protein ZO-1 was correlated with PCC exposure, and a method was developed to quantify ZO-1 staining using CSLM images.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. The Caco-2 cell line originating from a human colorectal carcinoma and IEC-18 cells originating from rat ileum epithelium were obtained from the American Type Culture Collection. Cells were maintained at 37°C in an atmosphere of 5% CO₂ and 90% relative humidity. Maintenance medium was DMEM with high glucose (4.5 g/l) and 25 mM HEPES, supplemented with 1% (v/v) MEM nonessential amino acids, 6 mM L-glutamine, 50 µg/ml gentamycin (Gibco, Paisley, Scotland) and 10% (v/v) FCS (Integro, Zaandam, Netherlands) for Caco-2 and 5% (v/v) FCS and 0.1 U/ml insulin (Sigma, Beers, Belgium) for IEC-18 cells.

TER measurements. Filter-grown cells were adapted to room temperature to allow for more standardized TER measurements using the Millicell-ERS epithelial voltohmmeter (Millipore Co., Bedford, MA). The TER was measured before and during the experiments to monitor cell layer confluence and the integrity of the tight junctions. The TER of the cell layers was calculated according to the equation: where R_total is the resistance measured, R_blank is resistance of control filters without cells (approximately 140  · cm²) and A is the surface area of filter (1 cm²).

Exposure to PCC. Caco-2 cells with a TER > 600  · cm² and IEC-18 cells with a TER > 40  · cm² (both determined at room temperature) were transferred to new wells with 1.8 ml of basal medium (HEPES-buffered DMEM with nonessential amino acids, 6 mM L-glutamine, gentamycin and 0.1% BSA). The apical medium was changed for 0.5 ml of basal medium supplemented with 0 to 1 mM freshly dissolved PCC (Aldrich, Milwaukee, USA). After 30 min of preincubation at room temperature, the apical test medium was removed and used for LDH leakage measurements. Subsequently, the inserts with cells were used for assessment of transepithelial transport or immunocytochemistry.

Exposure to Triton X-100 and EGTA. The protocol for Triton X-100 exposure (0-0.05% v/v) was exactly the same as that for PCC incubation. The studies with EGTA (0-5 mM) were performed by preincubation with EGTA for 45 min at room temperature. During the preincubations, EGTA was applied apically and basolaterally. After the preincubation, the TER was measured and apical medium was collected for LDH measurements. Transport of mannitol and PEG-4000 was performed with EGTA basolateral.

Transepithelial transport. Transport experiments were carried out at 37°C in Transwell inserts with 0.5 ml of basal medium with tracers apically and 1.8 ml of basal medium basolaterally. A reduction of the unstirred water layer and homogeneous mixing of the probes were achieved by rotating the two-compartment transport system at 30 rpm on a rotating platform device in an incubator (Dulfer et al., 1996). Apparent permeability coefficients (Papps) were determined on the basis of appearance of the probe in the basolateral receiver compartment before 10% of the probe was transported (under sink conditions) according to the following equation (Artursson and Karlsson, 1991): where P = permeability rate (mol/s), C₀ = initial apical concentration of test substance (mol/ml), A = area of filter (cm²). The probes ¹⁴C-polyethylene glycol-4000 and ³H-mannitol (Amersham, Little Chalfont, UK) were mixed with unlabeled compounds to yield final concentrations of 2.5 µM with specific activities of 2.5 GBq/mmol for ¹⁴C-polyethylene glycol-4000 and 1.5 GBq/mmol for ³H-mannitol. ³H- and ¹⁴C-labeled probes were analyzed using a Wallac 1409 Liquid Scintillation Counter after addition of Ultima Gold Scintillation liquid (Packard, Groningen, Netherlands).

Immunocytochemistry. For immunocytochemistry the cells were fixed in 1% formaldehyde in HBSS and permeabilized in 0.2% Triton X-100 in HBSS. F-actin was stained with phalloidin conjugated with TRITC or FITC (Sigma, St. Louis, MO). Polyclonal rabbit anti-ZO-1 and polyclonal anti-Occludin, anti-rabbit IgG-biotin and streptavidin-FITC (Zymed, San Francisco, CA) were used to visualize ZO-1 and occludin, respectively, in both cell lines.

CSLM and image processing. CSLM was performed via the Bio-Rad MRC600, with the PlanApoChromat 40× objective with variable diaphragm (Zeiss, Oberkochen, Germany). All images were recorded at standardized acquisition conditions; per-cell-line and per-experiment settings for gain and blacklevel were optimized and applied throughout that experiment. Images are single plains (a depth of 7 µm covers >90% of the total intensity) recorded using the Kalman filter mode at n = 5 with Bio-Rad software (COMOS version 7.0a). After transfer to an image-processing workstation, the images were processed by a custom-written program on the basis of the image-processing toolbox SCIL-Image (version 1.3, Free University of Amsterdam, Netherlands). The program developed to select junction-related immunolabeled ZO-1 consists of three sequential steps. First, two methods are applied to differentiate between borderline and background with a threshold value of 1.4. Both resulting binary images are combined into a usually honeycomb-like skeleton. This intermediate skeleton represents most of the junction-related ZO-1 pattern with small discontinuities and some appendixes not junction-related. In the second part of the program, the skeleton can be improved, first by closing the small discontinuities of the borderlines by using a maximum cost procedure (adding less than 0.5% to the total pixel count) and second by interactive and automatic removal of spurious line fragments. Finally, the total number of pixels is determined from the binary image of the optimized skeleton, and the original grey value image is used to determine the average pixel intensity over the borderlines.

Cytotoxicity and membrane damage assays. LDH leakage into the apical medium resulting from PCC, Triton X-100 or EGTA exposure was determined with the Boehringer Mannheim kit for LDH at the BM/Hitachi 911. Total LDH content was determined after sonicating the inserts with cells in basal medium. In several experiments, 5 µg/ml of propidium iodide was added to the PCC exposure medium, and uptake of this membrane-impermeant fluorescent dye during PCC exposure was determined using images made with a standard fluorescence microscope from Zeiss equipped with a PlanApoChromat 40/1.0 (Zeiss, Oberkochen, Germany) objective and a COHU monochrome CCD camera. Quantification was performed with the Colourmorph program (Perceptive Instruments, Halstead, UK). Each image yielded 130 to 160 Caco-2 cells or 40 to 60 IEC-18 cells; five images were recorded per filter (n = 6 per exposure group). NR uptake was determined directly after the transport experiments. For that purpose, cells were incubated with NR (50 µg/ml) in control medium for 30 min (37°C, 5% CO₂, 95% relative humidity). Subsequently, the cell layer was rinsed once with HBSS, and the NR was extracted with 1% (v/v) acetetic acid in a 1:1 water/ethanol mixture. The A_540nm was measured spectrophotometrically.

Statistics. We obtained data on PCC (TER, LDH leakage, transport of mannitol and PEG-4000 and NR uptake) in four independent experiments with both cell types using three filters per exposure group. Data on EGTA and Triton X-100 reflect the means of three independent experiments. Data on immunolabeling of ZO-1 were obtained in two experiments using two filters per exposure group per experiment. For each filter, five areas were recorded for junction-related ZO-1-specific fluorescence quantification. The data presented are pooled data of all experiments. Data were analyzed with a two-tailed Student's (paired) t test.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Effects of PCC on TER. In both cell lines, PCC decreased TER in a dose-dependent manner (fig. 1). The reduction of the TER caused by 30 min of PCC exposure was statistically significant at concentrations as low as 0.1 mM in IEC-18 and 0.4 mM in Caco-2. Most of the TER reduction was already accomplished within 5 min after exposure (data not shown). At the higher PCC concentrations, the decrease in TER was more pronounced in Caco-2 than in IEC-18. For instance, 1 mM PCC reduced the TER in Caco-2 to 29% (190  · cm²) and in IEC-18 to 62% (40  · cm²). Note that the absolute TER of the 1 mM exposed Caco-2 monolayer was still higher than the TER of control IEC-18 cell layers (64.5 ± 12  · cm²).

View larger version (22K): [in this window] [in a new window]   Fig. 1.   Effects of 30 min of PCC exposure on apparant permeabilities of mannitol () and PEG-4000 () and on TER () of Caco-2 (panel A) and IEC-18 (panel B) monolayers. TER was normalized with the TER before PCC exposure as 100%. Data are presented as mean ± S.D. (n = 4 experiments). * Significantly different from control, P < .05, according to two-tailed Student's paired t test.

Effects of PCC on transepithelial transport rates. Transport rates of mannitol and PEG-4000 were increased by PCC in a dose-dependent manner in both cell lines (fig. 1). The apparent permeabilities of both compounds were significantly increased at PCC concentrations 0.2 mM in IEC-18 and 0.4 mM in Caco-2 cells. Although the IEC-18 cells were affected at lower concentrations of PCC, the maximum relative increase in transport obtained with 1 mM PCC was larger in Caco-2 cells. The fluxes of mannitol (except for the higher PCC concentrations) and PEG-4000 were constant throughout the 4 h that transport was monitored (data not shown), which indicates that the effects of 30 min of PCC exposure on transport did not change within 4 h after its removal.

Effects of PCC on ZO-1 distribution. Immunocytochemical localization of ZO-1 in untreated layers of Caco-2 and IEC-18 cells showed a belt-like cell-circumscribing pattern throughout the total area (figs. 2A and 3A). In Caco-2 cells only local and subtle modifications of the ZO-1 pattern were seen at lower PCC concentrations. Up to a concentration of 0.4 mM PCC, no interruption of the cell-circumscribing pattern was detected. At 0.7 mM, small holes the size of only few cells were found in the monolayer, and the pattern of junction-related ZO-1 was locally disturbed (fig. 2B). Exposure to 1.0 mM PCC caused holes that varied in size from a few up to many cells, but cells with a cell-circumscribing ZO-1 pattern were still present (fig. 2C).

View larger version (128K): [in this window] [in a new window]   Fig. 2.   CSLM image showing immunolocalization of ZO-1 in control (panel A), 0.7 mM (panel B) and 1.0 mM (panel C) PCC-treated Caco-2 monolayers. Panels D, E and F are the binary images of panels A, B and C, respectively, after processing. The custom-written program selected the honeycomb-like tight junctional pattern.

View larger version (125K): [in this window] [in a new window]   Fig. 3.   CSLM image showing immunolocalization of ZO-1 in control (panel A), 0.4 mM (panel B) and 1.0 mM (panel C) PCC-treated IEC-18 cell layers. Panels D, E and F are the binary images of panels A, B and C, respectively, after processing. The custom-written program selected the honeycomb-like tight junctional pattern.

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Effects of PCC exposure on immunolocalization of ZO-1 in Caco-2 (panel A) and IEC-18 (panel B) cells. A custom-written program selected the junction-related ZO-1 pattern; shown are selected pixel number () and selected pixel density (). Data are presented as means ± S.D. (n = 10). * Significantly different from control, P < .05 according to Student's t test.

Effects of PCC on apical membrane integrity and cell viability. Cytotoxic effects of PCC were determined by three different assays. LDH leakage and PI staining are based on detection of apical membrane damage, and the NR assay is based on two processes: nonionic passive diffusion through the apical membrane and intracellular accumulation in lysosomes (Babich and Bohrenfreund, 1990). Figure 5 shows that the release of the cytosolic enzyme lactate dehydrogenase into the apical medium increased in a dose-dependent manner after exposure to PCC. Leakage of this enzyme (MW 140 kDa) is significantly increased at doses of 0.2 mM PCC in Caco-2 and 0.1 mM in IEC-18 cells. Staining of cells with the DNA dye propidium iodide (MW 668 Da) occurred at higher PCC concentrations and was significantly increased at 0.4 mM PCC in both cell lines. Next to apical membrane damage, PCC caused a dose-dependent reduction of cell viability as assessed by the NR assay. Because we found increased apical membrane permeability (LDH leakage and increased PI staining) after PCC exposure, we assume that lower A₅₄₀ readings in the NR assay indicate a reduced functionality of the lysosomes (NR retention), not a reduced uptake of NR.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Effects of 30 min of PCC exposure on apical membrane integrity and cell viability of Caco-2 (panel A) and IEC-18 (panel B) cells. LDH leakage (n = 4 experiments) is normalized with sonificated untreated cells as 100%. NR uptake (n = 4 experiments) and PI staining (n = 6 filters from two independent experiments) are presented with the control set to 100%. Data are presented as means ± S.D. * Significantly different from control, P < .05, according to Student's paired t test.

Comparison of effects of PCC, Triton X-100 and EGTA on Caco-2 cells. In figure 6A the LDH leakage into the apical medium is presented as a function of the transport-enhancing factor for PEG-4000. For PCC and Triton X-100 a similar relationship was found, showing increased LDH leakage with increased transport-enhancing factors. For EGTA the increase in transport-enhancing factors was independent of LDH leakage. The relationship between TER and the enhancing factor for PEG-4000 (fig. 6B) is almost linear after exposure to Triton X-100, whereas after exposure to low concentrations of PCC or EGTA, the limited decrease in TER is not accompanied by increased PEG-4000 permeability. Figure 6C shows that when Caco-2 cells are exposed to PCC or Triton X-100, increased LDH leakage correlates with decreased TER. Data on EGTA exposure show that it is possible to decrease TER without causing membrane damage.

View larger version (16K): [in this window] [in a new window]   Fig. 6.   Effects of various concentrations of PCC (), Triton-X-100 () or EGTA () on Caco-2 cells. Expressed are the LDH leakage (panel A) and Normalized TER (panel B) as a function of the enhancing factor for PEG-4000 and the normalized TER as a function of LDH leakage (panel C). Data are presented as best-fitted curves and mean ± S.D. (n = 4 experiments for PCC; n = 3 experiments for Triton X-100 and EGTA).

Recovery of TER and cell viability after PCC exposure. To investigate whether the effects of PCC on TER and cell viability were transient, we allowed both cell lines to recover for 22 h after 30 min of exposure to 0, 0.4 and 1 mM PCC. In Caco-2 cells, TER recovered completely within 6 h after exposure to 0.4 mM PCC (table 1), whereas recovery of TER after exposure to 1 mM PCC was only partial even after 22 h. Cell viability as determined with the NR assay was still not restored 22 h after exposure to 0.4 or 1 mM PCC (table 1).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

The present study was intended to correlate absorption-enhancing effects of PCC with effects on tight junction morphology and cytoxicity on the intestinal epithelium. For that purpose, a small intestinal (IEC-18) and a colonic (Caco-2) cell line were apically exposed to PCC and TER, and transport of mannitol and PEG-4000 were determined. The immunolocalization of ZO-1, the uptake of PI and NR and the release of LDH were assessed under similar conditions. It was found that PCC caused a decrease in TER and a concomitant increase in the permeability for mannitol and PEG-4000 in both cell lines. In IEC-18 cells, the PCC-induced transport enhancement of the hydrophilic markers mannitol and PEG-4000 was slightly lower than in Caco-2. The finding that the absorption enhancement was smaller for small intestinal cells than for colonic cells is supported by studies of Sutton et al. (1993a), who used isolated ligaments of the rat and showed that ileal bioavailability of the drug cefoxitin was increased less than colonic bioavailability after co-administration with PCC. This indicates that the IEC-18 cell line could be a useful model not only for small intestinal paracellular transport (Duizer et al. 1997) but also for the study of transport enhancement across the small intestinal epithelium.

Both parameters, increased transport of mannitol and PEG-4000 and decreased TER, are generally considered to be indicators of increased paracellular permeability in experiments using cell lines. Nevertheless, Bernards and Kern (1996) have shown that an increased transmeningeal flux of hydrophilic compounds caused by PCC exposure resulted from an increased transcellular flux. Furthermore, increased transport of PEG-4000 and decreased TER are seen as the result of effects on membrane integrity, rather than a specific effect on paracellular permeability (Fagerholm et al., 1996). Thus, even though hydrophilic macromolecular permeability and TER are, under physiological conditions, indicators of paracellular permeability, in case of increased membrane permeability these parameters can indicate increased transcellular permeability as well.

We examined immunolocalization of ZO-1 in an attempt to visualize alterations in the paracellular pathway as a result of PCC exposure. We hypothesized that if the permeation-enhancing effect of PCC was due mainly to increasing the paracellular permeability, changes would be visible in the distribution of the tight junctional proteins. Indeed, PCC caused small irregularities in the ZO-1 pattern of Caco-2 and discontinuities in the ZO-1 pattern of IEC-18 cells. In order to quantify the changes in the ZO-1 pattern, we have developed a method with which we were able to determine pixel number and intensity as a measure of the immunolocalization of this tight junctional protein. The very local disruptions caused by 0.1 to 0.2 mM PCC in IEC-18 cells could not be quantified by this method. However, as soon as the effects on ZO-1 pattern became more apparent (0.4 and 1 mM PCC), the pixel density and pixel number were reduced. Quantification of the immunolocalization of junction-related ZO-1 seemed to correlate with transport of mannitol and PEG-4000 in IEC-18 cells after exposure to PCC. It thus appeared that the quantification of this tight junctional protein might be a valuable tool to assess changes in the tight junctional complex complementary to the determination of TER and the apparent permeability of hydrophilic compounds.

In Caco-2 no reduction in pixel density as a result of PCC exposure was detected; either the ZO-1 protein was located in the cell periphery in a cell-circumscribing pattern, or whole cells were dislodged or lysed. This shows the heterogeneity of the Caco-2 cell line: clearly, some cells are more sensitive to PCC than others. But it probably also indicates that PCC does not selectively cause alterations in the intracellular ZO-1 distribution in Caco-2 cells. Because all cells with detectable defects in ZO-1 or occludin immunolocalization also had increased propidium iodide uptake, whereas not all cells with increased propidium iodide uptake showed an altered ZO-1 or occludin pattern, we conclude that increased permeability of the apical membrane is an earlier effect of PCC exposure than disruption of tight junction structure. In general, interactions of PCC with biological membranes are widely studied and are held responsible for most of the effects, such as the blocking of inward rectifier K⁺ channels in myocytes (Sato et al., 1993) and the inhibition of insulin receptor tyrosine phosphorylation (McCallum and Epand, 1995) and protein kinase C (Nakadate and Blumberg, 1987) but are also credited for enhancing drug absorption, as shown clearly in this study and in the study of LeCluyse et al. (1991).

To assess the severity of the membrane-damaging effects and cytotoxicity of PCC, three tests were performed. A clear indicator of apical membrane damage is leakage of cytosolic enzymes into the apical medium. This type of assay has been shown to be very sensitive in studies on the effects of surfactants on intestinal epithelial cells (Anderberg et al., 1992; this study). In the present study we show that in both intestinal cell lines, leakage of the 140-kDa enzyme LDH through the apical membrane started at lower PCC concentrations than transport enhancement of mannitol and PEG-4000 across the cell layer. Findings similar to those obtained with the LDH leakage assay were obtained with the PI uptake assay, although the latter was slightly less sensitive. In pilot studies we were unable to detect an increase in PI-stained nuclei in Caco-2 cells after exposure to PCC concentrations up to 0.4 mM, much as reported by Hochman et al. (1994). However, we detected an increased cytosolic PI staining before nucleus staining and therefore choose to determine the total PI fluorescence intensity in the final studies instead of performing a nuclei count. This method showed the PI uptake to be increased at 0.4 mM PCC in Caco-2 and IEC-18 cells. In addition to LeCluyse et al. (1991), who found that perturbation of the lipid order of biological membranes correlated well with the absorption-enhancing action of PCC, we report a good correlation between the absorption-enhancing action of PCC and the membrane-damage assays for LDH leakage and PI uptake.

The NR assay showed that in both intestinal cell lines, cytotoxicity of PCC is concomitant with increased transepithelial permeability of mannitol and PEG-4000. In Caco-2, cell damage as measured by LDH leakage and NR uptake was of the same order of magnitude, whereas in IEC-18 cells, the reduction in NR uptake is much larger than would be anticipated from the LDH leakage. One possible explanation is that PCC has additional (toxic) effects on IEC-18 cells that occur during or after the immediate lysis of the membrane. For example, it has been proposed that PCC depletes ATP (Hochman and Artursson, 1994), which might finally result in cell death and thus in lower NR uptake or retention in the cells.

In an attempt to establish the balance between selective paracellular transport enhancement and transcellular transport enhancement, we exposed Caco-2 monolayers to PCC, Triton X-100 and EGTA and compared TER, the transport-enhancing factors for PEG-4000 and the LDH leakage (fig. 6). We found that increased PEG-4000 transport correlated with LDH leakage into the apical medium in cells exposed to PCC or Triton X-100, whereas no correlation was found for EGTA treatment. This similarity between the effects of PCC and Triton X-100 suggests transcellular transport enhancement. On the other hand, at lower concentrations of PCC and EGTA, the decrease in TER did not correlate with an increased transport of PEG-4000, which suggests a selectivity of PCC and EGTA for increased electrolyte permeability (indicated by TER) before macromolecule permeability. This generally indicates involvement of the paracellular pathway. However, it also shows that the initial reduction in TER is not necessarily indicative of paracellular transport enhancement of macomolecules. The parallel in the correlation between LDH leakage and TER after exposure to PCC and Triton X-100 shows that the lytic effect of PCC contributes to the decrease in TER to almost the same extent as the lytic effect of Triton X-100.

In both cell lines, the reversibility of the PCC effect on permeability and cell viability was tested in additional experiments. In both cells the TER reduction caused by 0.4 mM PCC was transient (Hochman et al., 1994; Raeissi and Borchardt, 1993; this study). However, at this PCC concentration the cells were unable to recover full viability, as shown by decreased readings in the NR uptake assay. After exposure to 1 mM PCC, neither TER nor cell viability could be recovered in IEC-18 or Caco-2.

The present study shows clear signs of cell damage at all effective (i.e., mannitol or PEG-4000 transport-enhancing) concentrations of PCC. Although we show the cytotoxicity of PCC on intestinal epithelial cells, our results should not be directly extrapolated to the in vivo situation. It has been shown previously that the Caco-2 cells are much more sensitive and responsive to PCC than are intestinal mucosa mounted in Ussing chambers (Sutton et al., 1992; Hochman et al., 1994) and intestinal tissues of rats and beagle dogs in vivo (Fix et al., 1986). And even though PCC damages cells in the mucosa, the intestinal tract may be assumed to be able to deal with loss of cells to a certain extent because of the high rate of cell turnover. Thus further studies will be needed to elucidate the balance between absorption enhancement and cytotoxicity in the intestinal mucosa in vivo.

In conclusion, PCC is effective as an absorption enhancer for hydrophilic macromolecules. However, lytic effects on the apical cell membrane and reduced cell viability accompanied the enhancement of transport at all concentrations. Although the peripheral tight junctional network and the paracellular pathway are altered by PCC, our results indicate that this is not an initial effect but rather results from the interaction of PCC with the cell membrane. This suggests that enhancement of the transcellular route or cell lysis contributes considerably to the increased permeability.