INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

EPITHELIAL CANCERS FAR OUTNUMBER any other types of cancer, and by themselves constitute one of the most prominent causes of death in the US population (67). Cancers from organs such as colon, lung, breast, bladder, prostate, and uterus typically originate from epithelial cells. One of the most fundamental functions of epithelia is their ability to act collectively as a barrier. The second fundamental characteristic of epithelia is their polarity. Two structurally and functionally unique cell membranes face the two fluid compartments, which the epithelium separates. The tight junction, or zonula occludens, which encircles each epithelial cell, has a role in maintaining epithelial polarity and in selectively sealing the paracellular pathway and, thereby, maintaining the barrier, as discussed in recent reviews (2, 13). Any effect on the structure and, thereby, the permeability of the tight junction will have profound implications on the physiological functions of the tissue that the epithelia comprise. Evidence for physiological regulation of tight junctional permeability is becoming increasingly common. Parathyroid hormone-induced paracellular permeability increase of Ca²⁺ and Mg²⁺ in the loop of Henle (70), increased paracellular water permeability in the collecting duct in response to dehydration (22), and altered paracellular permeability of the renal proximal tubule in response to glucose elevation (27) are three examples.

Evidence for altered tight junctional structure in transformed epithelia has been available for many years (46). For example, inflammatory bowel disease linked with increased incidence of colon cancer has been suggested to be associated with a genetically based alteration of tight junctional permeability, producing increased paracellular mannitol flux (29). In addition, mice treated with chemical carcinogens have been observed to have decreased transepithelial impedance across their colonic mucosae (17). Also, freeze-fracture electron microscopy of epithelia from transitional carcinoma of the urinary bladder has shown a pattern of a decreased number of tight junctional strands (60). All these findings suggest a compromised barrier function of the epithelium.

As described in recent review articles, epithelial tight junctions are under the control of a wide variety of agents and signal transduction pathways (2, 63). Of particular interest in the regulation of tight junctional permeability is the specific signal transduction component, protein kinase C (PKC) (5, 20, 21, 28, 43, 45, 56, 64). The discovery of PKC was intertwined with the field of tumor promoter carcinogenesis, because the phorbol ester class of tumor promoters was recognized early as potent activators of PKC (8, 19). Phorbol esters act through PKC in their function as tumor promoters in the two-stage model of carcinogenesis originally developed by Boutwell (11). This model postulates that tumor development would require a "first-stage," heritable change in DNA caused by a one-time application of a primary carcinogen, such as dimethylbenzanthracene or dimethylhydrazine. If this is followed by a long-term, uninterrupted exposure to tumor promoters, such as phorbol esters, the eventual yields of tumors are greatly enhanced.

To further understand the relationship among PKC activity, tight junctional permeability, and epithelial cancer, it will be important to address the following questions: 1) Which junctional proteins are being targeted for phosphorylation with subsequent alteration of tight junctional permeability? 2) Which members of the PKC family may be responsible for this alteration? Our laboratory has previously shown that the level of PKC activity and the amount of membrane-associated PKC- correlate with tight junctional permeability (50, 51, 53, 54). Specifically, we showed that PKC- expression, particularly in the membrane-associated fraction, correlates with tight junctional leakiness and that overexpression of PKC- in conjunction with phorbol ester exposure results in complete loss of barrier function (55, 59). In this report we present evidence that the -isoform of PKC is also capable of inducing tight junctional leakiness, but without the requirement of phorbol esters.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Construction of expression vectors. The tetracycline-responsive expression system described by Gossen and Bujard (24) was used to overexpress the wild-type form of PKC-. cDNA encoding the entire reading frame of mouse PKC- was cloned into the tetracycline-responsive vector pUHD10-3 creating pUHD10-3wt. This PKC- construct has been used previously to overexpress PKC- in rat basophilic leukemia cells (26, 66). In addition to the entire PKC- reading frame, the construct also contains a COOH-terminal epitope tag derived from mouse PKC-.

Cell cultures and transfections. LLC-PK₁ cells, originally derived from pig kidney cortex (30), and their derivatives were cultured in -modified minimum essential medium (JRH) supplemented with 10% fetal bovine serum (HyClone Laboratories), as previously described (55). LLC-PK₁tTA cells expressing the tetracycline-responsive transactivator tTA (59) were transfected with pUHD10-3wt using one pulse in a Bio-Rad Gene Pulsor set at 300 V and 500 µF. pSVzeo (Invitrogen) was included to confer resistance to the antibiotic Zeocin. Drug selection was at 1 mg/ml Zeocin, and subclones were screened by Western analysis. One clone expressing exogenous PKC, LLC-PK₁wt (P5), was selected for further analysis. When tetracycline was present in the medium, its concentration was 1 µg/ml.

PKC immunoblotting. Differentiated cell sheets cultured on 75-cm² tissue culture flasks (Falcon), as described above, were scraped into 2 ml of buffer A (20 mM Tris · HCl, pH 7.5, 0.25 M sucrose, 10 mM EGTA, 2 mM EDTA, 20 µg/ml leupeptin, 10 µg/ml aprotinin, and 50 µM phenylmethylsulfonyl fluoride) at 4°C, sonicated, and separated as described previously into cytosolic (C) and membrane-associated (Triton X-100-soluble, M) fractions (55). As a check on the completeness of the Triton X-100 extraction, a second membrane extraction was performed by addition of another 150 µl of buffer A at 4°C with 1% Triton X-100 to the pellet. Samples were placed on a rotator for 1 h at 4°C, then centrifuged for 1 h at 39,000 rpm. This supernatant (M2) was removed to a separate tube. The pellet (F3) was solubilized in lysis buffer as described above and represents the Triton-insoluble fraction.

Immunofluorescent detection of PKC- and PKC-. After transepithelial resistance (R_T) was read across cell sheets cultured on Falcon 3102 filter rings, cell sheets were rinsed in PBS, then fixed in 3.2% formaldehyde and permeabilized in Triton X-100, as described previously (55). After additional rinses, cell sheets were exposed to 10% normal goat serum, then incubated overnight with a rabbit polyclonal antibody to PKC- (Research and Diagnostics Antibodies) or with a mouse monoclonal antibody to PKC- (Upstate Biotechnology). After rinses in PBS, cell sheets were incubated with a 1:100 CY3 goat anti-rabbit or goat anti-mouse secondary antibody (Jackson ImmunoResearch Labs, West Grove, PA) for 1 h in the dark. After a final rinse in PBS and distilled water, cell sheets were mounted in glycerol and viewed with epifluorescence illumination using rhodamine filters.

Freeze-fracture electron microscopy. Cell monolayers growing on the Falcon 3102 filter inserts were fixed in 2% glutaraldehyde in PBS for 2 h, washed thoroughly in PBS, scraped from the surface of the filter, and progressively equilibrated with 30% glycerol in PBS for cryoprotection. After 2-3 h in the glycerol solution the samples were placed on top of a gold disk and rapidly frozen by immersion in liquid Freon cooled by liquid nitrogen. After they were frozen, the samples were freeze fractured at 110°C in a Balzers 301 apparatus, shadowed with platinum-carbon, and viewed in a Zeiss 902 electron microscope. Micrographs were printed with reverse contrast so that the platinum deposits are white.

Transmission electron microscopy and ruthenium red staining. Cell sheets were exposed on their apical surface to 0.2% ruthenium red, as described previously (55). After they were washed, cell sheets were postfixed with osmium tetroxide containing ruthenium red. All materials used for electron microscopy were purchased from Electron Microscopy Sciences (Fort Washington, PA).

R_T measurements. As described previously (55), 4 days after cells were seeded into Millicell PCF filter rings (Millipore), cell sheets were refed fresh culture medium and incubated for an additional 1 h, then initial R_T measurements were performed. Cell sheets were again refed fresh medium plus the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA), and changes in R_T were followed for an additional 2 h.

Transepithelial flux of D-[¹⁴C]mannitol and [¹⁴C]dextran (2 ×10⁶ mol wt). As described previously (55), after initial R_T measurements, cells were refed culture medium containing 0.1 mM D-[¹⁴C]mannitol (0.6 µCi/ml) or ¹⁴C-methylated dextran (1.5 µCi/ml) in the basolateral compartment. At 30-min intervals, 25-µl samples were taken from the apical fluid compartment for liquid scintillation counting. This process continued for 2.5 h. Flux rates were expressed as counts per minute per hour per square centimeter or micromoles per hour per square centimeter of cell sheet surface area. Radiolabeled mannitol and dextran are products of NEN (Boston, MA) and Sigma Chemical (St. Louis, MO), respectively.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

As described in a previous publication (59), we sought in these studies to approach the problem of clonal variation (in the generation of transfectant sublines involving single cell dilution) by utilizing an inducible transfectant expression system. In the tetracycline-responsive expression system of Gossen and Bujard (24), overexpression of wild-type PKC- is achieved only in the absence of tetracycline, whereas its presence reduces expression to the level in the parental cell line. Initially, the LLC-PK₁ cells were transfected with a plasmid expressing the tetracycline transactivating protein (tTA) along with a second plasmid conferring resistance to hygromycin. Drug-resistant clones were then screened for tetracycline-repressible expression of tTA. One such clone, LLC-PK₁tTA, was selected for further experimentation. LLC-PK₁tTA, used in a previous study on exogenous expression of PKC- (59), was then itself transfected with a plasmid, pUHD10-3wt, with Zeocin as a coselectable marker. Drug-resistant clones were selected and analyzed for expression of PKC-. One such clone, LLC-PK₁wt (P5), was selected for further study on the basis of its marked elevation of PKC- expression.

As shown in the Western immunoblot of Fig. 1, the P5 subline exhibited a very high level of PKC- expression in the absence of tetracycline. The LLC-PK₁tTA parental cell line and the P5 transfectant in the presence of tetracycline showed barely detectable levels of expression of endogenous PKC-. The transfected PKC- comigrated with the PKC- from a known PKC--positive rat brain lysate (Transduction Laboratories), each having an apparent molecular weight of ~75,000. It was necessary to maintain P5 in culture medium with tetracycline for at least 14 days to reduce transfected PKC- levels to the point of indetectability.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Western immunoblot of total protein kinase C (PKC)- in LLC-PK1tTA parental cells expressing tetracycline-responsive transactivator (tTA) and P5 cell sheets with and without tetracycline. Differentiated cell sheets cultured in 75-cm2 tissue culture flasks were rinsed, placed in an SDS lysis buffer, sonicated, and processed for PKC- SDS-PAGE Western immunoblot. Each lane received 50 µg of total protein. A commercially available positive control (+) from rat brain lysate (Transduction Laboratories) was run, and band observed comigrated with 75-kDa band shown here. Exclusion of primary antibody resulted in disappearance of all bands. These bands also disappeared when blocking peptide used to generate primary antibody was added before incubation with primary antibody. Lanes A and B, LLC-PK1tTA with and without tetracycline, respectively; lanes C and D, LLC-PK1P5 with and without tetracycline, respectively.

LLC-PK₁ epithelia typically grow in the subconfluent state as tightly associated islandlike formations. Exposure to TPA or other PKC activators (Fig. 2, A and B) causes the cells to partially disaggregate and develop a scalloped cell border (51, 53). In the absence of tetracycline, P5 manifests the morphology of a subconfluent LLC-PK₁ culture already treated with TPA (Fig. 2C). P5 cultures treated for 2 h with TPA then exhibit an exaggerated morphological response as individual cells begin to partially round up and become refractile (Fig. 2D). P5 cultures in the presence of tetracycline (Fig. 2E) appear morphologically similar to parental cultures (Fig. 2F).

View larger version (109K): [in this window] [in a new window]   Fig. 2.   Phase-contrast micrographs showing effect of protein kinase C- overexpression on morphology and response to acute phorbol ester exposure of subconfluent LLC-PK1 cell sheets. A and B: parental LLC-PK1-tTA cell line in absence and presence of 107 M 12-O-tetradecanoylphorbol-13-acetate (TPA, 2 h), respectively. C and D: P5 transfectant (without tetracycline) in absence and presence of TPA (2 h). E and F: P5 transfectant (with tetracycline) in absence and presence of TPA (2 h). Parental cells (A and B) organize in characteristic epithelial islandlike formations, which then partially disaggregate in presence of phorbol ester as individual cell membrane edges begin to ruffle and scallop. P5 transfectant in absence of tetracycline (C and D) is overexpressing PKC-, and its cells never associate in islandlike formations. In presence of TPA its cells scallop and ruffle but also eventually round up and detach from culture dish. P5 transfectant in presence of tetracycline (E and F) is virtually indistinguishable from control.

For confluent LLC-PK₁ cultures on a solid plastic substratum, the hallmark morphological characteristic is the appearance of fluid-filled domes or hemicysts randomly across the cell sheet (Fig. 3A). These structures indicate the presence of vectorial salt and water transport, as well as the presence of intact tight junctional seals. Exposure of confluent LLC-PK₁ cell sheets to TPA for 2 h results in the near-complete disappearance of domes (Fig. 3B), a phenomenon that was previously shown to be due to the increased leakiness of tight junctions (51, 53). Overexpression of PKC- in the P5 cell line produces cell sheets with generally fewer and smaller domes or, in some cases, no domes at all, even in the absence of TPA (Fig. 3C). Exposure of P5 cell sheets to TPA for 2 h causes partial rounding and detaching of cells, a response never seen in parental cells regardless of the duration of exposure to TPA (Fig. 3D). The morphology of P5 cell sheets in the presence of tetracycline is nearly identical to that of parental cell lines in control conditions or in the presence of TPA (Fig. 3, E and F).

View larger version (123K): [in this window] [in a new window]   Fig. 3.   Phase-contrast micrographs showing effect of PKC- overexpression on morphology and response to phorbol ester exposure of confluent LLC-PK1 cell sheets. A and B: parental LLC-PK1tTA cell line in absence and presence of 107 M TPA (2 h), respectively. C and D: P5 transfectant (without tetracycline) in absence and presence of TPA. E and F: P5 transfectant (with tetracycline) in absence and presence of TPA. Parental cell sheets (A) show 3-dimensional fluid-filled domes characteristic of polarized, differentiated LLC-PK1 cell sheets. Domes totally disappear after 2 h in 107 M TPA as cells assume an arabesque-like morphology but never actually round up or detach (B). P5 transfectant in absence of tetracycline (C) typically manifests fewer or, in some areas, no domes at confluence. After 2 h in TPA medium, its cells begin to round and detach (D). Behavior is similar in P5 transfectant in presence of tetracycline and in parental cell line in absence (E) and presence (F) of TPA.

Because it has been reported that PKC- has a regulatory role in the cell cycle (68) and its overexpression decreased growth rate (47), the effect of PKC- overexpression on the log-phase growth rate of LLC-PK₁ cells was examined. As shown in Fig. 4, the P5 transfectant did have a slightly slower growth rate in the absence of tetracycline (PKC- overexpressing) than in the presence of tetracycline. In addition, although possibly unrelated, the confluent density of P5 was significantly higher in the absence than in the presence of tetracycline: 2.3 × 10⁵ vs. 1.76 × 10⁵ cells/cm². This likely arises in part from the multilayering of the LLC-PK₁ cells overexpressing PKC-.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Growth curve of P5 cells in presence and absence of tetracycline (tet). Cells from a previously confluent culture were trypsinized and seeded at a density of 1 × 106 cells per 75-cm2 culture flask per 25 ml of culture medium. Each day for 3 days, cells were again trypsinized and counted individually in hemocytometers. Values are means ± SE of 3 flasks.

When cell sheets were cultured on permeable filters, the effect on tight junctional permeability of overexpression of PKC- and exposure to TPA could be evaluated by measuring R_T. The confluent LLC-PK₁ cell sheet is characterized by an R_T of ~200  · cm². This value was unaffected by the presence of tetracycline (Table 1). A 2-h exposure to TPA caused a >90% decrease in R_T. Tight junctions in the P5 subline were already leaky, as evidenced by an R_T that was only 17% of that of the parental LLC-PK₁tTA cell sheets under control conditions. Exposure to TPA caused the R_T of P5 cell sheets to decrease even further. In P5 cell sheets cultured in the presence of tetracycline, R_T values were 86% of the values in parental cell sheets, and the response of the P5 cell sheets to TPA was similar to the response of the parental cell lines.

View this table: [in this window] [in a new window]   Table 1.   Effect of PKC- overexpression on RT

To corroborate that the above differences and changes in R_T are indeed due to changes in paracellular permeability (as opposed to changes in transcellular conductance), the effects of PKC- overexpression and TPA exposure on the transepithelial flux of mannitol were also measured. Because mannitol is limited to the paracellular route in its flux across the cell sheet and cannot be metabolized by LLC-PK₁ cells (49), it is a suitable marker to measure tight junctional permeability. Previous work has shown that a 2-h exposure to a phorbol ester, such as TPA or phorbol 12,13-dibutyrate, can increase transepithelial (paracellular) mannitol flux across these cell sheets by as much as 20-fold (53). Overexpression of PKC- had a similarly dramatic effect (10.6-fold) on mannitol flux (Table 2), even in the absence of TPA. Inhibition of PKC- expression by tetracycline reduced the difference to only a 27% increase in mannitol flux. Demonstrating junctional leakiness to mannitol (182 mol wt) does not address whether the tight junctions are leaky to large macromolecules. However, the flux of 2 × 10⁶-molecular-weight ¹⁴C-methylated dextran across P5 cell sheets was substantially greater in the absence than in the presence of tetracycline (Fig. 5). This signifies that overexpression of PKC- confers paracellular leakiness to a wide range of macromolecules.

View this table: [in this window] [in a new window]   Table 2.   Effect of PKC- overexpression on transepithelial flux of 0.1 mM D-[14C]mannitol

View larger version (20K): [in this window] [in a new window]   Fig. 5.   Transepithelial flux of 14C-methylated dextran across parental and P5 cell sheets in presence and absence of tetracycline. After measurement of transepithelial electrical resistance (RT), 1.5 µCi/ml of 14C-methylated dextran (2 × 106 mol wt) was added to basolateral fluid compartment. At 1-h intervals, 25-µl samples were withdrawn from apical fluid and analyzed by liquid scintillation counting. Samples were also analyzed by gel filtration chromatography (Sephadex G50) to determine fraction of counts per minute (cpm) that was undegraded dextran. Values are means ± SE of 3 cell sheets, with cpm corrected for percentage of undegraded dextran.

An additional means of ascertaining tight junctional leakiness in the PKC--overexpressing P5 cells is to observe the penetration of an apically administered, electron-dense probe into the intercellular space. As shown in Fig. 6 and Table 3, the electron-dense dye ruthenium red is able to permeate across all tight junctional bands of the P5 cell sheet. We previously observed that phorbol ester exposure of LLC-PK₁ cell sheets results in the ability of ruthenium red to penetrate across all tight junctions (55). A P5 cell sheet maintained in the presence of tetracycline is similar to the parental cell line, in that ruthenium red is unable to penetrate across 95% of the tight junctional bands. In addition to its effect on tight junctional permeability, overexpression of PKC- in P5 cell sheets also resulted in multilayering of cells. Multilayering has likewise previously been observed for chronic exposure of LLC-PK₁ cell sheets to phorbol esters (54, 55). However, the multilayering in LLC-PK₁ cell sheets chronically treated with TPA was heterogeneous, whereas in the absence of tetracycline the multilayering occurred uniformly throughout the P5 cell sheet.

View larger version (100K): [in this window] [in a new window]   Fig. 6.   Effect of PKC- overexpression on permeability of LLC-PK1 tight junctions to ruthenium red. Electron micrographs of P5 transfectant cells in presence (A) or absence (B) of tetracycline were exposed to electron-dense dye ruthenium red, on apical surface only, for evaluation of tight junctional permeability. A: monolayered cells showing no penetration of ruthenium red through tight junctions into lateral intercellular space (arrows); B: multilayered cells with darkly stained intercellular membranes (arrow), indicating tight junctional leakiness to apical ruthenium red. Scale bars, 2 µm.

View this table: [in this window] [in a new window]   Table 3.   Quantitation of electron-microscopic evaluation of tight junctional leakiness using ruthenium red on cell sheets of P5 transfectants with or without tetracycline

Although the PKC--overexpressing cell sheets are transepithelially leaky (as assessed electrically, by transport of radiolabeled D-mannitol and dextran, and by penetration of apical ruthenium red into intercellular spaces), the cells do in fact possess tight junctions, as observed by routine transmission electron microscopy. PKC- overexpression, however, does cause a structural change in the tight junctions, as shown by freeze-fracture electron microscopy. A comparison of Figs. 7 and 8 illustrates this fact, showing that with overexpression of PKC- the normal parallel arrays of junctional strands seen in control cells are altered to a more disorganized network of strands with many discontinuities. This provides a structural correlation to the functional permeability changes described earlier (31).

View larger version (96K): [in this window] [in a new window]   Fig. 7.   Images from freeze-fracture replicas showing cross-fracture (A) and en face membrane fracture views (B and C) of tight junctions from confluent P5 transfectant cells in presence of tetracycline. Tight junctions (arrows) consist of a network of 2-5 long parallel strands and short interconnecting strands forming a continuous band separating apical membrane (am) from lateral membrane (lm) and sealing luminal compartment (lu). mv, Microvilli. Scale bar, 0.2 µm.

View larger version (207K): [in this window] [in a new window]   Fig. 8.   Freeze-fracture views of tight junctions from confluent P5 transfected cells cultured in absence of tetracycline and overexpressing PKC-. About one-third of cell contacts have tight junctions with morphology very similar to wild-type cells, where long parallel strands are connected by short strands to form a continuous belt structure (arrows in A and B) that separates apical membrane (am) from lateral membrane (lm) closing luminal compartment (lu). However, in majority of cell contacts, tight junctional strands are not organized in parallel but form a loose network with discontinuities (C and arrowheads in D). Loose tight junctional strands that are clearly not forming a barrier structure are also found dispersed along lateral plasma membrane (E and F).

After 6 h of TPA exposure, we did not observe pronounced disappearance of PKC- from the cytosolic fraction of P5 (without tetracycline), nor did PKC- exhibit significant translocation to a membrane-associated fraction (Fig. 9). Cytosolic downregulation and translocation to membrane-associated fractions occurred dramatically for PKC- (50). The amount of PKC- in the Triton-insoluble fraction (F3), however, decreased noticeably after TPA exposure.

View larger version (32K): [in this window] [in a new window]   Fig. 9.   Western immunoblot of PKC- (without tetracycline) in cytosolic, membrane-associated, and Triton-insoluble fractions of P5 cell sheets exposed to 107 M TPA. Differentiated cell sheets cultured in 75-cm2 tissue culture flasks were refed control medium + 107 M TPA and reincubated at 37°C for 6 h. Cell sheets were rinsed, placed in buffer A, sonicated, and processed for PKC- SDS-PAGE Western immunoblot. Each lane received 50 µg of total protein. Cytosolic (C) fraction is supernatant of buffer A centrifugation. Membrane-associated fraction (M1) is supernatant of a subsequent centrifugation in 1% Triton X-100. This step was then repeated once (M2) to more completely extract fraction. Pellet of second 1% Triton X-100 centrifugation (F3) was solubilized in 0.1% SDS lysis buffer. Exclusion of primary antibody resulted in disappearance of all bands.

Immunofluorescent localization of PKC- in the parental and P5 transfectant subline in the presence of tetracycline shows similar patterns of diffuse PKC- staining. In the absence of tetracycline and consequent overexpression of PKC- (Fig. 10, A and B), one can see not only a dramatically increased level of expression amidst a multilayering cell sheet but also a uniform pattern of expression across all cells of the cell sheet.

View larger version (153K): [in this window] [in a new window]   Fig. 10.   Immunofluorescent localization of - and -isoforms of PKC in PKC--overexpressing transfectant LLC-PK1 subline. A and B: PKC- expression in P5 cells in presence and absence of tetracycline, respectively. C and D: PKC- expression in presence and absence of tetracycline. Note high expression of PKC- in multilayered cells in absence of tetracycline (B) compared with low level of expression in monolayered cells in presence of tetracycline (A). Expression of PKC- is the same with (C) and without (D) tetracycline. Scale bar, 25 µm.

Overexpression of one PKC isoform has been shown to affect levels of other isoforms (69). Coupled with the fact that PKC- expression was correlated with tight junctional permeability of LLC-PK₁ cell sheets (50), the level of PKC- was also measured in the transfectant subline. Immunofluorescence detection of PKC- does not show any dramatic difference of PKC- content in the P5 subline with or without tetracycline (Fig. 10, C and D). This lack of any difference concerning PKC- expression in the P5 subline with or without tetracycline was then confirmed by Western immunoblots (data not shown). Overexpression of PKC- did not affect PKC- expression in this transfectant subline. These results suggest that the observed changes in tight junctional permeability in P5 cells are a consequence of PKC- overexpression and not an indirect effect mediated by PKC-.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The PKC- isoform was overexpressed in the LLC-PK₁ epithelial cell line, which by itself was sufficient to increase tight junctional permeability. Previously, we showed that overexpression of the PKC- isoform affected transepithelial permeability only in the presence of phorbol esters (59). These two studies suggest that 1) PKC- and PKC- can exert a regulatory role in transepithelial paracellular permeability and 2) the unusual intracellular localization of PKC- in the P5 subline may obviate the need for exogenous PKC activators, such as phorbol esters, to achieve subsequent transepithelial leakiness. Not only do the radiotracer flux and electron-dense dye studies indicate that the transepithelial leakiness is due to paracellular leakiness, but they also further define the R_T data to show that the increased paracellular leakiness is due to greater tight junctional permeability and not a change in the lateral intercellular space (4, 36, 65).

In tight junctional research, some data suggest that the paracellular pathway is not simply "open" or "closed." Instead, a wide variety of intermediate stages may exist, depending on the varying physiological needs and roles of each specific epithelial tissue. For example, the need to move glucose paracellularly across the gastrointestinal tract after a meal (44) is very likely to involve a tight junctional mode different from that during the chemotaxis of neutrophils through the paracellular pathway in response to infection (57). Both of these states would also differ from the relatively closed mode, wherein the junctions can even discriminate between the permeability afforded to Na⁺ and that afforded to Cl (23). Physiological need for a variety of tight junctional "states" suggests that many signal transduction elements exist to create and dissipate these states. Our studies provide evidence that PKC- and PKC- are two such elements that may normally regulate tight junctions. Other PKC isoforms are likely to regulate transepithelial permeability as well (20, 61).

Our studies also indicate that PKC- and PKC- act independently to regulate tight junctional permeability. Furthermore, the fact that exposure of cells to phorbol esters is required to make tight junctions leaky in PKC--overexpressing cells, whereas PKC--overexpressing cells do not share this requirement to achieve a similar physiological change, suggests that these two members of the PKC family use different mechanisms to regulate tight junctional states. This reflects back on the known differences in the molecular structure of PKC- and PKC-, differences that are likely to affect substrate specificity, intracellular localization, and the mechanisms by which the kinase activity of each of these molecules is regulated.

Like all PKC isoforms, PKC- is a single polypeptide chain with separate regulatory and catalytic domains. Although catalytic domains (ATP and substrate binding sites) are relatively conserved among all three major classes of PKC (Ca²⁺ dependent, Ca²⁺ independent, and atypical), regulatory domains vary from one class to the next as well as among individual isoforms of a given class. Thus, although phospholipid and diacylglycerol (phorbol ester) binding sites are present in the regulatory domains of PKC- and PKC-, this domain of PKC- lacks a Ca²⁺-binding site and has V1 and V2 variable regions unique from those in PKC-. It is then noteworthy that PKC- and PKC- display fourfold differences in dissociation constants for the phorbol ester phorbol 12,13-dibutyrate, with PKC- having the higher affinity (34).

PKC- and PKC- are frequently found in different cellular compartments in most cell types. PKC- tends to localize in the cytosolic compartment, whereas PKC- tends to be found in membrane-associated compartments (10, 14, 35). In the P5 transfectant, even in the absence of TPA, PKC- already is distributed in membrane-associated (Triton-soluble) and cytoskeletal-associated (Triton-insoluble) fractions, whereas PKC- is normally almost totally found in the cytosolic fraction of LLC-PK₁ cells (50). This might therefore place PKC- in a significantly different lipid environment from that in which PKC- exists in these cells. This may in turn lead to a difference in the normal occupancy of phospholipid and/or diacylglycerol binding sites of these two isozymes, which then determines their relative degree of activation.

Regardless of whether the level of expression of these isoforms in the membrane-associated compartment is increased by TPA-induced translocation or by transfection, increased tight junctional permeability results. It may be possible, therefore, that PKC- (after translocation induced by phorbol esters) and PKC- (after transfection) are targeting the same substrate in the membrane-associated fraction and that phosphorylation of this substrate leads to increased paracellular permeability. This putative substrate may be a junctional protein or a protein regulating the insertion of junctional proteins into the membrane. In any event, the mechanisms of PKC- and PKC- downregulation from the membrane-associated compartment (38, 39) may, therefore, prove just as physiologically important as their induced increases (1, 3).

Considering the phosphorylation state of PKC, there is only a single report of Tyr phosphorylation of PKC- (41), despite strong evidence of phosphorylation of PKC- on Ser/Thr residues (9, 12). However, Tyr phosphorylation of PKC- has been shown in numerous studies, where it appears to play an important stimulatory role in its kinase activity. Two PKC- bands, ~6 kDa apart, have been seen in Western immunoblots of various cell extracts (18, 40), and in one cell type a second band was attributed to phosphorylation of PKC- by the protein product of the v-src oncogene (25). The doublet PKC- bands seen in P5 extracts before but not after phorbol ester exposure (Fig. 9) may therefore assume added significance. Future work in our laboratory will determine whether these doublet PKC- bands are in fact due to a difference in PKC- phosphorylation state and whether their presence and disappearance correlate with changes in paracellular permeability.

Because of the suspected role of PKC and tumor promoters in carcinogenesis, the aforementioned compromise of epithelial barrier function in tumorigenesis, and the demonstrated regulation of epithelial barrier function by PKC- and PKC-, it may be very significant that the - and -isoforms of PKC exhibit levels of expression in human adenocarcinoma markedly different from those in adjacent normal mucosa (14, 16). Other reports have demonstrated differences in PKC activity and individual PKC isoform content between tumor tissue and normal tissue (33, 37). Previous studies from our laboratory have shown a correlation between phorbol ester induction of tight junctional leakiness and multilayering of epithelia (50). This same correlation was likewise found to exist in this study with overexpression of the -isoform of PKC. Because a strong correlation exists between PKC activity and tight junctional permeability, a key question becomes whether tight junctional seals are leaky as a result of tumor development or whether such leakiness is causally involved in tumor development. Prolonged activation of one or several PKC isoforms may be responsible for maintaining a focus of epithelial tight junctions in an open state, which will allow passage of growth factor proteins across the epithelium for an extended period (48, 52). The altered pattern of tight junctional strands shown for the PKC--overexpressing cells (Fig. 8) suggests a structural change that is significant and not instantaneously reversible.

Prolonged flux of growth factors from luminal fluid compartments into the intercellular space and interstitium may have the potential to produce abnormal growth kinetics and architecture and, thereby, be causally involved in epithelial cancer. The intrinsic polarity of epithelia places their growth factor receptors normally on their basolateral cell surface (7, 58, 62). Luminal fluid compartments, delimited by epithelial tissues, however, frequently contain a very high concentration of growth stimulatory proteins. The high levels of epidermal growth factor in urine and in the upper gastrointestinal lumen are two examples (6, 32, 71). The epithelial barrier and epithelial polarity together operate to separate ligand from receptor. If, however, a focus of epithelial cells develops a condition of chronic tight junctional leakiness and this same group of cells has been rendered vulnerable to tumor promotion by a previous exposure to an initiating carcinogen, a sufficient condition may exist for tumor development. In this case, tumor development would be due specifically to growth factor access to basolateral growth factor receptor sites (48, 52).