1. biology the Cell, (1989) 145-148
of 66 145
@ Elsevier, Paris
Water permeability in different epithelial barriers
C. CAPURRO, E. ESCOBAR, C. IBARRA, M. PORTA and M. PARIS1
Section de Biomembranas, Departamento de Fisiologia, Facultad de Medicina, Universidad de Buenos Aires, Argentina
Thewater permeability properties of a series of epithelial barriers(the toad urinary bladder [TUB], the rat caecum CRC],
the distal human colon [DHC], and the human amnion [DA] were studied in different experimental conditions. Three
parameterswere simultaneously determined: t?le water permeability coefficient in the presence of a transepithelial
hydrostaticgradient (P,,); the water permeability coefficient in the presence of an osmotic gradient (PO,,); and the
transepithelialpotential difference (dV). AR experiments were performed with the same experimental device, allowing
comparison of the permeability properties of the barriers tested. The results obtained were : (1) TUB (N= 8) :
= 0.079 f 0.008 cm/s ; P,,, = 0.0004 f 0.0002 cm/s; dV = 31 f 5 mV ; (2) TUB after ADH (N= 8) :
FlE= 0.093 f 0.012 cm/s; P =0.0065~0.0011 cm/s; dV=52&8; (3) RC (N=lO): P,,,=O.18&0.02 cm/s;
= 0.0019 f 0.0004 cm/s; iv= 3.9 f 0.1 mV ; (4) RC adapted to a high K diet (N = 10) : Pbydr 0.21 f 0.02 cm/s;
=
:= = 0.0018 f 0.0006 cm/s ; dV = 4.5 f 0.5 mV ; (5) DHC (N = 6) : P,, dr= 0.22 f 0.03 cm/s ; Posm 0.002 f 0.05 cm/s ;
=
dv%5f3 mV; (6) HA (N=lO): P,,,=O.32&0.05 cm/s; Posm=0.0154 f 0.9015 ; dV = 0. The results show a good cor-
relation between Pnvac and dV, but not between dV and P,,, or between Posmand P,,,.
toad bladder - rat caecum - distal human colon - human amnion
INTRODUCTION MATERIALS AND METHODS
Fromter Diamondin a worknow considered
and classical The tested epithelial barriers were placed as a diaphragm bet-
161,
classifiedepithelialbarriers “tight” and “leaky”.
as ween 2 lucite chambers, and the transepithelialnet water flux
The criterion employed to state, for example, that the pro- (Jw) was measuredat 1-minintervals by a techniquepreviously
ximal convoluted tubule of the kidney is a leaky barrier, described[I]. In most experiments2 fragmentsof the same tissue
whereas the collector tube is a tight one, was transepi- were tested simultan!zously. The transepithelial potential dif-
thelial conductance. This situation is reflected in the ference (dV) was measured through conventional agar bridges
and calomel electrodes.
transepithelial potential difference (dV) that, in general,
increases with barrier tightness. The net water movements
observed in epithelia were also divided, according to their Tissuesemployid
origins, into 2 main groups : (1) the so called “isotonic
transfer”, which is associated with an ionic transport in Human amniotic membranes were obtained immediatelyafter
the absence of any transepithelial gradient [3, 1l] ; and uncomplicatedterm vaginaldelivery. They were separatedfrom
(2)the osmoticallydriven net water flux, of which the the other placental tissuesand mounted, as previouslydescrib-
classical
example is the flux controlledby antidiuretic ed, for permeabilitymeasurements.Fragmentsof human distal
hormone (ADH) in some tight epithelia. There is general colons (sigmoidor rectum) were obtained from surgicallyextir-
agreement that in ADH-sensitive tissues water moves pated organsin patientswithcanceror other diseases.Immediate-
mainly transcellularly, with the regulatory barrier located ly after ablation, apparently nonaffected regionswere dissected
in the apical border of the target cell [7,8]. Much less clear free, and the mucosaland submucosal layerswereseparatedfrom
is the case of isotonic transfer, where the relative the underlying tissues and mounted for permeability
significance of transcellular and paracellular routes for measurements. Human materialswereobtainedfrom the Univer-
sity Hospital (Hospital de Clinicas)at Buenos Aires, according
water movement is still under discussion [S, 141. to the formal rules of this institution.
We have now studied a series of epithelial barriers (toad The rat caecum was removed immediately after decapitation
urinary bladder, rat caecum, distal human colon, and of the animal, washed, opened sagitally, and placed in the ex-
human amnion) in different experimental conditions. perimental chamber, Two groups of rats (Wistar) were studied :
Three parameters were simultaneously determined : water the first group was fed normally, whilst the other received a high
permeability in the presence of a transepithelial hydrostatic potassium diet (HKD) for 10 d before the experiments.
gradient (P,, &; water permeability in the presence of Toad urinary bladders were obtained from pithed animals
a transepitheiial osmotic gradient (PO,,); and dV. All ex- (Bufo arenarum) originating from the Buenos Aires area and
periments were performed by means of the same ex- mounted for.permeability measurements as previously describ-
ed. Mammalian tissues were incubated at 37°C during the ex-
perimental device, thus allowing a valid comparison of the periments and toad bladders at 20°C. Both sides of the
permeability properties of the different epithelial barriers. preparations were bathed with the same saline solution. The
The results obtained show a good correlation between 2 basic compositions were the following : Mammaliah tissues
l&, and dV, but not between dV and Posm or between (mM): 114 NaCl, 5 KCI, 1.2 CaCl,, 25 NaHCOJ, 5 glucose,
‘OS~and phydr* 2.4 K,HPO,, 1.O KH,PO, ; Toad urinary bladder MM) :

2. 146 C. Capurro o t a l .
112 NaCI, 1.0 CaCI2, 5 KCI, 2.5 NaHCO3. In all cases osmotic Hydraulic permeability
gradients were created by adding to the serosal side of the
epithelia (maternal side in the case of the amnion) different con- The Jw observed in the different epithelial tissues is
centrations of polyethyleneglycol (PEG, mol wt 8000). represented in Figure 2 as a function of the applied
Hydrostatic pressure was always applied to the mucosal bath hydrostatic pressure. In all cases linear correlations were
0uminal or fetal baths in the case of the amnion). This pressure
applied the tissue against a nylon mesh placed on its serosal obtained, and Phydr was cal~lsted in each case from the
surface. slope of the regression line. These lines intercept the or-
dinate at positive values, except for that for the human
amnion. The intercept represents the Jw value observable
Hydraulic (Phd,) and osmotic (Posm) permeability in the absence of any osmotic, chemical, or hydrostatic
coefficients "
gradient, and probably represents the Jw associated with
the ionic active transport. This is strongly supported by
The volume flow (Jw) across a membrane in the presence of a the following observation: after sodium removal (NaC!
hydrostatic (AP) or osmotic (All) gradient is described by: was isosmotically replaced by choline-chloride on both
sides), the regression fines went through the origin in all
J w = L p . dP and cases, whereas the positive ordinates disappeared (results
J w : • . p . AH not shown).
No statistically significant differences in Phydr were
where Lp and p are phenomenological coefficients and ~ is the
Staverman reflexion coefficient. If Jw is measured in tool/ observed between toad urinary bladders at rest and after
cm2/s, the hydraulic permeability coefficient (Phydr) and the stimulation with ADH, or between normal and K-adapted
osmotic permeability coefficient (Posm) can be defined: rats.
Phydr-"Lp.R. 7'/Vw
Posm: ¢'P"R. 7'/Vw
where R and T have the usual meanings and Vwis the volume
of 1 mol of water. Both coefficients are expressed in units of
centimeters per second. On this basis, Ph.,dr and Posm can be
• 3 . .
calculated from the slope of the regression line obtained when
the volume flow values are plotted against AP or AH. 1. 0.32.*0.05 (n.9)
2. 0.23*0.04 (n.8) HUMAN AMNION
A 3. O. 17 tO.03 (n.8)
E 4. 0.03 *0.04 (n.8)
o
RESULTS .S RAT COLON
E
Figure 1 shows the minute-by-minute recording of Jw HUMAN COLON
across the human amnion. The height of each stroke is
proportional to the l-rain Jw. This parameter, as can be
TOAD BLADDER
observed, was the function of the applied hydrostatic
pressure and osmotic gradient. The same types of ex-
periments were performed with the other tested
epithelia-human colon, rat caecum, and toad urinary
bladder. | L
A p (©m X20)
FIGURE 2. - The observed Jw as a function of the applied
hydrostatic gradient in different epithelial barriers. Regression
lines were obtained from the experimental values. Each point
is the mee.n of at least 6 experiments. See Table I for the ex-
perimental dispersion of the calculated slopes. 1, 2, 3, and 4
~cmX:O) 13 28
RhYdr. 21 13 13 13 13 represent the calculated ordinate intercept.
ARosm. 0 0 0 0 20 0 40
(mOsm)
I I I
I
I
I
I
I
o
'illlluLiiliIi[Iqbll;n
il
il ll ' '1,,it,:'
I I
l
' ,~0'
!
Time (rain)
|
1--00 ' ' " ' 150-
' -
Osmotic permeability
Figure 3 shows the observed 3w as a function of the ap-
plied osmotic gradient in different tissues. Again, linear
correlations were observed. The osmotic permeabilities
(Posm) were calculated from the regression lines. As ex-
FIGURE 1. -- Net water transfer across the human amnion in pected, Posm in ADH-treated toad urinary bladders was
vitro. The height of each stroke indicates the 1 rain net water significantly higher than in the control ones. No dif-
flow. The 2 horizontal rows indicate the transepitheliai ferences in Posm were observed between normal and K-
hydrostatic (Phydr)and osmotic (Posm) gradients. adapted rats.

3. Water permeability in epithelia 147
/
1. 0.26.G00 ( n . 6 ) 9"HUMAN AMNION tal device with different tissues and it can be accepted that
2- O. 18-~G03 ( n . 7) / in all cases, the leaky path was similar. Two observations
- 3. O. 17.G02 ( n . 7 ) /
/ AO
indicate that it does not bring a major contribution to the
4. O. 1 6 * G 0 4 ( . . 6 ) / .,
5BO.O3t(X02 (n.lO) ~ / observed Jw: (1) The lowest Phydr values were observed
/ 4p~TOAD BLADDER in the toad urinary bladder, the tissue that shows the
highest mechanical fragility; (2) The observed dV values
So~ in different tissues were similar to those previously
¢
reported in acceptable experimental conditions [2, 10]. It
can also be mentioned here that in the human amnion the
observed Ph..dr was sensitive to changes in the pH of the
medium [9]. This result indicates that in this tissue, which
AN COLON shows the highest Phydr values, the parameter represents
the paracellular path.
~P'~/~ ~"~'- - TOAD BLADDER
- CONTROL Figure 2 and Table I indicate that Phydr ranged from
7.9:i:0.8×10 -2 cm/s in the toad urinary bladder to
0 I/ I I J_ . - 3 2 + 5 × 10 -2 cm/s in the human amnion. These values
:f 2o 40 co - can be compared with the 15 × 10 -2 cm/s previously
PEG (mOsm)
reported in the rabbit gallbladder epithelium [13]. When
FIOURE3. -- The observed Jw as a function of the applied a correlation plot was made between P hydr and dV a
osmotic gradient in different epithelial barriers. Regression lines negative and statistically significant potential correlation
calculated as in Fig. 2. 1, 2, 3, 4, and $ represent the calculated was observed (r = 0.9 + 0.1, P < 0.05; Fig. 4). This would
ordinate intercept• indicate that in general, leaky barriers have higher
hydraulic permeabilities and conversely, tight epithelia
Spontaneouspotential difference
Table I shows the spontaneously observed potential dif-
ferences, 10 rain after mounting, in different epithelial bar- I MAN AMNION
riers. The Phydr and Po,m values are also represented. It
can be observed that K-adaptation induced an important
U
increase in dV with no change in Phydr or Posm" On the 0
W
other hand, ADH action increased dV and Posm in the ,~A.-,-.u,-(1) Control
toad urinary bladder, with no change in Phydr" 6 20 ...... "~(a) HKD
ul
,g HUMAN COLON
w TOAD BLADDER
lo " ~,,,~.Control' "ADH
DISCUSSION
z
It is generally accepted that hydrostatic pressure, in the
range employed in this study (24.4 cm H20 = lmOsM), I I I _
0 25 50 -
does not move water transcellularly. The observed Jw can Transepitheilal potential (mV)
thus take place only in a paracellular or a leaky pathway
(the last resulting from damage to cells or from any other FIOURE 4. - Correlation plot between the observed Phydr and
artifactual path generated during membrane manipulation transepithelial potential values in different epithelial barriers. The
and mounting). We employed here the same experimen- curve represents the regression potential function (r = 0.9 ± 0.1).
TABLEI. - Water permeability coefficients and potential differences in epithelial barriers.
N Phydr* (cm'sec-l'102) Posm**(cm'sec-t'102) V (mV)
Toad bladder control 8 7.94-0.8 O.04+0.O2 31.0±5.0
Toad bladder + ADH 8 9.3 4-1.2 0.65+0.11 52.0 + 8.0
Human colon 6 22.0 4-3.0 0.2O+0.O5 15.0+3.0
Rat caecum 10 18.0±2.0 0.19±0.04 3.9+0.1
Rat caecum + HKD 6 21.0±2.0 0.18±0.06 8.5+0.5
Human amnion 10 32.04-5.0 1.544-0.15 0
*Phydr: water permeability coefficient (Pf) under hydrosmotic pressure•
** Posm: water permeability coefficient (Pf) under osmotic gradient.

4. 148 C. Capurro et al.
show low Ph-dr values. It must be remarked however, that
. • 3' . • • • • •
m 2 cases m which dV was increased m a specific barrier ~ HUMANAMNION
(ADH action in toad urinary bladder or K-adaptation in
the rat caecum), there were no changes in the correspond-
ing Phydr- .
The observeo values for Posm ranged from O.08x 2
10 -2 cm/s for the toad bladder at rest to 1.5 x 10 -2 cm/s ~" 1.O
|
in the human umnion (compared to 0.93 x 10 -2 cm/s in
the rabbit gall bladder epithelium [13]• These results would
indicate that the hydrostatic pressure was between 30 and g
100 times more effective than the osmotic gradient in driv- 8 o.~ I ADH
ing a net water flux in the amnion. It must be considered,
however, that our Posm values (measured in steady-state i-.. ; ; COLON
HUMAN
, TOAD t
B'AODE.
conditions) are probably underestimated because of the I --~_ co.Jro0
"sweeping away" and "solute polarization" phenomena ,AT=ECU. . . . . ,
associated with the presence of unstirred layers [3]. Never- O 25 50 "
theless, and because all experiments were made in the same Transepithelial potential (mV)
expe"nmental conditions, we can accept, as a first approach
to the problem, that the values obtained in different tissues FIGURE6. -- Correlation plot between osmotic permeability and
(all representing a single cell layer and having similar total transepithelial potential in different epithelial barriers.
thickness) can be compared.
Figures 5 and 6 show that no clear correlation was
observed between Ph-dr and Posm or between Posm
• . .v
and Vd. This ts probably due to the fact, in addition to
the unstirred layer problem previously described, that
water may be moved osmotically either between the cells
or through the cells [4, 14]. Interestingly enough, some REFERENCES
type of correlation can be observed (dotted lines) if the
values corresponding to the ADH-stimulated toad bladders 1 Bourguet J. & Jard S. (1964) Un dispositif automatique de
are deleted. There is general agreement that in challenged mesure et d'enregistrement du flux net d'ean Atravers la peau
bladders, water is osmotically driven transcellularly. On et la vessie des amphibiens. Biochim. Biophys. Acta 88,
the other hand, it has been reported that in the human 442-444
amnion, transepithelial [14C] sucrose movements and Jw 2 Clauss W., Schafer H., Horch I. & Hornicke H. (1985)
evolve in parallel in the presence of both hydrostatic and Segmental differences in electrical properties and Na-
an osmotic gradient [9]. This parallelism was also observed transport of rabbit caecum, proximal and distal colon in
when both Ph dr and Posm changed under medium vitro. Pflugers Arch. 403, 278-282
acidification. ~hese results have been interpreted as 3 Diamond J.M. (1979) Osmotic water flow in leaky epithelia.
J. Membr. Biol. 51,195-216
indicating that the osmotically driven Jw is paracellular 4 Fischbarg J., Liebovitch L.S. & Koniarek J.P. (1985) Cen- ,
in human amnion [9]. fral role for cell osmolarity in isotonic fluid transport across
The permeability of cell membranes has been recently epithelia. Biol. Cell 55, 239-244
estimated in different epithelial barriers [12]. These values, 5 FischbargJ., Warshavsky C.R. & Lim J.J. (1976) Pathways
taken together with those observed in total tissues, will for hydraulicallyand osmoticallyinduced water fluxes across
give further information on the relative importance of epithelia. Nature 266, 71-73
paracellular and transcellular routes in transepithelial 6 Fromter E. & Diamond J.M. (1972) Route of passive ion
water transfer. permeation in epithelia. Nature 235, 9-11
7 Parisi M. & Bourguet J. (1983) The single file hypothesis
and the water channels induced by antidiuretic hormone.
Y. Membr. Biol. 71, 189-193
8 Parisi M. & Bourguet J. (1985) water channels in the animal
40 ~ cells: a widespread structure? Biol. Cell 55, 155-158
9 Porta M., Capurro C., Escobar E. & Parisi M. (1989) The
HUMAN AMNION human amnion epithelium: a model of paracellular water
"--'------4 transport. Biol. Cell (in press)
~ 30 10 Reuss L. & Finn A.L. (1974) Passive electric properties
of toad urinary bladder epithelium. J. Gen. Physiol. 64,
w;'
0
0
1-15
m
11 Sackim H. & Boulpaep E.L. (1975) Model for coupling of
salt and water transport. Proximal tubular reabsorption in
UMAN COLON Necturus kidney. J. Oen. Physiol. 66, 671-734
} 12 Van Heeswijk M.P.E. & Van Os C.H. (1986) Osmotic
permeabilities of brush border and basolateral membrane
vesicles from rat renal cortex and small intestine. J. Membr.
n~nt rol~ ~ADH I Biol. 92, 183-193
i TOAD BLADDER • 13 Van Os C.H., Wiedner G. & Wright E.M. (1979) Volume
o-I u I l I flow across gallbladder epithelium induced by small
0 (15 1.O 1.5
hydrostatic and osmotic gradients. J. Membr. Biol. 49, 1-20
RosA. (CA. sec-1.10-|) 14 Whittembury G., Paz Alliaga A., Biondi A., Carpi
Medina P., Gonzalez E. & Linares H. (1985) Pathways for
FIGURE5. - Correlation plot between hydraulic and osmotic volume flow and volume regulation in leaky epithelia.
permeabilities in different epithelial barriers. Pflugers Arch. 405 (suppl. 1), S17-$22