Electrical parameters and water permeability properties of monolayers formed by T 84 cells cultured on permeable supports

T84 is an established cell line expressing an enterocyte phenotype whose permeability properties have been widely explored. Osmotic permeability (POSM), hydraulic permeability (PHYDR) and transport-associated net water fluxes (JW-transp), as well as short-circuit current (ISC), transepithelial resistance (RT), and potential difference (deltaVT) were measured in T84 monolayers with the following results: POSM 1.3 +/- 0.1 cm.s-1 x 10-3; PHYDR 0.27 +/- 0.02 cm.s-1; RT 2426 +/- 109 omega.cm2, and deltaVT 1.31 +/- 0.38 mV. The effect of 50 microM 5,6-dichloro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (DCEBIO), a "net Cl- secretory agent", on T84 cells was also studied. We confirm the reported important increase in ISC induced by DCEBIO which was associated here with a modest secretory deltaJW-transp. The present results were compared with those reported using the same experimental approach applied to established cell lines originating from intestinal and renal epithelial cells (Caco-2, LLC-PK1 and RCCD-1). No clear association between PHYDR and RT could be demonstrated and high PHYDR values were observed in an electrically tight epithelium, supporting the view that a "water leaky" barrier is not necessarily an "electrically leaky" one. Furthermore, the modest secretory deltaJW-transp was not consistent with previous results obtained with RCCD-1 cells stimulated with vasopressin (absorptive fluxes) or with T84 cells secreting water under the action of Escherichia coli heat stable enterotoxin. We conclude that, while the presence of aquaporins is necessary to dissipate an external osmotic gradient, coupling between water and ion transport cannot be explained by a simple and common underlying mechanism.


Introduction
Epithelial barriers are classified as "tight" or "leaky" according to their electrical conductance (1), which in general decreases with barrier tightness.Two different perme-ability coefficients of water net fluxes can be measured: osmotic (P OSM ) and hydrostatic (P HYDR ) (2).Moreover, diffusional water permeability can be measured from unidirectional fluxes using tritiated water.It is accepted that fluxes generated by hydrostatic pressure gradients (J W -hydr) move across the intercellular spaces, while the main route for those induced by an external osmotic gradient (J W -osm) is controversial.
Correlation between electric and water permeability properties is also especially important when a so-called electrogenic ionic transport is observed, resulting in a net salt transfer across the barrier.A net fluid transfer (J W -transp) is frequently associated with salt movement by a process that is not clearly understood (3,4).These water movements, driven by a salt transport-generated osmotic gradient, could also occur among or across epithelial cells.
The accepted water pathway for transcellular movements is the lipid bilayer itself (5), or specific water channels called aquaporins (6).Alternatively, solute-water cotransport, a controversial mechanism conceptually different from those previously mentioned, has been proposed (7)(8)(9).In the present study, water fluxes and electrical parameters were measured in T84 monolayers, a cell line derived from a human colon carcinoma.The original colon epithelium expresses aquaporins in different species, for example humans (10) and rats (11).However, the T84 cell line does not express aquaporins (12) like other established cell lines (10,13).The Na + /H + and Cl -/HCO 3 -exchangers and the Na + /HCO 3 -cotransporter have been described in these cells (14).In the present study, minute by minute recordings of the transepithelial net water fluxes (J W -hydr; J W -osm or J W -transp) were associated with the measurement of the transepithelial potential difference (∆V T ), transepithelial resistance (R T ) and short-circuit current (I SC ) under different experimental conditions.The results were compared with those reported for Caco-2, LLC-PK1 and RCCD-1 cells.Furthermore, the effects on water transfer of 5,6-dichloro-1-ethyl-1,3-dihydro-2H-benzimidazol-2-one (DCEBIO), recently reported as a "net Cl -secretory agent" in T84 cells (15), and of bumetanide were also tested and compared with the reported effects of agents that increase active absorption or secretion (12,13,16,17).

Measurement of water fluxes
In order to perform water flux measurements across the T84 monolayers, the Transwell holders with their bottom covered with the confluent cell layer were directly inserted between two Lucite hemi-chambers so as to define two independent compartments, as previously described (12).One of them (serosal) was open to the atmosphere, while the other (mucosal) was hermetically sealed.A hydrostatic pressure difference (4.5 cm of H 2 O) was continuously applied; pressure on the mucosal side is reported relative to that on the serosal side.The closed chamber was connected with a small diameter polyethylene tube to the net-water-measurement system where the net water flux (J W ) was recorded every minute, as described elsewhere (18).Briefly, the position of a liquid meniscus inside a capillary tube was photoelectrically detected.Displacements to the right or to the left were proportional to the amount of water moving across the tissue layer.The system sensitivity was 50 nl.The data were computed in units of µl min -1 cm -2 .Mucosa to serosa net movements (absorptive) were considered to be positive fluxes while serosa to mucosa net movements (secretory) were considered to be negative fluxes.The serosal bath was continuously bubbled with the appropriate CO 2 /O 2 mixture to maintain pH at 7.4 ± 0.1 (37ºC).J W -hydr and J W -osm were the J W values observed in the presence of a transepithelial hydrostatic pressure gradient or a transepithelial osmotic gradient, respectively.J Wtransp is the remaining J W when the effects of the external osmotic and hydrostatic gradients are subtracted.

Electrophysiological studies
Transepithelial voltage difference (∆V T ) and I SC were continuously recorded employing an automatic voltage clamp system (Physiological Instruments, San Diego, CA, USA) and Navycite (ME2AG4) electrodes.Before mounting the preparations in the voltage clamp system, R T was measured with a Millicell-ERS electric resistance system (Millipore, Bedford, MA, USA).During the experiments, R T was measured every 90 s from current deflections in response to a 1-mV/ second pulse.Regarding ∆V T , positive or negative values indicate "serosal side positive or negative".Electrophysiological and Jw measurements were carried out simultaneously in the same epithelial layer.
DCEBIO was a generous gift from Prof. Robert J. Bridges, University of Pittsburgh, USA.DMSO was used as solvent to test the effects of DCEBIO (50 µM on the mucosal side, and 50 and 75 µM on the serosal side) on water and ion transport.

Statistical analysis
Data are reported as means ± SEM.Statistical significance was determined using the paired or unpaired t-test, and a P value <0.05 was considered to be statistically significant.

Osmotic and hydraulic permeability of T84 cells
Figure 1 shows the J W values observed when the osmotic gradient across T84 cell monolayers was increased.Gradients were generated by adding mannitol to the serosal bath (the spontaneous water movement, measured in the presence of a 4.5-cm differ-Figure 1. Water flux as a function of osmotic gradients (J W -osm) applied from the serosal side (12.5, 50 and 100 mM mannitol).Data are reported as means ± SEM (N = 7).The inset shows a typical experiment in which water flux did not increase along the time course in the presence of an osmotic gradient.Osmotic permeability (P OSM = 1.30 ± 0.09 x 10 -3 cm s -1 ) was calculated from the slope of the corresponding regression curve (R = 0.99 ± 0.04, P < 0.001, t-test).Table 1).In this case, the spontaneous J W was a secretory one and reverted to absorptive values under the action of the applied hydrostatic pressure.

Effects of DCEBIO on electrical parameters and water movements in T84 cells
It has been reported that DCEBIO induces an important secretory Cl -response in T84 cells, associated with a significant increase in I SC as well (15).We explored the possible effects of DCEBIO on water transfer and also tested the effects of bumetanide, an inhibitor of Cl -secretion, on this cell line (17).
DMSO was employed as a solvent for both DCEBIO and bumetanide.It was added in a similar concentration to both sides of the cells and did not induce significant changes in J W , I SC , R T or ∆V T .The addition of DCEBIO (50 µM in both mucosal and serosal baths) was followed by a secretory response (∆J W -transp) together with the expected increase in both I SC and ∆V T (Table 2), effects that were reversed by 10 µM bumetanide.Figure 3 illustrates the results obtained when the concentration of DCEBIO increased on the serosal side after the initial stimulation of both sides.Finally, 10 µM bumetanide was added to the serosal bath.

Discussion
In the present study, we characterized the electrical parameters and water permeability properties of monolayers formed by T84 cells in culture.Low osmotic permeability was observed in this high resistance epithelium, together with rather high P HYDR values.An important difference in potential was associated with these properties, as previously reported (Table 1).We also report that the important increase in short circuit current induced by 50 µm DCEBIO was associated with secretory J W while both responses were inhibited by 10 µM bumetanide (Table 2).Figure 2. Hydraulic water flux (J W -hydr) across T84 monolayers as a function of different hydrostatic gradients (∆Ph, atmospheres, N = 3).Gradients were applied from the mucosal side.The hydraulic permeability coefficient (P HYDR = 0.27 ± 0.02 cm.s -1 ) was calculated from the slope of the corresponding regression curve (R = 0.99 ± 0.17, P < 0.02, t-test).The inset shows that hydrostatic flux did not increase along the time course in the presence of a hydrostatic gradient.
ence of H 2 O hydrostatic pressure, was not significantly different from zero in this experimental series).P OSM (Table 1) was calculated from the slope of the regression curve obtained from J W values and the applied osmotic gradients (R = 0.99 ± 0.04, P < 0.001).
Mean J W values as a function of the applied hydrostatic gradient are presented in Figure 2. P HYDR was calculated from the regression line (R = 0.99 ± 0.17, P < 0.02;

Comparison of passive electrical and water permeability properties in different epithelial cell lines: the role of aquaporins
The results obtained with T84 cells correlated with data previously reported for different epithelial cell lines cultured on permeable supports.All experiments were performed in our laboratory under similar experimental conditions, permitting comparison in a well-controlled manner.Figure 4 presents P HYDR values plotted against R T for Caco-2, RCCD-1, T84, LLC-PK1, and LLC-PK1 cells transfected with AQP-2 (Table 1).It can be seen that high resistance values (about 2380 Ω.cm 2 ) were associated with also rather high P HYDR values (about 1 cm s -1 ) in RCCD-1 cells.On the other hand, in LLC-PK1 cells much higher conductances were associated with lower P HYDR values.For both P HYDR and R T , the paracellular route has been previously reported to be the most important one (2,19).
When P OSM and R T were compared (Figure 5, Table 1), no clear correlation was observed.Values for LLC-PK1 cells transfected with AQP-2 clearly showed that an important increase in P OSM can be associated with no significant changes in R T .
Finally, no statistically significant correlation between P HYDR and P OSM was observed in non-transfected cells.Furthermore, it should be pointed out that after aquaporin transfection in LLC-PK1 cells, there was a huge increase in P OSM values with no changes in P HYDR (Table 1).Previous studies have demonstrated that T84, LLC-PK1 and RCCD-1 cell lines in culture lose aquaporin expression (13,17,20).Several straightforward interpretations can be proposed on the basis of these observations and of previous and present results: 1) P HYDR is a parameter that represents water transfer across the paracellular pathway.Nevertheless, rather high P HYDR values can be observed in electrically tight epithelia (T84, RCCD-1), confirming that a "water leaky" barrier is not necessarily

Water and ion coupling during absorption and secretion
The evaluation of the correlation of I SC , a parameter frequently employed to evaluate ionic transport in epithelial barriers, and the associated increases in J W -transp is more complex.Three experimental situations were evaluated in previous work from our laboratory and in the present study: the effects of ADH on RCCD-1 (13), the effects of Escherichia coli heat stable enterotoxin on T84 (17), and the effects of DCEBIO on this last cell line reported in this paper (Table 3).Figure 6 shows the reported ∆J W -transp as a function of the simultaneously measured ∆I SC .The observed ∆J W -transp values (open columns) were compared with theoretical values (filled columns), which were calculated assuming that: 1) the electrogenic transport, measured by the I SC , represents the total ionic net transport across the barrier in a pure absorptive or secretory process, and 2) this ionic movement is coupled to an isotonic fluid transfer.In the case of RCCD-1 + ADH, a modest increase in I SC was accompanied by an important absorptive J W -transp, higher than that predicted if the previously described conditions are accepted.Furthermore, when the effects of Escherichia coli heat stable enterotoxin on T84 were analyzed, the observed increase in I SC was coupled to a rather important secretory ∆J Wtransp.In fact, to accept isotonic movements, a non-electrogenic ionic transport must be postulated in both cases (17).
A completely different situation was observed in the DCEBIO-T84 experiments reported here.An important increase in I SC was associated with a modest ∆J W -transp.If we assume that the fluid transfer is isotonic and that the I SC is the result of electrogenic Cl -secretion, we can expect an induced ∆J W- transp of 0.281 µl min -1 cm -2 .The observed ∆J W was 20 times lower (Table 3, Figure 6).
These results can be explained if the observed increments in I SC correspond to the  2. an "electrically leaky" one (21); 2) in the tested native cells (not expressing aquaporins) no clear correlation was observed between P OSM and R T .A relatively high P OSM value was observed in the absence of aquaporins in a very leaky epithelium (Caco-2), and 3) after aquaporin transfection, P OSM strongly dissociated from P HYDR , probably because water pathway becomes mainly a transcellular process.
combination of Na + and Cl -DCEBIO-induced currents.If the increases in Na + and Cl -currents were in opposite directions (mucosa to serosa versus serosa to mucosa fluxes), current values would add up but the associated water movements would be canceled, at least partially.In this context, the Jw-transp calculated from the I SC (accepting isotonic movement) would be lower than that observed.On the other hand, if the increases in Na + and Cl -currents were in the same direction, the result would be that these ionic currents would subtract their values while water fluxes would be added.Now the Jw-transp calculated from the I SC (accepting isotonic movement) would be higher than that observed.If pure Na + or Cl - increases were detected, appropriate counterions should assure electroneutrality under open circuit conditions.We conclude that the associated water fluxes cannot be accurately determined from I SC measurements.

Water pathways in epithelial barriers
Net water and ionic fluxes can move across epithelial barriers through either transcellular or paracellular routes (2,22,23; for reviews, see Refs.19,24).Our results indicate that in an epithelium showing important P HYDR values (Caco-2), and in the absence of aquaporins, the paracellular route could be a significant pathway for the osmotic flux.Conversely, in an epithelium presenting lower P HYDR values (LLC-PK1), the presence of aquaporins was crucial to dissipate the external osmotic gradient (25).This last condition would be similar to that described for ADH target epithelia (12).
The route for the "transport-associated" water movement is a highly controversial topic (4), especially regarding secretory processes (26).The "standing gradient model" (27) postulates that the intercellular spaces play a central role in the so-called "isotonic transport-associated water movement".Nevertheless, it has been reported that there is no significant fluid flow through the tight junction of MDCK cells, even when fluid absorption is accelerated by the imposition of an external osmotic gradient (22).Discovery of aquaporin expression in different epithelial cells and particularly in leaky epithelia such as the proximal renal tubule has emphasized the role of transcellular channel-mediated water transport.However, a report from our laboratory demonstrated that, in the case of RCCD-1 cells, an ADH-induced absorptive J W -transp could develop in the absence of aquaporins (13).
An additional pathway and mechanism for water transport was proposed (i.e., Na + / glucose cotransporter, SGLT), which would consist of a cotransport of water and solute with a strict stoichiometry of two Na + , one glucose and ~220-250 water molecules (9,28).This mechanism was also proposed to explain coupling of water and ions in secretory processes.Although this hypothesis has generated strong controversy (29), the authors consider that the local osmotic gradient generated by the solute transfer can explain water transfer (30).Alternatively, it has been recently proposed that electro-osmosis can play a central role in water and ion coupling during active transport in a leaky epithelium (31).However, it should be pointed out that, although results obtained with epithelial cell monolayers represent useful experimental models, they cannot always be extrapolated to natural epithelia.
We conclude that J W -transp is not always the result of a simple "secondary effect" due to the generation of an osmotic gradient anywhere inside the epithelial barrier.The role of water-solute cotransport in the phenomena described here remains an open question.

Table 1 .
Osmotic permeability (P OSM ), hydraulic permeability (P HYDR ), transepithelial resistance (R T ), and potential difference (∆V T ) observed in different epithelial cell lines cultured on permeable supports.