Introduction

Top Abstract Introduction Procedures Results Discussion References

The sodium salt of the MCFA C10 is used as an absorption enhancer in drug products marketed in Japan, Denmark and Sweden, but only limited information is available regarding its mechanism of action (Anderberg et al., 1993; Lindmark et al., 1995; Tomita et al., 1995). Morphological studies have demonstrated that C10 regulates the paracellular permeability of the tight junctions in Caco-2 monolayers and in rat and human intestinal segments, indicating that it has a specific effect in the intestinal epithelium (Anderberg et al., 1993; Sawada et al., 1991; Söderholm et al., 1995). Recent studies on viable epithelial cell monolayers using fluorescent marker molecules support the hypothesis that C10 enhances drug permeability through the tight junctions (Lindmark et al., 1997b). Inhibition studies suggested that the effects of C10 are regulated through a phospholipase C-dependent pathway (Tomita et al., 1995). In general, these studies were performed after long incubations (>1 hr) with C10, and limited information is available on its immediate effects on the intestinal epithelium. This is perhaps surprising because the immediate effect of an absorption enhancer is likely to be more relevant than the long-term effect in vivo. When an absorption enhancer (together with a drug) is released from a formulation in the gastrointestinal tract, it will be rapidly diluted in the gastrointestinal fluids. As a result, the absorption enhancer will be present only in concentrations sufficiently high to increase epithelial permeability immediately after the release from the formulation. We therefore also investigated the short-term effects of the MCFAs in the present study.

A recent investigation identified the sodium salt of the MCFA C12 as another interesting regulator of epithelial permeability (Lindmark et al., 1995). In contrast to C10, C12 did not induce detectable changes in tight junction morphology, but electrophysiological measurements indicated that C12 also specifically regulates paracellular permeability in the intestinal epithelium.³ The absence of morphological changes suggests that C12 has at least a partly different mechanism of action from C10. However, no information is available on the mechanism by which C12 regulates paracellular permeability.

In this study, the short- (t = 0-12 min) and long- (t = 12-60 min) term effects of C10 and C12 on epithelial permeability to hydrophilic model drugs were studied in Caco-2 monolayers. The simplicity of this cell culture model makes it suitable for studies of the direct regulation of paracellular permeability in intestinal epithelial cells (Hochman et al., 1994; Lindmark et al., 1995; Tomita et al., 1995; Yen and Lee, 1995). Effects on the energy status of the epithelial cell monolayers as well as effects on intracellular signalling pathways were investigated.

	Experimental Procedures

Top Abstract Introduction Procedures Results Discussion References

Materials. ¹⁴C- Mannitol (molecular weight, 182; specific radioactivity, 300 mCi/g) was obtained from New England Nuclear Research Products (Boston, MA). C10 and C12 (99-100% purity) and Flu (molecular weight, 376.3) were obtained from Sigma Chemical (St. Louis, MO). Compound 48/80 (an inhibitor of PLC) (Bronner et al., 1987), W7 (a calmodulin antagonist (Hidaka et al., 1981; Itoh and Hidaka, 1984), ML7 (an inhibitor of MLCK) (Saitoh et al., 1987) and H7 (an inhibitor of PKC) (Hidaka et al., 1984) were obtained from Sigma Chemical. U73122 (an inhibitor of PLC) (Bleasdale et al., 1990; Yule and Williams, 1992), KN62 (an inhibitor of Ca²⁺/calmodulin-dependent protein kinase II) (Tansey et al., 1992) and calphostin C (an inhibitor of PKC) (Kobayashi et al., 1989) were obtained from Calbiochem (San Diego, CA). BAPTA-AM (a cell-permeant chelator of Ca⁺⁺) and diC8 (dioctanoylglycerol, a DAG analog) were obtained from Molecular Probes (Eugene, OR). All tissue culture media were obtained from GIBCO through Laboratorie Design AB (Lidingö, Sweden).

Cells. Caco-2 cells originating from a human colorectal carcinoma (Fogh et al., 1977) were obtained from American Type Culture Collection (Rockville, MD) and cultivated as described previously (Anderberg et al., 1993; Artursson, 1990; Artursson et al., 1996). In brief, 0.4 to 0.6 × 10⁶ cells/cm² were seeded onto 12-mm-diameter polycarbonate filters (Transwell cell culture inserts, mean pore diameter of 0.45 µm; Costar, Badhoevedorp, The Netherlands). The cells were used for experiments 21 to 35 days after seeding. Passages 90 through 106 were used.

Transport studies. The transport of ¹⁴C-mannitol (trace amounts) or Flu (2 mg/ml), with and without the addition of C10 or C12, across Caco-2 cell monolayers was studied as described previously (Artursson, 1990; Lindmark et al., 1995). All experiments were performed in HBSS (containing 25 mM HEPES buffer, pH 7.4) at 37°C. Ca⁺⁺/Mg⁺⁺-free HBSS was always used in the donor chamber to avoid precipitation of C10 and C12. Exclusion of Ca⁺⁺/Mg⁺⁺ from the donor solutions did not affect the integrity of the monolayers because the HBSS in the basolateral chamber always contained Ca⁺⁺/Mg⁺⁺ (Anderberg et al., 1993). The cell monolayers were equilibrated in prewarmed HBSS for 20 min before the transport experiments. At time = 0, the inserts were placed in new wells containing HBSS, and HBSS containing Flu or ¹⁴C-mannitol with or without C10 or C12 was then added to the apical side of the cell culture inserts. Samples were taken from the basolateral side by moving the cell culture inserts to new basolateral chambers containing fresh HBSS. The apparent permeability coefficient (P_app) was determined according to the following equation: P_app = dQ/dt·1/AC_o, where dQ/dt is the permeability rate (steady-state flux; mol/sec), C_o is the initial concentration in the donor chamber (mol/ml) and A is the surface area of the membrane (cm²).

Transmission electron microscopy. Monolayers were fixed with 1.5% glutaraldehyde, immersed consecutively in 1% osmium tetroxide and 1% uranyl acetate, dehydrated and embedded in Epon. Thin sections, stained with uranyl acetate and lead citrate, were examined with a Philips 420 electron microscope operated at 60 kV.

Intracellular enzyme activity (MTT). Intracellular dehydrogenase activity was determined using the MTT method (Mosman, 1983). Cells were incubated with serial dilutions of C10 or C12 in a 96-well plate for 60 min and assayed according to Anderberg et al. (1992).

Measurement of ATP. The ATP assay was based on the luciferin/luciferase reaction, using an ATP detection kit (BioThema AB, Dalarö, Sweden). This kit has a monitoring reagent formulated to provide a time-independent signal (Lundin, 1990). After incubation, ATP was extracted from the monolayers by the instantaneous addition of 1 ml of 2.5% trichloroacetic acid, and ATP in the aliquot was measured. Bioluminescence was measured with a 1250 Luminometer (LKB Wallac, Turku, Finland). An internal standard was used to determine the ATP concentration in each sample. ATP depletion, by exposing the cells to antimycin A and 2-deoxyglucose (Bacallo et al., 1994), was used as a positive control, decreasing the ATP levels to 2% of those in control monolayers.

TER. Caco-2 cells (grown on Snapwell cell culture inserts) were mounted in Ussing-type chambers equipped with a four-electrode system; one pair of Pt electrodes was used for current passage, and one pair of Ag/AgCl electrodes was used to measure the transepithelial potential difference. Measurements were made at 37°C in HBSS (Karlsson, 1995). After mounting and equilibration in HBSS, the cell monolayers were preincubated with 48/80 (2.5 µg/ml) for 20 min. Half of the apical volume of HBSS was replaced with HBSS containing C10 or C12 (final concentrations in the apical chamber were 5 and 0.375 mM, respectively) with or without 2.5 µg/ml 48/80; control cells contained HBSS only with 2.5 µg/ml 48/80. Measurements were made at 2- to 5-sec intervals the first 2 min and every minute thereafter. Average TER of the monolayers (n = 17) at time = 0 was 308.7 ± 59.2  × cm².

Immunohistochemistry and confocal microscopy. Before staining, the monolayers were exposed to buffer only (control), C10 (13 mM) or C12 (0.75 mM) for 12 or 60 min at 37°C. Further steps were performed at room temperature. Monolayers were fixed in 3% paraformaldehyde for 2 to 3 min and permeated with 0.2% Triton X-100 for 10 min. Rabbit polyclonal antibodies to ZO-1 (Zymed Lab, San Francisco, CA), diluted 1:1000 in PBS, or rabbit polyclonal antibodies against occludin (Zymed), final concentration of 10 µg/ml, were added to the apical side for 60 min. Excess solution was rinsed off, and the monolayers were incubated with PBS for 10 min. FITC-anti-rabbit antibody (Amersham, Buckinghamshire, England), diluted 1:100, was added to the apical side for 30 min. After a final incubation in PBS for 15 min, monolayers were mounted in Dako mounting medium (Dakopatts AB, Copenhagen, Denmark) on glass slides and examined under a Leica TCS 4D confocal laser scanning microscope equipped with a Leica Plan Apo 63× water-immersion lens with a numerical aperture of 1.20 and an argon/krypton laser, using the 488-nm line for excitation and a 530/30 emission filter. The images were processed with Image Space software (Molecular Dynamics, Sunnyvale, CA).

Statistics. All values are expressed as mean ± S.D. One-way analysis of variance (followed by the Fisher LSD for comparisons between two mean values) was used for the statistical analysis.

	Results

Top Abstract Introduction Procedures Results Discussion References

Short- and long-term effects on epithelial permeability. The MCFAs were studied at concentrations of 5 and 13 mM (C10) and 0.375 and 0.75 mM (C12), respectively. The lower concentrations (5 and 0.375 mM) were included to study more settled effects of the MCFAs. The two MCFAs had comparable time-dependent effects on epithelial permeability (fig. 1). Exposure to the lower C10 and C12 concentrations of 5 and 0.375 mM, respectively, revealed that both MCFAs rapidly and significantly enhanced the permeability of the monolayers within 2 to 4 min but that the epithelial integrity was restored after 20 to 40 min (fig. 1).

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Effects of C10 and C12 on the apparent permeability coefficient (Papp) of Flu in Caco-2 monolayers. Values are mean ± S.D. (n = 3 or 4). a, Control, no enhancer (), 5 mM () and 13 mM () C10. 5 mM C10 increased Papp within 3 to 4 min (P < .05, vs. control). No further increase was observed at later time points; 13 mM C10 increased Papp within 2 to 3 min of addition (P < .001). Papp was further increased at later time points (P < .01). b, Control, no enhancer (), 0.375 mM () and 0.75 mM () C12; 0.375 mM C12 increased Papp within 2 to 3 min (P < .05, vs. control), and no further increase was observed at later time points. Papp was increased by 0.75 mM C12 at 1 to 2 min (P < .01), and a further increase in Papp was observed at later time points (P < .01).

Effects on cellular metabolism. Routine transmission electron microscopy showed that some mitochondria in Caco-2 cells exposed to C10 (13 mM) for 60 min were swollen and rounded compared with cells exposed to C12 (0.75 mM) for 60 min or control cells (fig. 2), indicating changes in mitochondrial metabolism in the C10-treated cells. To further investigate this issue, the intracellular dehydrogenase activity and cellular ATP levels were investigated.

View larger version (101K): [in this window] [in a new window]   Fig. 2.   Transmission electron micrographs of mitochondria in Caco-2 cells. The bar indicates 1 µm. a, After 60-min exposure to buffer only (control), the mitochondria had a normal appearance (arrowhead). b) After 60-min exposure to 13 mM C10, some mitochondria were swollen (arrowhead). c, A 60-min exposure to 0.75 mM C12 did not change the morphology of the mitochondria compared with control (arrowhead).

View larger version (36K): [in this window] [in a new window]   Fig. 3.   Effects of C10 and C12 on the ATP content of Caco-2 cells. The figure shows the average ATP content per filter insert after exposure to buffer only (control), C10 (shaded bars, 13 mM; filled bars, 16 mM) or C12 (shaded bars, 0.75 mM; filled bars, 1.0 mM) for 60 min. Values are mean ± S.D. (n = 4). Percentage values denote decrease in ATP compared with control. **P < .01 and ***P < .001 vs. control.

Inhibition of increase in epithelial permeability. The effects of a nonspecific inhibitor of PLC, 48/80, on the short-term effects of C10 and C12 on epithelial permeability were investigated. C10 and C12 induced an instantaneous decrease in TER. Recovery in the direction of base-line values after the initial decrease in TER occurred for both 5 mM C10 and 0.375 mM C12 in the presence of 48/80. Thus, recovery of epithelial integrity was observed within 10 min (table 2). However, in contrast to a previous observation (Tomita et al., 1995), 48/80 failed to inhibit the effects of 13 mM C10 and 0.75 mM C12 (data not shown).

View larger version (21K): [in this window] [in a new window]   Fig. 4.   Schematic drawing of possible intracellular pathways activated by C10 and C12 leading to regulation of tight junction permeability. The investigated sites of inhibition are denoted with dashed arrows. See text for further explanations. CaM, calmodulin; Ca/CaM/PKII, Ca++/calmodulin-dependent protein kinase II.

View larger version (30K): [in this window] [in a new window]   Fig. 5.   Reduction in the permeability of Caco-2 monolayers to 14C-mannitol in the long-term interval (12-60 min) for 13 mM C10 (filled bars) or 0.75 mM C12 (shaded bars) after incubation with U73122, BAPTA, W7, KN62, ML7, diC8, H7 or calphostin C. The figure shows the percentage of the Papp value compared with C10 or C12 in the absence of additional modulators of signal transduction. Values are mean ± S.D. (n = 3-6). *P < .05, **P < .01, ***P < .001 compared with C10 or C12 only.

Effects on tight junction-associated proteins. Exposure to C10 (13 mM) or C12 (0.75 mM) for 12 min did not change the distribution of the intracellular (ZO-1) and the extracellular (occludin) tight junction proteins (data not shown). Exposure to C10, but not C12, for 60 min changed the appearance of both ZO-1 and occludin. In control monolayers, continuous bands of ZO-1 and occludin were observed at the cell borders (fig. 6, a and a). After exposure to C10 for 60 min, the ZO-1 and occludin staining became less even and fragmented (fig. 6, b and b). However, all neighboring cells remained in contact. C12 did not change the staining pattern of ZO-1 or occludin compared with control (fig. 6c and c).

View larger version (76K): [in this window] [in a new window]   Fig. 6.   Immunofluorescent staining of ZO-1 (a-c) and occludin (a-c) in Caco-2 monolayers. a and a, Control cells exposed to buffer only for 60 min have smooth and continuous staining of ZO-1 and occludin at the sites of cell-cell contact. b and b, Cells exposed to 13 mM C10 for 60 min have a fragmented and less-even staining of ZO-1 and occludin at the sites of cell-cell contact (arrowheads). c and c, The staining of ZO-1 and occludin in cells exposed to 0.75 mM C12 for 60 min was indistinguishable from that of controls. Scale bar, 10 µm.

	Discussion

Top Abstract Introduction Procedures Results Discussion References

This report shows that C10 and C12, two absorption enhancers specifically acting on the tight junctions, regulate paracellular permeability through PLC-dependent IP₃/DAG pathways. Surprisingly, the IP₃ and DAG pathways were found to have opposing effects on paracellular permeability. Prolonged exposure to the enhancers also reduced cellular dehydrogenase activity and ATP levels, but this reduction was most likely not sufficient to mediate the increased permeability in the present study. These findings indicate that C10 and C12 have more complex mechanisms of action than previously recognized, which has several implications for their use as absorption enhancers.

The introduction of Flu as a sensitive paracellular marker molecule in the studies of time-dependent absorption enhancement allowed the identification of a previously unobserved rapid increase in paracellular permeability. This rapid increase was followed by a second phase characterized by a continued but slower increase comparable to that observed previously after 20- to 60-min exposure to C10 and C12 (Anderberg et al., 1993; Lindmark et al., 1995). Although these results are in agreement with the rapid decrease in TER in the present and a previous study (Anderberg et al., 1993), recent investigations have indicated that changes in TER and mannitol permeability are not always correlated (Balda et al., 1996; Karlsson, 1995). It was therefore important to establish that the rapid decrease in TER was reflected in a rapid increase in marker permeability.

The increase in permeability was faster than that reported for other tight junction selective absorption enhancers such as palmitoyl carnitine and chitosans (Hochman et al., 1994; Schipper et al., 1996). A rapid onset of action is an advantageous property in an absorption enhancer because after oral administration, the intestinal epithelium is exposed to sufficiently high concentrations of the absorption enhancer together with the drug for a short time period only. Dilution in the luminal fluids and transport along the gastrointestinal tract will reduce the concentration of the enhancer and drug at the epithelial surface. This does not mean that the long-term effects are unimportant after local (e.g., rectal) administration because exposure of the mucosal tissue to the enhancer is more sustained in this case (Lindmark et al., 1997a).

Several reports suggest that a decrease in ATP levels may increase paracellular permeability (e.g., Canfield et al., 1991; Mandel et al., 1993). However, a 50% reduction in the ATP levels was required to increase the permeability of rat ileal mucosa in vivo (Madsen et al., 1995). Thus, it is unlikely that the mechanism of absorption enhancement was related to the modest changes in ATP levels observed in this study. Further support for this hypothesis is provided by the fact that C12 did not reduce the dehydrogenase or ATP levels although it increased the permeability. A possible explanation for the effect of these MCFAs on the ATP levels is that they may uncouple oxidative phosphorylation in a way similar to that observed for another MCFA, octanoic acid (C8) (Soboll et al., 1984).

Previous studies on the mechanism of C10 in Caco-2 monolayers suggested that a pathway activated by PLC could be involved (Tomita et al., 1995). The fact that the PLC inhibitor 48/80 aided recovery of TER after short-term exposure to C10 or C12 in the present study supports this hypothesis. In contrast to previous reports, no inhibitory effect was observed of 48/80 on ¹⁴C-mannitol permeability (data not shown). A possible explanation for this discrepancy is that a much larger marker molecule was used in the previous study (Tomita et al., 1995). A slight decrease in the tight junction pore size would reduce the permeability of a large molecule, whereas it may not be sufficient to reduce the permeability of a smaller marker molecule, such as mannitol. However, because changes in mannitol permeability reflect those of conventionally sized hydrophilic drugs (Anderberg et al., 1993, Artursson and Karlsson, 1991; Lennernäs et al., 1996), we retained mannitol as a marker throughout this study. The finding that a more selective inhibitor of PLC, U73122, inhibited the long-term effect of C10 on ¹⁴C-mannitol permeability gave further support for the involvement of PLC in the mechanism of C10. Because there was no corresponding inhibition of C12-induced absorption enhancement, the role of PLC in the mechanism of action for this enhancer warrants further investigation. The difference in the effect of U73122 on C10- and C12-mediated increase in permeability may be related to the higher lipid solubility of C12, resulting in a more extensive accumulation of C12 (compared with C10) in the cell membrane and a requirement of higher concentration of U73122 to obtain inhibition.

Activated PLC cleaves PIP₂ to the two intracellular mediators IP₃ and DAG. IP₃ releases Ca⁺⁺ from intracellular stores, thus increasing the intracellular Ca⁺⁺ concentration, which could result in increased paracellular permeability across the epithelium (Nathanson et al., 1992; Tai et al., 1996). To investigate whether elevated intracellular Ca⁺⁺ concentrations are involved in the mechanisms of action of C10 and C12, intracellular Ca⁺⁺ ions were buffered with BAPTA. This inhibited the effects of both C10 and C12, supporting previous suggestions that elevated intracellular Ca⁺⁺ levels is an important factor in the mechanism of C10 (Tomita et al., 1995). The lack of effect of BAPTA on the short-term effect of C10 remains unexplained, but it is possible that use of a larger marker molecule or further titration of the BAPTA concentration could provide evidence for the involvement of elevated Ca⁺⁺ concentration in this case (Tomita et al., 1995).

It is well established that contraction of the cytoskeletal structure adjacent to the tight junction and adherence junction results in increased paracellular permeability (Madara et al., 1986, 1988). Contraction of this structure is driven by ATP-dependent interaction of myosin with the actin filaments that form the core structure of the terminal web (Citi and Kendrick-Jones, 1987; Keller and Mooseker, 1982). The contraction is regulated by several mechanisms, including phosphorylation of the regulatory light chain of myosin by Ca⁺⁺/calmodulin-activated MLCK (Citi and Kendrick-Jones, 1987; Keller and Mooseker, 1982). The finding that an antagonist of calmodulin and a specific inhibitor of MLCK inhibited the long-term effects of C10 and C12 on the permeability of the epithelium to ¹⁴C-mannitol indicates that this pathway is at least partly involved in mediation of the absorption-enhancing effects of C10 and C12 in the intestinal epithelium. Interestingly, inhibition of Ca⁺⁺/calmodulin kinase II, a multifunctional Ca⁺⁺/calmodulin-dependent kinase (Schulman, 1993; Walsh, 1994), resulted in only inhibition of the effect of C10. Because previous studies indicate that C10 (but not C12) induces morphological changes in the perijunctional F-actin ring and in the tight junctions (Anderberg et al., 1993; Lindmark et al., 1995), it may be speculated that Ca⁺⁺/calmodulin kinase II is involved in these changes. Because none of these inhibitors affected the short-term increase in the permeability to ¹⁴C-mannitol caused by C10 and C12, we tentatively conclude that the short-term effect of the MCFAs is mediated via a partly different mechanism. This is supported by results indicating that separations of the tight junctions induced by C10 were first observed after long-term exposure to C10 (Anderberg et al., 1993).

The second product of PLC-induced cleavage of PIP₂ is DAG. Along with Ca⁺⁺, DAG activates PKC, which in turn regulates tight junction assembly as well as tight junction permeability (Balda et al., 1991, 1993; Stenson et al., 1993; Stuart and Nigam, 1995). The effects of PKC activation on tight junction permeability are complex and vary with different experimental settings and cell types (Ellis et al., 1992). Activation of PKC has been shown to up- or down-regulate tight junction permeability depending on the experimental environment (Balda et al., 1993; Nathanson et al., 1992; Stenson et al., 1993; Stuart and Nigam, 1995)_. The results of the present study suggest that PKC activation down-regulated the C10/C12-mediated effects on the paracellular permeability of Caco-2 cells. First, the application of PKC inhibitors at nontoxic concentrations increased rather than decreased the C10- or C12-mediated long-term effect on paracellular permeability. Second, the DAG agonist diC8 inhibited the effects of the two MCFAs; this is in agreement with studies on tight junction assembly in which diC8 was found to increase the recruitment of the tight junction protein ZO-1 to the tight junctions (Balda et al., 1993). We conclude that PKC-mediated phosphorylation counteracts both the short- and long-term effects of C10- and C12-mediated absorption enhancement in Caco-2 cells. Whether these effects are mediated via PKC isoenzymes recently shown to be colocalized with ZO-1 at the tight junctions remains to be seen (Dodane and Kachar, 1996; Stuart and Nigam, 1995).

The staining patterns of the tight junction proteins ZO-1 and occludin after exposure to C10 suggest that these proteins are involved in the morphological changes in tight junction structure after C10 exposure in cell culture as well as in intestinal tissue (Anderberg et al., 1993; Söderholm et al., 1995). Although no clear effects on the tight junctions were observed after C12 exposure, it cannot at present be excluded that changes below the resolution of the confocal microscope occurred after exposure to C12.

The results of this study have several implications for the application of C10 and C12 as absorption enhancers. First, the time- and concentration-dependent effects on cellular ATP levels indicate that prolonged interaction with an epithelial barrier (as obtained after local administration, e.g., to the rectal mucosa; Lindmark et al., 1997a) may result in nonspecific cell damage and cell death caused by energy depletion. Development of delivery systems that control the release of the absorption enhancer may solve this problem. Second, the multiple and partly different intracellular effects of C10 and C12 suggest that the efficacy and specificity of these absorption enhancers may vary with the expression of, for example, protein kinases and cytoskeletal components in different epithelia (Ellis et al., 1992). Third, because many drug molecules regulate intracellular events, such drugs could decrease or increase the enhanced paracellular permeability. This could provide an alternative explanation for the observed drug-dependent variability in the efficacy of C10 in vivo (Aungst et al., 1996).