Renal ischemia-reperfusion injury causes intercalated cell-specific disruption of occludin in the collecting duct.

Abstract

Renal ischemic events open tight junctions and disrupt epithelial polarity. The purpose of this study was to examine the effects of ischemia-reperfusion (IR) injury on expression and distribution of the tight junction proteins, occludin and ZO-1, in the rat kidney. IR injury was induced by clamping both renal pedicles for 30 min and animals were killed at 6 h after the reperfusion. IR injury decreased blood bicarbonate level, but did not persistently alter pH, Na(+), K(+), or Cl(-). In control kidneys, occludin immunoreactivity was intense in the tight junctions in the thick ascending limb, distal convoluted tubule, and collecting duct, moderate in the thin limbs of the loop of Henle, and was not detected in the proximal tubule, glomerulus, and blood vessels. ZO-1 was expressed in the same sites in which occludin was expressed, and additionally was also expressed in the proximal tubule, glomerulus, and vascular endothelial cells. IR kidneys exhibited damaged renal tubular epithelial cells in both proximal tubule and collecting duct segments in the outer medulla. In the collecting duct, the response of intercalated cells and principal cells differed. Following IR injury, intercalated cells, but not principal cells, lost their normal epithelial polarity and were frequently extruded into the tubule lumen. Occludin, instead of being localized to tight junctions, was localized diffusely in the cytoplasm in intercalated cells of IR kidneys. Principal cells, in contrast, were not detectably affected and neither occludin nor ZO-1 expression were altered in response to IR injury. The normal localization of ZO-1 expression to tight junction sites in both the proximal tubule and collecting duct was altered in response to IR, and, instead, ZO-1 expression was present diffusely in the cytoplasm. IR injury did not alter detectably either occludin or ZO-1 localization to the tight junction of the thick ascending limb cells. The abundance of total occludin protein by immunoblot analysis was not changed with IR injury. These results demonstrate that renal IR injury causes tight junction disruptions in both the proximal tubule and the collecting duct, and that altered distribution of the tight junction protein, occludin, may play a critical role in the collecting duct dysfunction which IR induces.

Full text

Page 1

Histochem Cell Biol (2011) 136:637–647
DOI 10.1007/s00418-011-0881-4
123
ORIGINAL PAPER
Renal ischemia–reperfusion injury causes intercalated cell-speciWc
disruption of occludin in the collecting duct
Su-Youn Lee · Jung-A Shin · H. Moo Kwon ·
I. David Weiner · Ki-Hwan Han
Accepted: 23 October 2011 / Published online: 3 November 2011
© The Author(s) 2011. This article is published with open access at Springerlink.com
Abstract
disrupt epithelial polarity. The purpose of this study was to
examine the eVects of ischemia–reperfusion (IR) injury on
expression and distribution of the tight junction proteins,
occludin and ZO-1, in the rat kidney. IR injury was induced
by clamping both renal pedicles for 30 min and animals
were killed at 6 h after the reperfusion. IR injury decreased
blood bicarbonate level, but did not persistently alter pH,
Na+, K+, or Cl¡. In control kidneys, occludin immunoreac-
tivity was intense in the tight junctions in the thick ascend-
ing limb, distal convoluted tubule, and collecting duct,
moderate in the thin limbs of the loop of Henle, and was not
detected in the proximal tubule, glomerulus, and blood ves-
sels. ZO-1 was expressed in the same sites in which occlu-
din was expressed, and additionally was also expressed in
Renal ischemic events open tight junctions and the proximal tubule, glomerulus, and vascular endothelial
cells. IR kidneys exhibited damaged renal tubular epithelial
cells in both proximal tubule and collecting duct segments
in the outer medulla. In the collecting duct, the response of
intercalated cells and principal cells diVered. Following IR
injury, intercalated cells, but not principal cells, lost their
normal epithelial polarity and were frequently extruded into
the tubule lumen. Occludin, instead of being localized to
tight junctions, was localized diVusely in the cytoplasm in
intercalated cells of IR kidneys. Principal cells, in contrast,
were not detectably aVected and neither occludin nor ZO-1
expression were altered in response to IR injury. The nor-
mal localization of ZO-1 expression to tight junction sites
in both the proximal tubule and collecting duct was altered
in response to IR, and, instead, ZO-1 expression was pres-
ent diVusely in the cytoplasm. IR injury did not alter detec-
tably either occludin or ZO-1 localization to the tight
junction of the thick ascending limb cells. The abundance
of total occludin protein by immunoblot analysis was not
changed with IR injury. These results demonstrate that
renal IR injury causes tight junction disruptions in both the
proximal tubule and the collecting duct, and that altered
distribution of the tight junction protein, occludin, may play
a critical role in the collecting duct dysfunction which IR
induces.
Keywords
Kidney · Occludin · Tight junction
Collecting duct · Ischemia/reperfusion injury ·
Introduction
Renal tubular cells have unique Xuid and electrolyte trans-
port capabilities across the distinct apical and basolateral
membrane domains. Tight junctions maintain the functional
S.-Y. Lee · J.-A. Shin · K.-H. Han (&)
Department of Anatomy, Ewha Womans University
School of Medicine, 911-1 Mok-6-dong, Yangcheon-ku,
Seoul 158-710, Korea
e-mail: khhan@ewha.ac.kr
H. M. Kwon
Division of Nephrology, University of Maryland School
of Medicine, Baltimore, MD, USA
I. D. Weiner
Division of Nephrology, University of Florida College
of Medicine, Gainesville, FL, USA
I. D. Weiner
Nephrology Section, NF/SGVHS, Gainesville, FL, USA

Page 2

638 Histochem Cell Biol (2011) 136:637–647
123
integrity and polarity of tubular surface membrane, and reg-
ulate paracellular and transcellular transports (Fish and
Molitoris 1994; Lee et al. 2006; Weinstein 2003). The tight
junction complex consists of multiple adhesion proteins
linking it to the cytoskeleton of the individual cells.
Occludin, ZO-1, and the claudin family are major com-
ponents of the growing number of tight junction proteins
(Furuse et al. 1993; Mitic et al. 2000; Schneeberger and
Lynch 2004; Stevenson et al. 1986). Occludin carries out an
adhesive function when interacting with ZO-1 in kidney
cells (Furuse et al. 1994; Van Itallie and Anderson 1997).
Although the exact roles of occludin and ZO-1 in the kid-
ney are not completely understood, their distribution varies
markedly along the tubule segments. The amount of occlu-
din and ZO-1 expression is signiWcantly greater in more
distal tubular segments than in proximal tubules, which
correlates with the greater transepithelial electrical resis-
tance in those segments (Gonzalez-Mariscal et al. 2000). In
cultured cells, overexpression of occludin dramatically
increases transepithelial resistance (Balda et al. 1996;
McCarthy et al. 1996). Thus, occludin and ZO-1 may play
an important role in increasing the adhesive strength of the
tight junction in renal tubular epithelia.
Renal ischemic injury is a commonly observed compli-
cation in hospitalized patients that results in multiple
impairments of water and ion transport (Clarkson et al.
2008). The disruption of tight junctions is an important
component of the pathogenesis of renal ischemic injury and
causes polarity alteration of tubular epithelium (Fish and
Molitoris 1994; Lee et al. 2006; Molitoris et al. 1989,
1992). Distorted distributions of the tight junction proteins
occludin and ZO-1 have been observed with ATP depletion
(an in vitro model of renal ischemia) in cultured cells
(Bacallao et al. 1994; Gopalakrishnan et al. 2007). ZO-1
staining in renal allograft recipients with sustained acute
renal failure exhibits diminished intensity and redistribu-
tion in the proximal tubule (Kwon et al. 1998). Ischemic
injury also has been shown to induce glomerular podocyte
eVacement in association with ZO-1 dissociation (Wagner
et al. 2008). However, the eVects of ischemic injury on
occludin and ZO-1 distribution in renal epithelial cells have
not yet been fully investigated in the kidney.
In the present study, we show that renal ischemia–reper-
fusion injury causes cellular damages not only in the proxi-
mal tubule but also in the collecting duct and that the
disruption of tight junctions is cell-speciWc to intercalated
cells in the collecting duct. The collecting duct is composed
of two histologically diVerent cell types, the principal cells
and intercalated cells. The principal cells reabsorb water,
whereas the intercalated cells are involved in acid–base
homeostasis (Madsen and Tisher 1986). Renal ammonia
excretion is a central component of net acid excretion and
ischemia–reperfusion injury decreases renal ammonia
excretion (Han et al. 2007; Wagner et al. 2011; Weiner and
Verlander 2011). We show that disruption of tight junctions
is associated with the polarity loss or delocalization of the
ammonia transporting Rhesus protein, Rhcg, as well as
other acid–base transporter proteins in intercalated cells.
The results may have signiWcant value in understanding
clinical manifestations of renal ischemic injury such as
metabolic acidosis.
Materials and methods
Animals and surgical procedures
Male Sprague–Dawley rats weighing 200–250 g were used
in all experiments. Animals were maintained in a tempera-
ture-controlled room with alternating 12:12-h light–dark
cycles and had free access to water and food throughout the
experiment. These studies were approved by the Ewha
Womans University Institutional Animal Care and Use
Committee (EMRI 07-0073, 08-0083).
Animals were anesthetized with an intraperitoneal injec-
tion of tiletamine/zolazepam (10 mg/kg; Zoletil 50, Virbac
Laboratories, Carros, France) and 2% xylazine hydrochlo-
ride (2 mg/kg; Rumpun, Bayer Korea, Ansan, Korea). After
the abdomen was opened through a midline incision, both
renal pedicles were exposed and cleaned by blunt dissec-
tion. Microvascular clamps (RS-5422, Roboz Surgical
Instrument Co., Inc., Gaithersburg, MD, USA) were placed
on both renal arteries to completely block renal blood Xow.
After 30 min, the clamps were removed and blood Xow
returned to the kidneys. Control animals underwent sham
operation. The animals were killed 6 h after reperfusion.
Arterial blood samples of the animals were collected from
the abdominal aorta. The collected blood sample was
placed on ice, and the blood gas was immediately measured
(Beckman Instruments, Palo Alto, CA, USA).
Tissue preservation
The kidneys were Wrst perfused brieXy through the abdomi-
nal aorta with cold PBS and subsequently perfused with the
Wxative solution, periodate-lysine-2% paraformaldehyde
(PLP), for 10 min. The kidneys were cut transversely into
1- to 2-mm thick slices and immersed in PLP overnight at
4°C. The tissue was rinsed in PBS, dehydrated in a series of
ethanol, and embedded in wax (Polysciences, Inc.,
Warrington, PA, USA) for light microscopy.
Antibodies
Mouse anti-occludin monoclonal antibodies (Zymed Labo-
ratories, San Francisco, CA, USA; catalog No 33-1500) and

Page 3

Histochem Cell Biol (2011) 136:637–647639
123
rabbit anti-ZO-1 polyclonal antibodies (Zymed Laborato-
ries, San Francisco, CA, USA; catalog No 40-2300) were
used. Aquaporin 1 (Chemicon Inc., Temecula, CA, USA;
catalog No AB3065), bumetanide-sensitive Na+/K+/2Cl¡
cotransporter (NKCC2; Jeon et al. 2007), aquaporin 2
(Chemicon Inc., Temecula, CA, USA; catalog No
AB3066), Rh C glycoprotein (Rhcg; Han et al. 2007), and
anion exchanger 1 (AE1; Alpha Diagnostics, San Antonio,
TX, USA; catalog No AE11-A) were used to identify the
proximal tubule and descending thin limb, thick ascending
limb, collecting duct principal cells, collecting duct interca-
lated cells, and type A intercalated cells, respectively.
Light microscopic immunohistochemistry
Four micrometer-thick wax sections were processed for
immunohistochemistry using immunoperoxidase proce-
dures, as described previously (Han et al. 2003, 2004,
2005). The sections were dewaxed with xylene and ethanol
and rinsed in PBS. Endogenous peroxidase activity was
blocked by incubating the sections in 3% H2O2 for 30 min.
The sections were treated with proteinase K (DakoCytoma-
tion, Carpinteria, CA, USA) for 10 min and blocked with
the blocking solution (DakoCytomation, Carpinteria, CA,
USA) for 30 min before incubation with primary antibody
(occludin 1:200; aquaporin 1:1000; NKCC2 1:500) over-
night at 4°C. The sections were washed in PBS and incu-
bated for 1 h with peroxidase-conjugated secondary
antibodies against mouse or rabbit IgG (Jackson Immuno-
Research laboratories, West Grove, PA, USA) diluted
1:100 in PBS. The sections were again washed with PBS
and then exposed to a mixture of 0.05% 3,3?-diaminobenzi-
dine and 0.01% H2O2. The sections were washed in Tris–
HCl buVer, dehydrated with graded ethanol and xylene,
mounted, and observed by light microscopy.
Double immunohistochemistry
Colocalization was accomplished using sequential immun-
operoxidase procedures, as described previously (Han et al.
2007). BrieXy, the Wrst immunoperoxidase procedure used
the protocol for single labeling described above. After the
diaminobenzidine reaction, the sections were washed in
PBS and incubated in 3% H2O2 for 30 min and blocked
with the blocking solution. The sections were treated for
1 h with the second primary antibodies (aquaporin 1
1:1000; NKCC2 1:500; Rhcg 1:1000), washed in PBS,
incubated with the secondary antibodies, and then washed
with PBS. Vector SG (Vector Laboratories, Burlingame,
CA, USA) was used as the chromogen to produce a blue
label easily distinguishable from the brown label produced
by the diaminobenzidine used for detection of the Wrst pro-
tein. The sections were washed in Tris–HCl buVer, dehy-
drated with graded ethanol and xylene, mounted, and
observed by light microscopy.
Confocal microscopy
Kidney sections were dewaxed and rehydrated in a graded
series of ethanol. The sections were treated with proteinase
K (DakoCytomation, Carpinteria, CA, USA) for 10 min
and blocked with the blocking solution (DakoCytomation,
Carpinteria, CA, USA) for 30 min before incubation with a
mixture of primary antibodies (occludin 1:100; AE1 1:100)
overnight at 4°C. The sections were washed in PBS and
incubated for 1 h with a mixture of Cy3-conjugated donkey
anti-mouse IgG and FITC-conjugated donkey anti-rabbit
IgG (Jackson ImmunoResearch laboratories, West Grove,
PA) diluted 1:100 in PBS, respectively. Coverslips were
mounted with VectaShield mounting media. Images were
captured on a Zeiss LSM5 PASCAL.
Immunoblot analysis
Kidneys were divided into the cortex and medulla and pro-
cessed for immunoblot analysis, as described previously
(Han et al. 2005). Tissues were homogenized in lysis buVer
containing 20 mM Tris–HCI, 1% Triton X-100, 150 mM
sodium chloride, 0.5% sodium deoxycholate, 0.1% sodium
dodecyl sulfate, 0.02% sodium azide, 1 mM EDTA, 10 ?M
leupeptin, and 1 mM phenylmethylsulfonyl Xuoride. The
homogenate was centrifuged at 3,000g for 20 min at 4°C.
After determination of protein concentration in the superna-
tant by the Coomassie method (Pierce, Rockford, IL, USA),
samples were loaded (20 ?g/lane) and underwent electro-
phoresis on sodium dodecyl sulfate-polyacrylamide gels
under reducing conditions. Proteins were transferred to
nitrocellulose membranes by electroblotting. To reduce
nonspeciWc antibody binding, the membranes were blocked
with 5% nonfat dried milk for 30 min at room temperature,
and then incubated for 24 h at 4°C with anti-occludin
antibody (1:1,000). After three washes, the blot was
incubated with a peroxidase conjugated donkey anti-mouse
IgG (1:1,000) for 2 h at room temperature. Samples were
visualized using an enhanced chemiluminescence system
(Amersham Life Science, Buckinghamshire, UK) after a
5- to 30-min exposure at room temperature. Densitometric
analysis was performed using the Zero-Dscan software of
the Eagle EYETMII Still Video System (Stratagene, La
Jolla, CA, USA).
Statistics
Results are presented as mean § SD. Statistical analyses
were performed using Student’s unpaired t test, and
P < 0.05 was taken as statistical signiWcance.

Page 4

640Histochem Cell Biol (2011) 136:637–647
123
Results
Physiological parameters
The results of the arterial blood gas are summarized in
Table 1. Renal ischemia–reperfusion injury decreased sig-
niWcantly blood bicarbonate level compared with control
animals (20.6 § 2.36 vs. 25.1 § 2.02 mmol/L). Arterial
carbon dioxide tension (PaCO2) was slightly reduced indi-
cating a respiratory compensation in ischemic rats. There
were no signiWcant diVerences in serum sodium, potassium,
and chloride concentrations.
Localization of occludin and ZO-1 in the normal rat kidney
Occludin was expressed from the cortex to the tip of the
inner medulla. In both normal and sham-operated control
kidneys, strong occludin immunoreactivity was observed in
the distal nephron segments, including the thick ascending
limb, distal convoluted tubule, and collecting duct. High-
magniWcation microscopic examination demonstrated a
classic dot or thread-like expression pattern at the apical
end of the lateral membrane in the tubular epithelial cells,
as expected for the location of the tight junction (Fig. 1a, b).
Occludin immunolabeling was not detected in the glomeru-
lus, proximal tubule, and blood vessels in the rat kidney
(Fig. 1a, b). The thin limbs of the loop of Henle showed
moderate occludin immunoreactivity. To identify the
descending thin limb, we used aquaporin 1 as a marker.
Occasionally, the transition point from the descending thin
limb to the ascending thin limb was observed in the inner
medulla, and occludin was expressed in both thin limbs
(Fig. 1f–h).
ZO-1 immunoreactivity was strong in the thick ascend-
ing limb, distal convoluted tubules, and collecting duct and
was moderate in thin limbs of the loop of Henle in the nor-
mal kidney (Fig. 1c–e). In contrast to occludin, ZO-1
expression was also observed in the parietal epithelium of
Bowman’s capsule and in vascular endothelial cells, includ-
ing the arcuate artery, interlobular artery, glomerular capil-
laries, peritubular capillaries, and vascular bundles
(Fig. 1c–e). Also in contrast to occludin, ZO-1 immunola-
bel was observed in the proximal tubule, although the
intensity of immunolabel was less than that observed in dis-
tal nephron segments (Fig. 1c).
Abnormal localization of occludin and ZO-1 after renal
ischemia–reperfusion injury
After ischemia–reperfusion injury, abnormal localization of
occludin was observed in the outer medulla region (Fig. 2).
Typical dot or thread-like patterns of occludin immunore-
activity was disrupted. Instead, occludin was distributed
diVusely in the cytoplasm in a subset of tubular epithelial
cells in the outer medulla of IR injury kidneys (Fig. 2d).
Because there are three diVerent tubular segments (the
proximal straight tubule, thick ascending limb, and collect-
ing duct) in the outer medulla, we used speciWc marker pro-
teins to identify the diVerent epithelial segments and cells
aVected by IR injury. We used AQP1 to identify proximal
tubule segments, NKCC-2 to identify the thick ascending
limb of the loop of Henle and Rhcg, and AQP2 and AE1 to
identify collecting duct segments.
Many aquaporin 1-positive proximal tubule cells were
damaged and detached into the tubular lumen in IR kidneys
(Fig. 2e). However, occludin immunoreactivity was not
detected in the proximal tubule cells in either control or
ischemic kidneys (Fig. 2c, f).
NKCC2-positive thick ascending limb cells were rela-
tively intact and did not show detectable morphological
damage in this model of ischemia–reperfusion injury
(Fig. 3). Occludin expression was strong in the thick
ascending limb in both control and IR kidney, and there
was no change in occludin distribution in the thick ascend-
ing limb in response to IR injury (Fig. 3b, d).
Ischemia–reperfusion injury caused considerable dam-
age to a subset of cells in the outer medullary collecting
duct (OMCD). To determine whether the damage is speciWc
to either intercalated cells or principal cells, the two pri-
mary cell types present in the collecting duct, we double-
immunolabel with Rhcg, to identify intercalated cells, and
AQP2, to identify principal cells. We observed that Rhcg-
positive intercalated cells were detached frequently from
the basement membrane, whereas principal cells, in con-
trast, appeared intact and undamaged (Fig. 4a, f). To iden-
tify whether disturbances in occludin localization were
present and whether changes in its localization were spe-
ciWc to intercalated cells, we performed double-immunola-
bel Xuorescent microscopy using antibodies to occludin and
the intercalated cell marker, AE1. Confocal Xuorescent
microscopy conWrmed that occludin disruption in the
Table 1 EVects of ischemia–reperfusion injury on blood pH and elec-
trolyte concentration
Values are mean § SD of Wve rats (*P < 0.05)
Control IR
pH
HCO3
pCO2 (mmHg)
Plasma Na+ (mmol/L)
Plasma K+ (mmol/L)
Plasma Cl¡ (mmol/L)
7.38 § 0.04
25.1 § 2.0
44.1 § 7.1
131 § 3.0
4.6 § 0.6
108 § 2
7.37 § 0.05
20.6 § 2.4*
37.1 § 6.3
132 § 2.1
5.1 § 0.4
111 § 1
¡ (mmol/L)

Page 5

Histochem Cell Biol (2011) 136:637–647641
123
OMCD was speciWc to damaged intercalated cells and that
occludin lost its typical apical tight junction localization
and was localized diVusely in detached intercalated cells.
Occludin expression was only slightly delocalized in cells
remaining in the OMCD (Fig. 4b–e, g–j).
IR injury also resulted in signiWcant changes in ZO-1
immunolabel in the outer medulla. However, in contrast to
occludin, ZO-1 distribution was disrupted in both the prox-
imal tubule and the collecting duct, but not in the thick
ascending limb of the loop of Henle (Fig. 5). In both the
Fig. 1 Occludin and ZO-1 expression in control kidney. Representa-
tive of occludin (a, b), ZO-1(c–e), and occludin-AQP1 confocal
(f–h) immunostaining images obtained from sham-operated control
kidneys. Occludin (brown) immunoreactivity was observed as dot-
and-line patterns at the apical end of the lateral membrane in tubular
epithelial cells in the cortex (a) and outer medulla (b). Occludin was
expressed primarily in the distal tubule (DT) including the thick
ascending limb (TAL). No occludin immunolabel was detected in
glomeruli (G), proximal tubule (PT), or blood vessels, including the
vascular bundles (VB). ZO-1 (brown) immunoreactivity was observed
in all tubular segments, including the proximal tubule (PT) and distal
tubule (DT) (c–e). ZO-1 labeling also appeared in the glomerular pari-
etal epithelial cells (arrows) and in endothelial cells (arrowheads) of
all blood vessels including the glomerular capillary, peritubular capil-
lary, arcuate artery (AA), and vascular bundle (VB). To identify the thin
limbs of the loop of Henle, confocal immunoXuorescence staining for
occludin (red) and AQP1 (green) were performed (f–h). Occasionally,
transition points (white arrows) from the AQP1-positive descending
thin limb (DTL) to AQP1-negative ascending thin limb (ATL) were ob-
served and occludin immunoreactivity appeared in both thin DTL and
ATL segments (g, h). Results are representative of Wndings in the Wve
control rat kidneys
fgh
DTL
ATL
Occludin + AQP1
OccludinZO-1
OccludinAQP1
ac
bde
VB
G
PT
G
PT
DT
AA
VB
DT
TAL
PT
20 µm
20 µm