Permeability of intestinal mucosa from urinary reservoirs inman and ratP. NEJDFORS, J . KOÈ NYVES*, T. DAVIDSSON*, M. EKELUND², W. MAÊ NSSON* and B.R. WESTROÈ M
Departments of Animal Physiology, *Urology and ²Surgery, Lund University, Lund, Sweden
Objective To evaluate the barrier properties of intestinal
mucosa chronically exposed to urine and to evaluate
possible differences between ileal and colonic seg-
ments used in the reconstruction of the urinary tract.
Materials and methods Mucosal specimens from patients
with continent reservoirs with an abdominal stoma,
or orthotopic neobladders constructed from colonic
segments, were obtained at revisional surgery. Control
segments were obtained during right-sided hemico-
lectomy. In addition, ileal and colonic segments
from enterocystoplasties in rats were assessed. The
mucosa-to-serosa passage of marker molecules, i.e.14C-mannitol, 3H-glucose, ¯uorescein isothiocyanate-
dextran 4400 and ovalbumin, was measured using
modi®ed Ussing diffusion chambers.
Results In man, there were no permeability differences
between segments exposed to urine and control
segments for any of the marker molecules. In rats,
there was less passage of markers in ileal and colonic
transplanted segments than in intestinal segments
from sham-operated animals.
Conclusions Intestinal mucosa that has been in chronic
contact with urine maintains its barrier function; in
the rat model the permeability was even decreased. In
addition, there were no detectable differences between
ileal and colonic segments in this model.
Keywords urinary tract reconstruction, ileum, colon,
Ussing chamber, mucosa, permeability
Introduction
Intestinal segments from different anatomical regions are
increasingly used in reconstructing the urinary tract.
Several studies have shown that intestinal mucosa
chronically exposed to urine undergoes morphological
changes, e.g. mucosal atrophy of varying degree,
in¯ammatory reactions and oedema [1±3]. Previous
studies showed that intestinal segments incorporated into
the urinary tract have different transport properties for,
e.g. electrolytes, than has normal bladder epithelium [4].
In addition, the absorptive pathways in these reservoirs
change when the luminal content is replaced by urine
[5±9]. Whether these changes also re¯ect changes in the
intestinal mucosal barrier properties is not known.
The barrier function of the normal intestinal mucosa is
of utmost importance to prevent potentially noxious
substances, e.g. antigenic molecules, endotoxins and
bacteria, from permeating [10,11]. Abnormal permea-
tion of such molecules may have undesirable conse-
quences, such as septicaemia. Hence, it is important that
the mucosa of intestinal segments interposed in the
urinary tract maintains its integrity and preserves its
barrier properties.
The barrier function of the intestinal mucosa has been
assessed mainly in vivo by studying the passage of orally
given marker molecules into the blood circulation or into
urine [12±14]. In vivo techniques have been used to study
the permeability of the intestinal mucosa incorporated
in urinary tract reconstruction [15]. However, in vivo
studies in humans with urinary reconstructions are
hampered by different methodological problems, e.g.
unknown mucosal surface area, residual urine and
continuous urine production. In vitro studies may
circumvent these problems and the permeation of
different molecular probes can be evaluated in controlled
settings. Thus the aims of this study were to evaluate the
barrier properties of intestinal mucosa chronically
exposed to urine, and to evaluate any differences between
ileal and colonic intestinal segments used in the
reconstruction of the urinary tract.
Materials and methods
Adult female Sprague-Dawley rats (B&K Universal,
Sollentuna, Sweden) weighing 250±300 g were kept in
polycarbonate cages under a 12-h day-night cycle at a
temperature of 20t2uC and a relative humidity of
50t10%. The rats had free access to a pelleted rat
standard diet (B&K Universal) and water. Three groups ofAccepted for publication 19 September 2000
BJU International (2000), 86, 1058±1063
# 2000 BJU International1058
rats were used: the animals in group 1 underwent a sham
operation; in group 2 an ileal segment was isolated and
an ileocystoplasty constructed; and in group 3 an isolated
colonic segment was used in a colocystoplasty. There
were no differences in body weight among the groups
when the rats were killed after 2 months. The Ethics
Review Committee on Animal Experiments at Lund
University approved the study.
To construct the enterocystoplasty, after 12 h fasting,
the rats were anaesthetized using ketamine HCl
(7.5 mg/100 g body weight, intramuscular) and xyla-
zine HCl (1.4 mg/100 g body weight, intramuscular). At
the beginning of the operation cefuroxim (2 mg/100 g
body weight, intramuscular) was given as prophylaxis.
The peritoneal cavity was opened, and the terminal
ileum and proximal colon identi®ed and mobilized. The
sham-operated animals in group 1 underwent no
further dissection. Using a microsurgical technique, a
distal ileal or a proximal colonic segment was isolated
on its mesentery. The intestine was divided to obtain a
15-mm long ileal (group 2) or colonic (group 3)
segment to construct the enterocystoplasty. The intest-
inal continuity was re-established by an end-to-end
anastomosis using a continuous single layer of 5/0
polyglycolic acid suture. The dome of the bladder was
excised and the proximal opening of the intestinal
isolate anastomosed to the bladder using a continuous
suture. The distal opening was closed. After surgery the
rats were given subcutaneous depots of Ringer glucose
solution until they had recovered. Water was given
freely on the second day and the diet was usually
reintroduced on the third day.
The animals were killed 2 months after surgery; the
ileal (six) and colonic (nine) segments of the enterocys-
toplasties were dissected from the urinary bladder using
microsurgery. From the animals in group 1, 15-mm long
segments were obtained from the distal ileum (six) and
proximal colon (nine). All specimens were immediately
placed into oxygenated modi®ed Krebs' buffer solution
and stored there until mounted in the Ussing chambers.
The segments were cut open close to the mesenteric
border before mounting. Full-thickness intestinal seg-
ments were used, and hence the serosal-muscle layer was
not removed from the mucosa.
Nine patients (mean age 55 years, range 26±69) with
continent reservoirs and an abdominal stoma, or
orthotopic neobladders constructed from detubularized
right colonic segments, were also included in the study.
Six patients had undergone surgery for bladder cancer
and three patients, one of whom had diabetes mellitus,
had undergone surgery for neurogenic bladder disorders.
A full-thickness (< 3r3 cm) sample of the pouch wall
was taken during a re-operation, required because there
was a stricture of the uretero-colonic anastomosis or a
malfunction of the reservoir outlet. As controls, similar
specimens were taken from eight right colonic segments
during right-sided hemicolectomy in patients with
colonic cancer, and two samples were taken during
urinary tract reconstruction. None of the patients with
urinary reconstructions or in the control group had
symptoms or signs of in¯ammatory bowel disease. The
specimens were placed into oxygenated modi®ed Krebs'
buffer solution and immediately transported to the
laboratory. The serosa-muscle layer was carefully
removed and the mucosa then mounted in Ussing
diffusion chambers as described below. Approval for
the human studies was obtained from the Ethics
Committee of Lund University.
In vitro permeability experiments
Ussing diffusion chambers, with apertures of 1.78 cm2
for human specimens, 0.9 cm2 for rat ileal specimens and
0.64 cm2 for rat colonic specimens (Navicyte, San Diego,
CA, USA) were used. The chamber consists of two half-
cells joined together with the mucosal specimen mounted
as a sheet between them. After mounting the tissue the
chamber was immediately placed in a heating block; the
serosal and mucosal reservoirs were ®lled with 5 mL
modi®ed Krebs' buffer continuously bubbled with carbo-
gen (95% O2 and 5% CO2), and thus circulated by gas lift
at a temperature of 37uC [16]. The experiments started
(at t=0) within 30 min after dividing the intestinal blood
vessels, by replacing the buffer in the mucosal reservoirs
with 5 mL Krebs' buffer containing the marker molecules
(Table 1). The experiment continued for 120 min; every
20 min a 1-mL sample was taken from the serosal
reservoir and replaced with 1 mL of fresh Krebs buffer. To
obtain information about mucosal viability in the human
specimens, especially after the stripping procedure, the
transepithelial potential differences were measured every
20 min throughout the experimental period, using a pair
of Ag/AgCl electrodes embedded in 3 mol/L KCl agar.
The amount of radiolabelled mannitol and glucose in
0.5 mL samples was directly measured in a scintillation
counter (LKB, Bromma, Sweden) after mixing with 5 mL
of a liquid scintillation cocktail (Ready Safe@, Beckman,
Fullerton, CA, USA). Fluorescein isothiocyanate (FITC)-
dextrans was measured using spectrophoto¯uorometry
(Cyto¯uor 2300, Millipore, Bedford, MA, USA) at an
excitation wavelength of 480 nm and an emission
wavelength of 520 nm, with FITC-dextran dissolved in
Krebs' buffer as a standard. Ovalbumin was quanti®ed by
electro-immunoassay [17] using a speci®c antiserum to
ovalbumin (Department of Animal Physiology, Lund
University), with puri®ed ovalbumin (A-7641, Sigma) as
a standard.
The apparent permeability coef®cients (Papp) of the
INTESTINAL MUCOSA PERMEABILIT Y 1059
# 2000 BJU International 86, 1058±1063
markers across the intestinal mucosa were calculated
from:
Papp (cm/sr10x6)=dc/dtr(V/(ArC0))
where dc/dt is the change of the serosal concentration
over 60±120 min (mol/L/s), V is the volume in the
reservoir of the serosal side (mL), C0 is the initial marker
concentration in the mucosal reservoir (mol/L) and A is
the exposed intestinal area in the chamber (cm2). The
Mann±Whitney rank sum test was used for the statistical
evaluation, with differences considered signi®cant when
P<0.05.
Results
As a measure of tissue viability, the potential difference
(PD) was recorded over the human colonic mucosal
specimens during the experimental 120 min in the
Ussing chambers (Fig. 1). The PD declined with time
for all specimens, but there were no signi®cant differences
among the transplanted segments from the urine
reservoirs and the control segments.
The cumulative mucosa-to-serosa passage of the
different markers molecules increased with time in the
human colonic specimens (Fig. 2). This passage was
linear for the low molecular weight markers 14C-
mannitol and 3H-glucose, while there was a lag period
for the medium-sized markers FITC-dextran 4400 and
the macromolecular marker ovalbumin. There were no
signi®cant differences in the cumulative passage (Fig. 2)
or the calculated Papp (Table 2) for any of the different
markers between mucosa from segments from the
urinary reservoirs and mucosa from the right colon
(control). In addition, there was no correlation with the
time elapsed since the construction of the reservoir.
In the rat enterocystoplasty model, the passage of both
mannitol and glucose was less in the ileal and colonic
transplanted segments than in the corresponding seg-
ments in the sham-operated animals (Table 2). However,
the permeability to FITC-dextran 4400 and ovalbumin
was signi®cantly less only in the colonic transplanted
segments.
Discussion
Intestinal mucosa in chronic contact with urine after
urinary tract reconstruction maintains its barrier
function, as shown by the unchanged or even decreased
permeability to different sized marker molecules in
humans and rats, respectively. These results con®rm
previous in vivo ®ndings in humans of unchanged
mucosal integrity of interposed intestinal segments in
the urinary tract, as indicated by no or negligible
mucosal permeation of the medium-sized marker
molecules 51Cr-EDTA and polyethylene glycol 400
instilled into human continent reservoirs [14]. The
conclusion in the latter study was based on measure-
ments of the markers in blood, a technique that may be
questioned because of dilution and rapid renal excretion
of the markers. In the present study, the Ussing
chamber technique provided a model in which the
intestinal mucosa was isolated and evaluated, permit-
ting investigations with no interference from other
factors [5,18,19].
The choice of intestinal segments used in reconstruct-
ing the urinary tract is based mainly on the clinical
problem to be solved, but personal preferences are also
important. At the authors' department the right colon is
used most often for reconstructing the lower urinary
tract, and thus there could be no comparison between
human ileal and colonic reservoir mucosa in the present
Table 1 Characteristics and speci®cations of the marker molecules used and the initial marker concentration (C0) at the mucosal (donor) side
in the Ussing diffusion chamber experiments
Marker Manufacturer MW Concentration (C0)
14C-mannitol NEN Life Science Products, Brussels, Belgium 183 11.5 kBq/mL3H-methyl-D-glucose NEN Life Sciences 194 01.48 kBq/mL
FITC-dextran 4.400 Sigma, St Louis, MO, USA 4400 01 mg/mL
Ovalbumin Sigma 45 000 25 mg/mL
20
Time (min)
40 60 1200
Po
ten
tial
dif
fere
nce
(m
V)
10
20
30
40
0
80 100
Fig. 1. Changes in the transmural potential difference in humantransplanted right colon (red) and in controls (green).
1060 P. NEJDFORS et al .
# 2000 BJU International 86, 1058±1063
study. To partly circumvent this problem, the mucosal
permeability of ileal and colonic segments interposed in
the urinary tract was also assessed in the rat entero-
cystoplasty model.
In accord with many reports of histological changes,
our group has shown that intestinal segments incorpo-
rated in the urinary tract undergo morphological
changes, e.g. mucosal atrophy with signs of a mild
40
Time (min)
60 80 12020
Mar
ker
pas
sag
e (%
)
0.01
0.02
0.03
0.04
0.00
100
c
40
Time (min)
60 80 12020
0.001
0.002
0.003
0.004
0.000
100
d
40 60 80 12020
Mar
ker
pas
sag
e (%
)
0.1
0.2
0.3
0.4
0.0
100
a
40 60 80 12020
0.25
0.50
0.75
1.00
0.00
100
b
Fig. 2. The transmural passage of: a, 14C-mannitol; b, 3H-glucose; c, FITC-dextran 4400; and d, ovalbumin in specimens from patients with
urinary continent reservoirs (red) and in right colon control segments (green) in Ussing diffusion chambers. The permeation of the markersfrom the mucosal to the serosal side is expressed as the mean (SEM) percentage of the serosal concentration (n=9±10).
Table 2 The apparent permeability coef®cients for marker molecules in ileal and colonic enterocystoplasty segments in rats, and in mucosa
from colonic segments in contact with urine and from control right colon specimens
Segment
Mean (SD) [n] apparent permeability coef®cient (cm/sr10x6)
14C-mannitol 3H-methyl-D-glucose FITC-dextran Ovalbumin
Rat model
Ileum
Enterocystoplasty (n=6) 5.16 (1.71)* 4.93 (1.61)* 0.07 (0.04) 0.007 (0.004)
Sham (n=6) 8.37 (2.32) 8.51 (2.00) 0.19 (0.14) 0.019 (0.025)
Colon
Enterocystoplasty (n=9) 3.90 (0.93)* 4.00 (0.82)* 0.10 (0.03)* 0.0091 (0.0058)*
Sham (n=5) 5.46 (0.97) 5.14 (0.96) 0.22 (0.14) 0.0187 (0.0099)
Human
Urinary reservoir 1.47 (0.56) [9] 4.91 (2.14) [3] 0.109 (0.057) [10] 0.0145 (0.0045) [7]
Controls 1.37 (0.63) [10] 5.95 (0.88) [4] 0.067 (0.044) [10] 0.0135 (0.0050) [8]
*P<0.05 between the enterocystoplasty and the sham group.
INTESTINAL MUCOSA PERMEABILIT Y 1061
# 2000 BJU International 86, 1058±1063
in¯ammatory reaction in the lamina propria and the
muscularis mucosa [3]. In rat ileum exposed to urine the
atrophic mucosal changes increase with time, with
decreasing total height of the lamina mucosa, the
height of the villi and of the villus-crypt index [20]. It
could be argued that 2 months exposure used in the
present study was too short, but atrophic mucosal
changes were detected in a rat enterocystoplasty model
using ileal and colonic segments with a similar follow-up
[21]. Moreover, rats treated with total parenteral
nutrition show signs of gastrointestinal mucosal atrophy
by 1 week after deprivation of luminal stimuli [22],
suggesting that a 2-month urine exposure is enough to
induce changes in the enterocystoplasty model.
Small water-soluble molecules such as mannitol and
glucose permeate the intestinal mucosa by passive
diffusion (and for glucose, by active uptake using speci®c
carrier proteins integrated in the plasma membrane)
[23]. Hence, the permeation of these molecules mainly
depends on the mucosal surface area. The permeability of
mannitol and glucose in the human colonic reservoirs
was of the same magnitude as that in the colonic control
segments. This was an unexpected ®nding considering
the mucosal atrophy with consequently reduced mucosal
surface area in the former preparation. The most likely
result would have been a reduced permeability for both of
the small marker molecules. The permeability to all
marker molecules was reduced in the rat colonic
reservoir segments. This was also true in rat ileal
reservoir segment, although the permeability to FITC-
dextran 4400 and ovalbumin was not signi®cantly
decreased.
Despite the decreased surface area of the mucosa in the
human colonic reservoirs, the permeability was
unchanged. This implies that the permeability per unit
area was increased. A possible explanation could be that
the mild in¯ammatory reactions observed in biopsies
from continent cutaneous reservoirs may cause an
increased permeation despite a reduced surface area
[3,24]. An increased permeation of mannitol was
reported for intestinal segments from patients with
in¯ammatory bowel disease [18,25].
Intact tight junctions are necessary for maintained
barrier function [26]. The preserved morphology of this
structure in human and animal mucosa of ileal and
colonic reservoirs has been shown using electron
microscopy [21,27,28]. The present ®ndings of unaltered
or decreased permeability to FITC-dextran 4400 in
segments exposed to urine indicate that the functional
integrity of the junctional complexes is preserved, as
marker molecules of this size are believed to mainly
permeate by the paracellular route [29]. The small
transmural passage of ovalbumin suggests an endocyto-
tic process, but at a low level and no different from that in
the normal intestine. This is an important ®nding, as
intact barrier properties to large molecules are vital in
preventing the passage of antigenic and toxic macro-
molecules. A case report of a shock syndrome caused by
toxin from Staphylococcus aureus in urine of a continent
pouch highlights the hazards of damaged barrier
function [30].
Despite ileum and colon reacting differently to
exposure to urine, e.g. as shown by the occurrence of
the enigmatic stenosis of the ileal conduit not previously
described for a colonic conduit [31], we conclude that the
barrier integrity was no different between ileal and
colonic segments after exposure to urine.
Acknowledgements
This study was supported by grants from the Swedish
Cancer Foundation (3789-B98±03XAC); G.A.E. Nilsson
Foundation, Helsingborg, Sweden and the Dir. Albert
PaÊhlsson Foundation, MalmoÈ, Sweden.
References1 HoÈckenstroÈm T, Kock NG, NorleÂn LJ, AÊ hreÂn C, Philipson
BM. Morphological changes in ileal reservoir mucosa
after long-term exposure to urine (Kock pouch). Scand
J Gastroenterol 1986; 21: 1224±34
2 Hall MC, Koch MO, Halter SA, Dahlstedt SM. Morphological
and functional alterations of intestinal segments following
urinary diversion. J Urol 1993; 149: 664±6
3 MaÊnsson W, WilleÂn R. Mucosal morphology and histo-
chemistry of the continent cecal reservoir for urine. J Urol
1988; 139: 1199±201
4 Nurse DE, Mundy AR. Metabolic complications of cysto-
plasty. Br J Urol 1989; 63: 165±70
5 Burnett AL, Donowitz M, Marshall FF. Inhibition of
transport processes of intestinal segments following aug-
mentation enterocystoplasty in rats. J Urol 1996; 156:
1872±5
6 Filipas D, Fisch M, Abol Enein A, Fichtner J, Bockisch A,
Hohenfellner R. Chloride absorption in patients with a
continent ileocaecal reservoir (Mainz pouch I). Br J Urol
1997; 79: 588±91
7 Riedasch G, GroÈschel J, Bubeck B, Staehler G. Absorption of
vitamin B12 administrated to continent ileal reservoirs for
urinary diversion. Urol Int 1996; 56: 224±9
8 AÊ kerlund S, Jagenburg R, Kock NG, Philipson BM.
Absorption of L-phenylalanine in human ileal reservoirs
exposed to urine. Urol Res 1988; 16: 321±3
9 AÊ kerlund S, Forssell-Aronsson E, Jonsson O, Kock NG.
Decreased absorption of 22Na and 36Cl in ileal reservoirs
after exposure to urine. Urol Res 1991; 19: 249±52
10 Aranow JS, Fink MP. Determinants of intestinal barrier
failure in critical illness. Br J Anaesth 1996; 77: 71±81
11 Gardiner KR, Anderson NH, Rowlands BJ, Barbul A. Colitis
and colonic mucosal barrier dysfunction. Gut 1995; 37:
530±5
1062 P. NEJDFORS et al .
# 2000 BJU International 86, 1058±1063
12 Bjarnason I, MacPherson A, Hollander D. Intestinal
permeability: an overview. Gastroenterol 1995; 108:
1566±81
13 Meddings JB, Gibbons I. Discrimination of site-speci®c
alterations in gastrointestinal permeability in the rat.
Gastroenterol 1998; 114: 83±92
14 Menzies IS. Transmucosal passage of inert molecules in
health and disease. In Skadhauge E, Heintze K, eds. Intestinal
Absorption and Secretion. Falk Symposium 36 Lancaster,
England MTP 52, 1984
15 Davidsson T, AÊ kerlund S, White T, Olaisson G, MaÊnsson W.
Mucosal permeability of ileal and colonic reservoirs for
urine. Br J Urol 1996; 78: 64±8
16 Grass GM, Sweetana SA. In vitro measurement of gastro-
intestinal tissue permeability using a new diffusion cell.
Pharmacol Res 1998; 5: 372±6
17 Laurell CB. Electroimmunoassay. Scand J Clin Lab Invest
1972; 29: 21±37
18 Nejdfors P, Wang Q, Ekelund M et al. Increased colonic
permeability in patients with ulcerative colitis: an in vitro
study. Scand J Gastroenterol 1998; 33: 749±53
19 SoÈderholm JD, Hedman L, Artursson P et al. Integrity and
metabolism of human ileal mucosa in vitro in the Ussing
chamber. Acta Physiol Scand 1998; 162: 47±56
20 Hanke P, Hofmann M, Schneider M, BickeboÈller R, Jonas D.
Histological and scanning electron microscopic alteration in
ileal conduit depending on the time factor. Urol Int 1990;
45: 84±7
21 Davidsson T, CarleÂn B, Bak-Jensen E, WilleÂn R, MaÊnsson W.
Morphological changes in intestinal mucosa with urinary
contact-effects of urine or disuse? J Urol 1996; 156: 226±32
22 Lo CL, Walker WA. Changes in the gastrointesinal tract
during enteral or parenteral feeding. Nutrition Rev 1989;
47: 193±8
23 Ferraris RP, Diamond J. Regulation of intestinal sugar
transport. Physiol Rev 1997; 77: 257±302
24 Chen KK, Ho DM, Chen MT, Chang LS. Histopathological
changes in Kock pouch. Br J Urol 1993; 72: 422±4
25 SoÈderholm JD, Holmgren Peterson K, Olaison G et al.
Epithelial permeability to proteins in the nonin¯amed ileum
of Crohn's disease. Gastroenterol 1999; 117: 65±72
26 Madara JL. Loosening tight junctions. Lessons from the
intestine. J Clin Invest 1989; 83: 1089±94
27 CarleÂn B, WilleÂn R, MaÊnsson W. Mucosal ultrastructure of
continent cecal reservoir for urine and its ileal nipple valve
2±9 years after construction. J Urol 1990; 143: 372±6
28 Philipson BM, Kock NG, HoÈckenstroÈm T, NorleÂn LJ, AÊ hreÂn
C, Hansson HA. Ultrastructural and histochemical changes
in ileal reservoir mucosa after long-term exposure to urine.
A study in patients with continent urostomy. (Kock Pouch).
Scand J Gastroenterol 1986; 21: 1235±44
29 Pantzar N. Characteristics of intestinal permeability to
different-sized molecules in the rat. Doctoral Thesis, Lund
University, Department of Animal Physiology 1995
30 McCahill PD, Whittle DI, Jacobs SC. Toxic shock syndrome:
a comparison of continent urinary diversion. J Urol 1992;
147: 681±2
31 Magnusson B, Carlen B, Bak-Jensen E, WilleÂn R, MaÊnsson
W. Ileal conduit stenosis ± an enigma. Scand J Urol Nephrol
1986; 30: 193±7
AuthorsP. Nejdfors, PhD.
J. KoÈnyves, MD.
T. Davidsson, MD, PhD.
M. Ekelund, MD, PhD, Associate Professor of Surgery.
W. MaÊnsson, MD, PhD, Associate Professor of Urology.
B.R. WestroÈm, PhD, Associate Professor of Animal Physiology.
Correspondence: Pernilla Nejdfors, Department of Animal
Physiology, Lund University, HelgonavaÈgen 3B, SE-223 62
Lund, Sweden.
e-mail: [email protected]
INTESTINAL MUCOSA PERMEABILIT Y 1063
# 2000 BJU International 86, 1058±1063
