ATLA 30, Supplement 2, 53–59, 2002
53
Transepithelial Resistance and Inulin Permeability as
Endpoints in In Vitro Nephrotoxicity Testing
Tracey Duff,1 Simon Carter,2 Gemma Feldman,3 Gordon McEwan,1 Walter Pfaller,4 Pauline
Rhodes,2 Michael Ryan3 and Gabrielle Hawksworth1
1Department of Medicine and Therapeutics and Department of Biomedical Sciences, Polwarth and IMS
Building, University of Aberdeen, Foresterhill, Aberdeen AB25 2ZD, UK; 2GlaxoSmithKline, Park Road,
Ware, Herts. SG12 0DP, UK; 3Department of Pharmacology, University College Dublin, Belfield, Dublin 4,
Republic of Ireland; 4Institute of Physiology, University of Innsbruck, Fritz-Preglstrasse 3, 6010 Innsbruck,
Austria
Summary — Transepithelial electrical resistance (RT) and the flux of fluorescein isothiocyanate (FITC) across
Madin Darby canine kidney (MDCK) strain 1 cells and porcine epithelial kidney (LLC-PK1) monolayers were
compared between three laboratories for a range of nephrotoxins. The precision of the REMS AutoSampler
was similar to that of the Ussing chamber and the ENDOHM® technique, but superior to using chopstick
electrodes, for measurements of resistance. The nephrotoxins used were selective for the proximal tubule,
and in all cases, LLC-PK1 cells were more sensitive than MDCK cells. In most cases, change in RT was a more
sensitive indicator of damage than alterations in FITC flux. The REMS system provides high intra-plate precision for RT measurements and is a higher throughput system, which is applicable to screening for nephrotoxicity in vitro.
Key words: in vitro testing, medium-throughput screening, nephrotoxicity, nephrotoxins, transepithelial
resistance, transmembrane flux.
Introduction
The movement of solutes, ions and water across the
renal epithelium occurs by both transcellular and
paracellular routes. Transcellular transport, through
specific membrane pumps, channels and transporters, actively generates the unique electroosmotic gradients characteristic of each segment of
the renal tubule (1). The maintenance of these gradients is dependent on limiting back-diffusion
between cells through the paracellular pathway. The
major barrier in the paracellular pathway is created
by the tight junctional complex (2). Tight junctions
display varying degrees of leakiness to ion flow, from
relatively high permeability to cations in the proximal tubule, to virtual impermeability to any species
in the distal tubule and collecting duct (1). The
transepithelial electrical resistance (RT) reflects this
variability, with the proximal tubular epithelium displaying relatively low values compared with those in
the collecting duct. Madin Darby canine kidney
(MDCK) strain 1 cells, which display distal tubular
characteristics, exhibit resistance values of
5000Ω.cm2 on semipermeable membranes, whereas,
for porcine epithelial kidney (LLC-PK1) cells, this
value is approximately 100Ω.cm2. The RT reflects
resistance to ion flow across both the paracellular
and transcellular routes. Paracellular permeability is
normally measured by the transepithelial transport
of molecules such as inulin, which are restricted to
the extracellular domain. Thus, a measured fall in
RT, coupled with enhanced permeability to inulin,
would be a reliable indicator of loss of epithelial barrier function (3).
The aims of this ECVAM-funded prevalidation
study were: a) to compare RT determinations by
using the REMS automated system with the Ussing
chamber/voltage clamp technique, the ENDOHM®
technique and the Chopstick electrode/Evometer
technique; b) to carry out an interlaboratory assessment of the reproducibility of the REMS system;
and c) to compare the differential effects of known
nephrotoxins on RT and fluorescein isothiocyanate
(FITC) flux across LLC-PK1 and MDCK monolayers. The nephrotoxins were selected according to
their different modes of action and different
nephrotoxicities across a series of analogues.
Materials and Methods
Cell culture
Two renal cell lines were used in the study, MDCK
strain 1 cells (from dog distal tubule) and LLC-
Address for correspondence: Professor Gabrielle Hawksworth, Department of Medicine and Therapeutics and
Department of Biomedical Sciences, University of Aberdeen, Polwarth Building, Foresterhill, Aberdeen AB25 2ZD, UK.
T. Duff et al.
54
PK1 cells (from pig proximal tubule). The cells
were obtained from the European Collection of
Animal Cell Cultures (ECACC). MDCK cells were
cultured in Dulbecco’s modified Eagle’s medium
(DMEM) containing GLUTAMAX (Gibco, Paisley,
UK), and LLC-PK1 cells were cultured in Medium
199 (Gibco). Both culture media were supplemented with 10% heat-inactivated fetal calf serum
and 1% penicillin/streptomycin (Gibco). The cultures were maintained at 37°C in a humidified
incubator gassed with 5% CO2 and 95% air, with
complete medium replacement every 72 hours.
The cells were passaged on a weekly basis (with a
split ratio of 1:5).
For experimental purposes, the cells were seeded
in 24-well Costar HTS polycarbonate filter plates at
a seeding-density of 106 cells/ml. The filter plate
was placed into a 24-well base plate with 1ml of the
appropriate culture medium in each well, and 250µl
of cell suspension was applied to the apical chamber
of each well (growth area = 0.33cm2). The cells
were examined three days after seeding, unless otherwise indicated.
Epithelial barrier function
The RT was assessed by using a novel apparatus,
the REMS AutoSampler, manufactured and produced by World Precision Instruments, Inc. (WPI;
New Haven, CT, USA; Figure 1). The performance
of the REMS machine was tested against three welldocumented methods, namely the Ussing Chamber/
voltage clamp, ENDOHM® and chopstick electrode/
Evometer techniques.
Ussing Chambers
Cultured epithelial monolayers, grown on 12mmdiameter polycarbonate supports (Snapwell,
Corning) for 72 hours, were mounted in modified
Ussing-type chambers. The transepithelial potential difference (VT) was measured by using two
calomel electrodes, connected to each half of the
chamber by 3M KCl/3% agar bridges. Two currentpassing Ag/AgCl electrodes were similarly connected to each half-chamber. The calomel and
Ag/AgCl electrodes were coupled to a voltage/current clamp apparatus (DVC-1000, WPI), from
which current output signals were recorded on a
paper-chart recorder. Monolayers were maintained
under short-circuit conditions by passing sufficient
current (ISC) to clamp VT at 0mV. The RT across the
monolayers was determined by intermittently
changing the voltage clamp from 0mV to 5mV (1
second duration every 5 seconds), and measuring
the change in current (∆I) required to achieve this.
RT was then calculated from Ohm’s law (RT =
∆Iclamp/ Vclamp).
The ENDOHM® technique
Epithelial monolayers were grown on 12mm-diameter polyester inserts (Transwell-clear, Corning) for
72 hours. The inserts were then transferred to the
ENDOHM® chamber (WPI). The chamber and the
cap each contained a pair of concentric electrodes: a
voltage-sensing Ag/AgCl pellet in the centre, and an
annular current electrode made of medical-grade
stainless steel. For the measurement of RT, the
ENDOHM® apparatus was connected to a combined voltmeter and constant current source
(EVOM, WPI). RT was determined from the VT
measured following a defined current pulse.
Chopstick electrodes
Epithelial monolayers were grown on Costar HTS
polycarbonate Transwell 24-well plates (growth
area = 0.33cm2) for 72 hours. RT measurements
were made by using a combined voltmeter and constant current source (Millicell-ERS; Millipore,
Watford, UK) connected to the external bathing
medium and to the insert interior by paired voltagesensing and current-passing Ag/AgCl (chopstick)
electrodes. RT was determined from the VT measured following a defined current pulse.
Poly(L-lysine) experiments
The time-course of the poly(L-lysine) (PLL) effect
on the RT of MDCK cells grown on Transwell HTS
filter plates (4 days after seeding) and LLC-PK1
cells (3 days after seeding) was determined. Test
solutions covering a range of PLL (molecular mass
approximately 20kD) concentrations (1–50µg/ml)
were prepared in serum-free FITC (2µg/ml)-con-
Figure 1: The REMS autosampler
Endpoints in nephrotoxicity testing
Nephrotoxicity experiments
The effects of the nephrotoxins on MDCK strain 1
and LLC-PK1 cells were investigated at concentrations of 0.3, 1, 3 and 10mM. Plated cells were
removed from the incubator on day 3 after seeding. Two plates were tested for each nephrotoxin,
with eight wells for each nephrotoxin concentration. Two sets of test solutions were prepared for
each drug, one set with and one set without FITC
(2µg/ml). Apical and basolateral media were discarded and replaced with 1ml of appropriate
serum-free medium in each basolateral well, and
250µl of FITC-free test solution in each apical
well. Columns 2 to 4 were replaced with a 0.3, 1, 3
or 10mM solution of test compound. Column 1
was used as the vehicle control. The plates and the
FITC-containing test solutions, along with any
remaining FITC working solution, were returned
to the incubator for 24 hours. RT measurements
were then made by using the REMS AutoSampler,
and the apical media were discarded and replaced
with the appropriate FITC-containing test solution. The serum-free solution in column 6 was
replaced with PLL (0, 1, 5, 20µg/ml), prepared in
the FITC working solution. The plates were
returned to the incubator for a further 90 minutes, before a final RT measurement was made.
The filter plate was discarded, and the FITC fluorescence in each well of the base plate was determined by using a fluorescence plate reader
(Wallac Victor2 1420 multilabel counter).
Results and Discussion
= 6), or between the REMS and ENDOHM techniques (n = 4; Figure 2). RT measurements made
by using the chopsticks electrodes were significantly higher than those made with the REMS
AutoSampler (p < 0.05; n = 6; Figure 2). This is
likely to have been due to difficulties experienced
with positioning of the chopstick electrodes accurately in the filter wells. The coefficients of variation for replicates using each system were similar
although, overall, the coefficient was lowest for
the REMS system.
FITC flux
In initial experiments, fluorescein-labelled inulin
was used to assess permeability, but the labelled
inulin was too large to cross the epithelial monolayer under any conditions. It was found that
FITC itself was a suitable marker, and this was
subsequently used in all the flux experiments.
PLL produced a dose-dependent increase in FITC
flux, over the range of 5–50µg/ml. This mirrored
Figure 2: Comparison of transepithelial
electrical resistance (RT)
determinations in MDCK strain 1
epithelial monolayers, made by
using the REMS, Ussing chamber,
ENDOHM® and chopstick
electrode techniques
7000
6000
5000
mean RT (Ω.cm2)
taining medium. Filter plates were removed from
the incubator, and both apical and basolateral
media were discarded and replaced with serumfree medium. After 20 minutes, the apical bathing
medium was replaced with 250µl of the appropriate test solution. Columns 1-6 of the filter plate
were treated with 0, 1, 5, 10, 20 and 50µg/ml PLL,
respectively. The cells were incubated with
PLL/FITC for 90 minutes, during which time RT
was measured at regular intervals by using the
REMS AutoSampler. The transepithelial FITC
flux was calculated by reading the FITC concentration in each well of the baseplate, by using a
fluorescence plate reader (Wallac Victor2 1420
multilabel counter [Turku, Finland]).
55
4000
3000
2000
1000
0
Comparison of REMS Autosampler with
standard resistance measurements
There were no significant differences between
resistance measurements made by using the
REMS AutoSampler and the Ussing Chamber (n
Ussing
chamber
(n = 6)
REMS
ENDOHM
(n = 4)
REMS
(n = 4)
Chopsticks
(n = 6)
REMS
(n = 6)
apparatus
Data represented as mean ± SEM for numbers of data
shown (n).
T. Duff et al.
56
the decrease in RT over the same concentration
range (Figure 3).
10µg/ml PLL was used as a positive control in
all the 24-well plates where the monolayers were
incubated with nephrotoxins. If this concentration of PLL did not produce a significant change
in resistance or FITC flux, the results were not
included.
Nephrotoxicity
An example of FITC flux after administration of a
nephrotoxin is given in Figures 4 and 5. Cisplatin
resulted in a significant increase in FITC flux at
300µM, whereas carboplatin did not produce a sig-
nificant effect on FITC flux until a 10mM concentration was used with LLC-PK1 cells.
The mean results are summarised in Table 1. For
all compounds which showed an effect on RT or
FITC flux, LLC-PK1 cells were more sensitive than
the MDCK strain 1 cells. This is not surprising, as
the nephrotoxins used were selective for the proximal tubule. For the heavy metals, changes in RT
were a more-sensitive indicator of damage than
alterations in FITC flux. Mercuric chloride, which
causes damage by oxidative stress, was more toxic
than bismuth nitrate or cadmium chloride (4). The
platinum analogues showed toxicity in the predicted order, with cisplatin exerting an effect at a
lower concentration than transplatin, and carboplatin being the least toxic (5).
Figure 3: Time-course for action of poly(L-lysine) (PLL) on transepithelial electrical resistance
(RT) in MDCK strain 1 epithelial monolayers
4000
3500
3000
mean RT (Ω.cm2)
2500
2000
1500
1000
500
0
0
5
10
15
20
25
30
35
40
time (minutes)
Data represent mean ± SEM for four experiments.
= 0µg/ml; = 1µg/ml; = 5µg/ml; = 10µg/ml; = 20µg/ml; = 50µg/ml.
45
50
55
60
65
Endpoints in nephrotoxicity testing
Table 1:
57
RT and FITC flux measurements in MDCK strain 1 cells and LLC-PK1 cells exposed to
known nephrotoxins and positive and negative controls
MDCK strain 1
LLC-PK1
Lowest
concentration
showing
significant
effect on RT
Lowest
concentration
showing
significant effect
on FITC flux
Lowest
concentration
showing
significant
effect on RT
Lowest
concentration
showing
significant effect
on FITC flux
Mercuric chloride
↓0.01mM*
n=8
↑0.5mM**
n=8
↓0.01mM*
n=8
↑0.1mM**
n=8
Cadmium(II) chloride
↓0.3mM*
n=7
↔
↓0.1mM**
n=8
↑0.1mM**
n=8
Bismuth nitrate
↓0.01mM**
n=8
↔
↑0.1mM**
n=8
↔
Cisplatin
↓10mM**
n = 12
↑0.3mM**
n =8
↓0.3mM**
n=8
↑0.3mM**
n=8
Carboplatin
↓10mM**
n = 16
↔
↓3mM**
n=8
↑10mM**
n=8
Transplatin
↓5mM*
n = 15
↑3mM**
n=8
↑0.5mM**
n=8
↔
Gentamicin
↔
↔
↔
↔
Kanamycin
↔
↔
↑1mM**
n=8
Neomycin
↔
↔
↑0.01mM*
n=8
Cefotaxime
↔
↔
Not Done
Cephaloridine
↑1mM*
n=8
↔
↓10mM**
n=8
↑10mM**
n=8
Sodium dodecyl
sulphate
↓0.1mM**
n=4
↑1mM**
n=4
↓0.01mM**
n=8
↑0.01mM**
n=8
Diquat dibromide
↓10mM**
n=4
↑10mM**
n=4
↓0.1mM**
n=8
↑1mM**
n=8
Mannitol
↔
↔
↔
↔
Puromycin
aminonucleoside
↑10mM**
n=4
↔
↑0.1mM**
n=8
↓1mM**
n=8
Compound
↑10mM*
n=8
*p < 0.05, **p < 0.01.
FITC = fluorescein isothiocyanate; LLC-PK1 = porcine epithelial kidney; MDCK = Madin Darby canine kidney; RT =
transepithelial electrical resistance.
↓ = decrease; ↔ = no charge; ↑ = increase.
T. Duff et al.
58
As with several in vitro systems, gentamicin
exerted no effect over the concentration range used,
and neomycin was more toxic than kanamycin to
LLC-PK1 cells. This agrees with the relative order
of toxicity seen in vivo (6).
Cephaloridine exerted an effect on LLC-PK1 cells
only at a 10mM concentration. Cephaloridine toxicity is known to be mediated by intracellular accumulation, through the OAT1 transporter. This
transporter is not expressed in LLC-PK1 cells or
MDCK cells. Sodium dodecyl sulphate was used as
a non-selectively nephrotoxic detergent, but the
LLC-PK1 cells were more sensitive to this compound than were the MDCK cells. This was also the
case for diquat dibromide, which is nephrotoxic, but
which exerts its effect by a free-radical mechanism.
Puromycin aminonucleoside, the effect of which is
primarily at the glomerulus, with secondary effects
on the proximal tubule, was again more toxic to
LLC-PK1 cells than to MDCK cells (7).
Overall, the LLC-PK1 cells, with a lower starting
resistance, were more sensitive to the compounds
studied than were the MDCK cells. However, no
selectively distal tubular toxins were used in this
study. In most cases, RT was a more sensitive indicator of damage than FITC flux. For compounds
such as cisplatin, where this was not the case, it
would be interesting to determine the mechanisms
responsible for the changes in RT and FITC flux.
The changes in RT occurred at concentrations
lower than those which resulted in toxicity, as
determined by the MTT assay on a 96-well plate
(data not shown), which, together with selective
effects on RT and FITC flux, confirms that the
effects were not simply due to damage to the cell
membrane. Both the LLC-PK1 cells and the MDCK
cells were only exposed apically to the nephrotoxins, both to minimise the amount of compound used
and to set up a system which could be directly applicable to the pharmaceutical industry as a mediumthroughput screen. Exposure of cells at either the
apical or basolateral membranes, or concomitant
exposure of both cell membranes to the test compounds, may provide a system with higher sensitivity and selectivity, but would not be suitable for
routine screening. The REMS system provides high
intra-plate precision for RT measurements and is a
far higher throughput system than conventional
techniques for measuring resistance. The applicability of this system is limited to directly-acting
nephrotoxins. It is not suitable for screening when
compounds require metabolic activation or are
Figure 4: Fluorescein isothiocyanate (FITC)
flux after exposure of cells to
cisplatin
Figure 5: Fluorescein isothiocyanate (FITC)
flux after exposure of cells to
carboplatin
700
700
600
mean FITC flux (mg/cm2/hour)
mean FITC flux (mg/cm2/hour)
600
500
400
300
200
400
300
200
100
100
0
500
0
0.5%
DMSO
control
0.3
1.0
3.0
10.0
cisplatin (mM)
Data represent mean ± SEM for eight experiments.
serumfree
media
control
0.3
1.0
3.0
10.0
carboplatin (mM)
Data represent mean ± SEM for eight experiments.
Endpoints in nephrotoxicity testing
accumulated by specific transporters, unless cell
lines transfected with these transporters are used.
59
3.
Acknowledgements
This work was supported by the European
Commission (ECVAM, JRC: Contract Number
15538-1999-12 F1ED ISP GB). The authors would
like to acknowledge the excellent support received
from Dr Pilar Prieto (ECVAM, Institute for Health
& Consumer Protection, European Commission
Joint Research Centre, Ispra, Italy).
4.
5.
References
1.
2.
Brown, D. (2000). Targeting of membrane transporters in renal epithelia: when cell biology meets
physiology. American Journal of Physiology 278,
F192–F201.
Cereijido, M., Shoshani, L. & Contreras, R.G. (2000).
Molecular physiology and pathophysiology of tight
junctions. I. Biogenesis of tight junctions and epithe-
6.
7.
lial polarity. American Journal of Physiology 279,
G477–G482.
Pfaller, W. & Troopmair, E. (2000). Renal transepithelial resistance (TER) and paracellular permeability (PCP) are reliable endpoints to screen for
nephrotoxicity. In Progress in the Reduction, Refinement and Replacement of Animal Experimentation
(ed. M. Balls, A-M. van Zeller & M.E. Halder), pp
291–304. Amsterdam, The Netherlands: Elsevier
Science.
Aleo, M.D., Taub, M.L. & Kostyniak P.J. (1992).
Primary cultures of rabbit renal proximal tubule
cells. III. Comparative cytotoxicity of inorganic and
organic mercury. Toxicology and Applied Pharmacology 112, 310–317.
Baliga, R., Zhang, Z., Baliga, M., Ueda, N. & Shah,
S.V. (1998). Role of cytochrome P-450 as a source of
catalytic iron in Cisplatin-induced nephrotoxicity.
Kidney International 54, 1562–1569.
Prins, J.M., Weverling, G.J. de Blok, K., van Ketel, R.J.
& Speelman, P. (1996). Validation and nephrotoxicity
of a simplified once-daily aminoglycoside dosing schedule and guidelines for monitoring therapy. Antimicrobial Agents and Chemotherapy 40, 249–299.
Diamond, J.R., Bonventre, J.V. & Karnovski, M.J.
(1986). A role for oxygen free radicals in aminonucleoside nephrosis. Kidney International 29, 478–483.