Cystic Fibrosis/Pulmonary Research and Treatment Center, University
of North Carolina, Chapel Hill, North Carolina 27599-7248
sodium-potassium-chlorine ion cotransporter; chloride secretion; bicarbonate secretion; jejunum; cecum
 |
INTRODUCTION |
INTESTINAL SECRETION IS
IMPORTANT in maintaining the liquidity of intestinal contents
during digestion, which aids the mixing of digestive enzymes with the
luminal contents. In addition, the liquid secreted in the crypts serves
to flush the mucus produced in the crypts to the surface epithelium,
where it protects and lubricates the mucosa. In most intestinal
regions, Cl
has been thought to be the primary anion
secreted, which induces liquid secretion isosmotically. Defective
regulation of Cl secretion underlies major intestinal
pathophysiology ranging from diarrhea (hypersecretion of liquid) to
cystic fibrosis (CF; hyposecretion of liquid).
Studies in the CF mouse have established that an absence of
Cl secretion is associated with crypt dilation, goblet
cell hyperplasia, and intestinal obstruction (see Ref. 8 for review).
In light of the Cl channel functions described for the CF
transmembrane conductance regulator (CFTR) and its apical membrane
localization in gut epithelia, the data from the CF mouse have
established CFTR as the major path for conductive Cl exit
across the enterocyte apical membrane during anion secretion.
Although the focus of much research has been on the apical
Cl channel (CFTR) responsible for intestinal secretion,
the basolateral pathway(s) by which Cl enters the cell
has received less attention. The basolateral Na+-K+-2Cl cotransporter (NKCC1)
is thought to play a key role as a Cl entry pathway in
intestinal epithelia (9, 11) to sustain Cl
secretion. In intestinal epithelia, NKCC1 electroneutrally transports Cl across the basolateral membrane along with
Na+ and K+, with a stoichiometry of 1 Na+:1 K+:2 Cl (9).
The loop diuretics bumetanide and furosemide are effective blockers of
NKCC1 and have been used to determine the contribution of the
cotransporter to the secretory response of the intestine. However, a
more specific strategy to identify the contribution of basolateral
NKCC1 to the basal and stimulated Cl secretion of the
intestine is the use of gene targeting. In the present study, we have
inactivated Slc12a2, the gene that codes for the NKCC1
cotransporter, and tested the intestinal epithelia of the mutant mice
for the expression of a gut phenotype in Ussing chamber studies
designed to identify the NKCC1 cotransporter contribution to basal and
regulated anion secretion in two different regions of the mouse
gastrointestinal tract.
| |
MATERIALS AND METHODS |
Adult animals (3-8 mo old) were used in this investigation.
All mice were allowed food and water ad libitum until the time of the
experiment. In all cases, homozygous normal littermates (referred to as
+/+, normal, or wild-type) were compared with the NKCC1-deficient mice
homozygous for mutant Slc12a2 (referred to as NKCC1 
/).
Two mutations in Slc12a2 were generated (20), and mice bearing these mutations were studied. Ion-transport data for
only one of the mutations
(Slc12a2
1065-1137) are reported, because
the physiological characterization of tissues from mice bearing the two
mutations was identical. Mice carrying each mutation (not all mice
studied were weighed) had a lower body mass than that of normal
littermates [19.7 ± 1.2 and 29.9 ± 1.8 g for
Slc12a21065-1137 mice vs. normal
littermates, respectively (n = 6 for each group); 18.7 ± 0.77 and 24.2 ± 0.72 g for
Slc12a2506-621 mice vs. normal
littermates, respectively (n = 5 for each group)]. Mice were killed with CO2, and the intestine was quickly
removed and placed in oxygenated Krebs buffer. The designated region of the intestine was opened lengthwise, flushed with buffer to remove intestinal contents, and then mounted on Ussing chambers having an
exposed surface area of 0.25 cm2 (7). All
preparations studied (ion transport and Northern analysis) were "full
thickness" with the muscle layer and the enteric nervous system still
intact. Tissues were studied under short-circuit current
(Isc) conditions for the duration of the experiment. A constant voltage pulse (1-5 mV, 1-s duration) was applied to the tissue every minute. Potential difference and resistance were calculated using Ohm's law from the changes in
Isc in response to the voltage pulse.
Solutions and drugs.
Several different solutions were used in this investigation. Unless
otherwise indicated, the tissues were mounted and bathed in
Krebs-HCO3 Ringer (KBR) solution having the following
composition (in mM): 140 Na+, 120 Cl, 5.2 K+, 1.2 Mg2+, 1.2 Ca2+, 2.4 HPO42, 0.4 H2PO4, and
25 HCO3. A Cl-free buffer was used to
determine the contribution of Cl to the basal and
stimulated Isc. For a Cl-free,
HCO3-replete buffer, 115 mM Na+ gluconate
replaced NaCl, 1.2 mM MgSO4 replaced MgCl2, and
1.2 mM CaCl2 was replaced by calcium gluconate (6 mM
calcium gluconate was added to offset the calcium-chelating effects of
gluconate). In some preparations, basolateral Na+-free
buffer was used. For an Na+-free buffer, the solution was
identical to KBR except that
N-methyl-D-glucamine replaced all
Na+. These three buffers were gassed with
95%O2-5% CO2 to maintain pH 7.4. The
Cl-free, HCO3-free buffer was similar
to the Cl-free buffer except that 125 mM Na+
gluconate was used to replace all Cl and
HCO3 was replaced by 10 mM HEPES buffer titrated to
pH 7.4 with 1 M Tris base. The HCO3-free and
Cl-free, HCO3-free buffers were gassed
with 100% O2. Whenever HCO3 was removed
from the medium, 103 M acetazolamide was added to the
buffer to prevent endogenous tissue HCO3 production.
Where indicated, some preparations were treated with TTX
(106 M basolateral) to block neurotransmitter release
from the enteric nervous system. All preparations contained 5 mM
glucose in the basolateral solution and 5 mM mannitol in the apical solution.
Selected drugs were used to characterize the basolateral anion entry
paths (see Fig. 9). Bumetanide, an inhibitor of NKCC1, was added to the
basolateral bath to test for inhibition of the basal
Isc or to test for inhibition of the
forskolin-stimulated current. The bumetanide sensitivity was calculated
as the magnitude of change in Isc over a 4-min
period after bumetanide addition. DIDS was used to block both
Cl/HCO3 exchange and
Na+-HCO3 cotransport. Dimethylamiloride
(DMA) was used to block Na+/H+ exchange.
Hydrochlorothiazide, which blocks NaCl-coupled exchange (see Ref. 9 for
review), was also tested in intestinal tissue. Drugs were used at the
following concentrations (in M): 104 bumetanide
(basolateral), 105 forskolin (bilateral),
103 acetazolamide (bilateral), 103 DIDS
(basolateral), 105 DMA (basolateral), 104
hydrochlorothiazide (basolateral), and 104 amiloride
(apical). All drugs and chemicals were purchased from Sigma Chemical
(St. Louis, MO) with the exception of DIDS, which was purchased from
Molecular Probes (Eugene, OR).
Northern analysis.
Total RNA was isolated from salivary and intestinal tract tissue of
wild-type animals. Tissue was frozen, homogenized, and phenol/chloroform extracted using RNAzol B (Tel-test, Friendswood, TX).
Total RNA (20 µg) was electrophoresed on a 1.1% formaldehyde, 1.2%
agarose denaturing gel in the presence of ethidium as described by
Kroczek and Siebert (16) and transferred to Immobilon-NC nitrocellulose membrane (Millipore, Bedford, MA) by capillary transfer.
Radiolabeled DNA probes were hybridized to Northern blots in Quick-Hybe
(Stratagene) for 1 h at 68°C.
Histological analysis.
Wild-type and Slc12a2506-621 mice were
killed, and intestinal tract tissues were fixed in 10%
neutral-buffered formalin. The fixed tissues were embedded in paraffin,
cut into 5-µm sections, and stained with aniline blue and periodic
acid-Schiff's reagent.
In situ analysis.
A probe for in situ analysis was prepared in the following manner. The
primers NKCC3a, 5'-CAG GGC CTG CTT TACTTCATCTTG-3', and NKCC3b, 5'-GCC
TTT CCG TGC GAC TGG-3', were used to generate a 1.2-kb cDNA probe from
salivary gland mRNA by RT-PCR. The fragment was cloned into pCR 2.1, the clone was digested with Hind III to remove a 700-bp
fragment of cDNA, and this clone was religated to give a cDNA probe of
~600 bp corresponding to bases 3127 to 3636 of the published mouse
NKCC1 sequence (18). Using this construct,
35S-labeled sense and antisense RNA was prepared
(Maxiscript SP6/T7 kit; Ambion, Austin, TX) and used to analyze tissue
sections prepared as follows. Tissues were fixed in 4%
paraformaldehyde and washed with 30% sucrose in PBS to remove the
fixative. Tissues were embedded in Tissue-Tek, and cryostat sections
cut at 8 µm were mounted on slides and stored at 80°C. Sections
were fixed in 4% paraformaldehyde and dehydrated in a graded series of
ethanol washes. Sections were then digested with proteinase K (10 µg/ml) for 30 min at 30°C. Proteinase K was inactivated by addition
of 4% paraformaldehyde, and the sections were rinsed in
triethanolamine and acetylated with 0.25% acetic anhydride for 10 min.
Sections were then rinsed in 0.2× SSC (1× SSC is 0.15 M NaCl and
0.015 M sodium citrate, pH 7.0) and dehydrated. A quantity of NKCC1
sense or antisense probe producing 1 × 107 cpm was
hydridized to the sections overnight at 54°C in 50% formamide, 1×
Denhardt's solution, 0.6 M NaCl, 10 mM Tris, pH 8.0, 1 mM EDTA, 0.1%
SDS, 10 mM dithiothreitol (DTT), 1 mg/ml yeast transfer RNA, and 10%
dextran sulfate. Slides were washed with 4× SSC and treated with 20 µg/ml RNase at 37°C, then washed four times with 2× SSC and 1 mM
DTT at room temperature and washed three times with 0.5× SSC and 1 mM
DTT at 54°C. Slides were dehydrated and hand-dipped in NTB2 emulsion
(Eastman Kodak, Rochester, NY), exposed for 2 wk at 4°C, developed,
and stained with hematoxylin and eosin.
Statistics.
All data are means ± SE. Only one preparation per animal (of each
intestinal region) was studied for each protocol. Data were compared by
a Student's t-test if only two groups were being compared. If more than two groups were compared, ANOVA was used and a
Student-Newman-Keuls test was used for multiple comparisons among groups.
| |
RESULTS |
Expression of NKCC1 in mouse intestinal tract.
To examine the expression of Slc12a2 in the mouse intestinal
tract, Northern analyses were carried out on RNA prepared from various
regions of the intestinal tract of normal animals. Expression was
detected in all regions examined; however, expression was higher in the
distal regions, with the highest levels seen in the cecum and colon
(Fig. 1A). In situ analysis of
a section from wild-type jejunum was performed to determine the
expression pattern of the message for the Slc12a2 gene
within this tissue. Slc12a2 expression was largely
restricted to the epithelium lining the crypts (Fig. 1C). No
expression was detected in the muscle layers of the intestinal tract.
View larger version (120K):
[in this window]
[in a new window]
|
Fig. 1.
A: expression of Slc12a2 [which codes for
Na+-K+-2Cl cotransporter
(NKCC1)] in the mouse intestinal tract as shown by Northern blot
analysis of RNA prepared from various regions of the intestinal tract
and the salivary gland of normal mice. Slc12a2 expression
increases in the distal regions of the intestinal tract; however, even
in these regions, expression levels are not as high as those observed
in the salivary gland. B and C: in situ analysis
of Slc12a2 expression in the mouse jejunum indicates that
expression is largely confined to the crypts; bars = 200 µm. The
section shown in B is hybridized with a probe corresponding
to the sense strand of the Slc12a2 transcript, whereas that
shown in C is analyzed with antisense probe. D:
normal adult mouse jejunum. E: NKCC1 / mouse jejunum
exhibiting scattered dilated crypts (arrows). D and
E: stained with aniline blue and periodic acid-Schiff's
reagent; bars = 50 µm.
|
|
Neonatal death associated with intestinal obstruction is common in mice
deficient in the CFTR gene, and the growth rate of the surviving pups
is often lower than that of normal littermates (see Ref. 8 for review).
No neonatal deaths were observed in the NKCC1 / mice in this study,
in contrast with data previously described (5). Mice
deficient in NKCC1 were growth retarded, and the majority of these
animals fail to reach the weight of littermates (20). To
determine whether loss of NKCC1 resulted in alterations in the
intestinal epithelia, histological analysis was carried out on tissue
from all regions of both neonates and adult NKCC1-deficient animals. As
can be seen in Fig. 1, D and E, the intestinal
epithelium was largely normal. Comparison with wild-type tissue did,
however, reveal a slight increase in the frequency of dilated crypts in
the NKCC1-deficient tissue.
Jejunum.
Ion transport across a midjejunal region was characterized for
mice carrying the
Slc12a21065-1137 and
Slc12a2506-621 mutations. "Basal"
refers to the Isc recorded 30 min after
mounting. There were no significant differences in either the basal or
stimulated Isc between the two
Slc12a2 mutations and their respective controls. Therefore,
only the jejunal data from the
Slc12a21065-1137 jejuna (and littermate
controls) are shown.
The basal bioelectric properties for the wild-type and / jejuna
bathed in KBR (as well as 0 HCO3 and 0 Cl) are shown in Table 1.
When bathed in bilateral KBR, the jejuna from the / mice exhibited
a net basal Isc that did not differ significantly from the control (+/+) jejuna (Fig.
2A; Table 1). The
basal Isc (determined at the midpoint of the
oscillations) of both the control and the / preparations exhibited
spontaneous oscillations that have been previously characterized as
oscillations in the basal rate of anion secretion mediated by
neurotransmitter release from the enteric nervous system (Refs. 7 and
25; Fig. 3, A and
C). The magnitude of these oscillations was virtually identical for control and / mice (Table
2). The basal Isc
of the +/+ jejuna exhibited a small decrease in response to basolateral bumetanide, whereas the basal Isc of the /
jejuna was unresponsive to bumetanide (Fig. 2B, KBR). The
bumetanide-insensitive basal Isc was similar for
the two genotypes (Fig. 2B). Bumetanide also significantly
reduced the magnitude of the oscillations in the +/+ preparation (see
Ref. 7) but had no effect on the oscillations in the / preparations
(data not shown). Interestingly, the transmembrane conductance in KBR
was significantly greater in the / jejunum compared with the
wild-type preparations (Table 1).
View larger version (18K):
[in this window]
[in a new window]
|
Fig. 2.
A: effect of ion-replacement protocol on net
basal short-circuit current (Isc) in jejuna.
Jejuna were incubated in the indicated buffer for 30 min before the
measurements were taken. Open bars, normal tissue; solid bars, NKCC1
/ tissue. KBR, Krebs-HCO3 Ringer buffer bilateral
(n = 14 normal; n = 18 NKCC1 /); 0 Cl, nominally Cl-free bilateral
(n = 5 normal; n = 8 NKCC1 /). 0 HCO3, bilateral HCO3-free buffer
bilateral (n = 8 normal; n = 9 NKCC1
/). 0 Cl/0 HCO3, nominally
Cl- and HCO3-free bilateral
(n = 7 normal; n = 4 NKCC1 /).
** P 0.01 vs. respective tissue in KBR. B:
effect of bumetanide on basal Isc. Open bars,
bumetanide-sensitive component of basal Isc in
the +/+ jejuna; hatched bars, bumetanide-insensitive component of basal
Isc in the +/+ jejuna; solid bars,
bumetanide-insensitive component of basal
Isc of the / tissue (none of the /
tissue responded to bumetanide). * P0.05,
bumetanide-insensitive Isc vs. basal
Isc (bumetanide insensitive) in tissue incubated
in KBR for the respective genotype;
+ P0.05, bumetanide-sensitive
Isc significantly different vs.
bumetanide-sensitive Isc in KBR and
bumetanide-insensitive Isc (+/+) vs. KBR (+/+)
jejuna. n = 4 for each group. All data are means ± SE.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
Fig. 3.
A and B: recorder trace of
Isc from normal jejunum without and with DIDS
application, respectively. C and D: recorder
trace of Isc from NKCC1 / jejunum without
and with DIDS application, respectively. Forsk, 105 M
forskolin serosal; Bumet, 104 M bumetanide serosal;
glucose, 10 mM mucosal; DIDS, 103 M serosal.
|
|
Most of the preparations in this investigation were studied as
full-thickness preparations with the enteric nervous system intact. To
determine the role of the enteric nervous system in maintaining the
Isc in wild-type and / tissue, some
preparations were treated basolaterally with TTX. In both the normal
and / jejuna, TTX caused a significant decrease in the magnitude of the basal Isc but the post-TTX
Isc did not differ between the genotypes (Fig.
4A). The post-TTX
conductance was again significantly greater in the / preparations
(data not shown). The influence of bumetanide on the basal
Isc was not studied in the TTX-treated preparations.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 4.
A: basal and post-TTX
Isc in jejunal preparations incubated in KBR or
bilateral HCO3-free buffer. * P0.05
vs. basal KBR for respective genotype. ** P0.01 vs.
same genotype post-TTX KBR. In KBR-incubated preparations,
n = 8 and 9 for +/+ and NKCC1 /, respectively; in 0 HCO3-incubated preparations, n = 4 for both genotypes. B: forskolin response of TTX-treated
tissue incubated in KBR or bilateral HCO3-free
buffer. ** P0.01 vs. +/+ tissue incubated in the same
buffer. In KBR-incubated preparations, n = 8 and 9 for
+/+ and /, respectively; in 0 HCO3-incubated
preparations, n = 4 for both genotypes. A
and B: open bars, +/+ tissue; solid bars, NKCC1 /
jejuna. C: effect of bumetanide on the forskolin-stimulated
Isc in TTX-treated jejuna. Open bars,
bumetanide-sensitive Isc +/+ tissue; hatched
bars, bumetanide-insensitive Isc +/+ tissue;
solid bars, bumetanide-insensitive Isc /
tissue. + P0.05, +/+ vs.
/bumetanide-sensitive Isc in KBR;
* P 0.05 bumetanide-insensitive
Isc / vs. +/+ jejuna in KBR;
++ P 0.01, bumetanide-sensitive
Isc +/+ vs. / in 0 HCO3;
** P 0.01 bumetanide-insensitive
Isc in 0 HCO3 buffer / vs.
bumetanide-insensitive Isc / in KBR buffer.
Sample sizes same as in B. Values are means ± SE.
|
|
To identify potential routes for Cl influx under basal
conditions across the basolateral enterocyte membrane, the jejuna of NKCC1 / mice were studied using ion-substitution protocols. Bilateral Cl replacement eliminated the spontaneous
oscillations in the basal Isc of both the
wild-type and the / jejuna (Table 2). In bilateral Cl-free buffer, net basal Isc was
significantly and proportionately reduced in both genotypes (Fig.
2A and Table 1), and tissues of neither genotype responded
to bumetanide (Fig. 2B, 0 Cl). As in KBR, the
conductance of the / preparations was significantly increased
compared with the +/+ tissue (Table 1).
In another series of ion-substitution experiments,
HCO3 was removed from the buffer bathing both tissue
surfaces. HCO3 substitution had no significant effect
on the magnitude of the net basal Isc in the
control jejuna. However, this maneuver significantly reduced the
magnitude of the Isc in the / jejuna
compared with tissue incubated in KBR (Fig. 2A and Table 1).
In HCO3-free buffer, the Isc
of the +/+ jejuna had a significantly greater response to bumetanide
than did the +/+ preparations in KBR (Fig. 2B). The /
tissue again was not inhibited by bumetanide in this buffer.
Interestingly, bilateral removal of HCO3
significantly reduced the magnitude of the Isc
oscillations in both the normal and / jejuna (Table 2). The
conductance was again significantly greater in the / tissue
compared with the +/+ preparations in HCO3-free
solutions (Table 1).
A number of TTX-treated preparations were studied in bilateral
HCO3-free buffer. HCO3-free buffer
had no significant effect on the magnitude of the net basal
Isc of the normal preparations (post-TTX)
compared with normal TTX-treated preparations in KBR (Fig.
4A). However, the post-TTX Isc in the
/ jejuna incubated in HCO3-free buffer was
significantly reduced (3-fold) compared with the TTX-treated /
jejuna incubated in KBR (Fig. 4A).
Removal of both HCO3 and Cl bilaterally
from the solutions bathing the tissues caused a significant reduction
in the magnitude of the Isc of the normal tissue
(compared with tissue in KBR and HCO3-free KBR but
not Cl free alone) but did not significantly reduce the
magnitude of the Isc in the / tissue
compared with tissues incubated in bilateral HCO3- or
Cl-free medium (Fig. 2A). We did not test the
effect of bumetanide on the basal Isc of these
tissues. However, because bumetanide had no effect on the basal
Isc in the bilateral Cl-free
group, this drug would not be expected to have an effect on this group
of tissues.
We also used selected drugs to test for other transport paths that
could account for basolateral anion influx in wild-type and mutant
jejunal epithelia. In the normal jejunum, basolateral DIDS caused a
large transient increase in Isc [
= 65.7 ± 2.7 µA/cm2 (n = 7)] (Fig.
3B). After DIDS application, the oscillations in the basal
Isc tended to be somewhat smaller than the
magnitude of the pre-DIDS oscillations; however, this difference was
not significant, and DIDS had no detectable effect on basal
Isc [56.6 ± 7.1 vs. 65.5 ± 4.4 µA/cm2 for pre- vs. post-DIDS (5 min); Fig.
3B]. In contrast, in the NKCC1 / jejuna, basolateral
DIDS caused a small, transient (2-3 min) increase in the
Isc (Fig. 3D). Three to five minutes
after DIDS application, the oscillations in the basal
Isc disappeared (n = 5) (Fig.
3D), and the magnitude of the basal
Isc was significantly reduced [74.1 ± 16 vs. 45.3 ± 7.9 µA/cm2 for pre- vs. post-DIDS
(n = 5 for both); P0.05].
Mice with mutant NKCC1 exhibited major differences from wild-type mice
in response to cAMP-stimulated Cl secretion. In KBR
buffer, the jejuna of the / mice responded to forskolin with an
increase in Isc that was markedly reduced compared with the forskolin response exhibited by the normal jejuna (Fig. 3, A and C, and Fig.
5A, KBR data). In the
+/+ tissue, virtually the entire forskolin response was inhibited by
bumetanide, whereas the forskolin response of the / jejuna
exhibited no significant response to bumetanide (Fig. 3, A
and C, and Fig. 5B, KBR data).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 5.
A: jejunal Isc response
to forskolin in preparations incubated in indicated buffers (see Fig.
1). Open bars, +/+ (normal) tissue; solid bars, NKCC1 / jejuna.
* P0.05 vs. normal jejuna in KBR;
§ P0.05 vs. /jejuna in KBR. In KBR,
n = 10 for both genotypes; in 0 Cl and 0 Cl/0 HCO3, n = 5 and 4 for normal and NKCC1 /, respectively; in 0 HCO3,
n = 4 and 5 for normal and NKCC1 /, respectively.
B: effect of bumetanide on the forskolin-stimulated
Isc in jejuna. Open bars, bumetanide-sensitive
component of the forskolin-stimulated Isc in +/+
jejuna; hatched bars, bumetanide-insensitive Isc
in +/+ preparations; solid bars, bumetanide-insensitive
Isc in the / jejuna.
** P0.01 bumetanide-sensitive
Isc in KBR (+/+) vs. bumetanide-sensitive
forskolin response in 0 Cl buffer (+/+). * P 0.05 bumetanide-insensitive response / vs. +/+ in KBR.
+ P0.05 /in 0 HCO3buffer vs. /in KBR. In KBR, n= 10 for both genotypes; in 0 Cl, n = 5 and 4 for normal and NKCC1 /, respectively; in 0 HCO3
and 0 Cl/0 HCO3, n = 4 for both genotypes.
|
|
In TTX-treated preparations bathed in KBR, the magnitude of the
forskolin response was also significantly reduced in the / jejuna
(Fig. 4B). Bumetanide again inhibited the forskolin response in the TTX-treated normal preparations but had no effect on the forskolin response in the TTX-treated / jejuna (Fig.
4C). However, in the normal TTX-treated preparations a small
bumetanide-insensitive forskolin-stimulated Isc
was present (Fig. 4C) that was significantly less than the
bumetanide-insensitive response of the / tissue, a pattern similar
to the forskolin tissues without TTX pretreatment (Fig. 5B).
The response to forskolin was significantly reduced by bilateral
Cl removal in normal tissue, whereas this maneuver had no
significant effect on the magnitude of the forskolin response in the
/ jejuna (Fig. 5A). In normal tissue, this was caused by
the elimination of the bumetanide-sensitive component of the forskolin
response (Fig. 5B).
Removal of HCO3 from the medium had no effect on the
magnitude of the forskolin response in the normal jejuna, which was
100% inhibitable by bumetanide (Fig. 5). However, this maneuver
significantly reduced the magnitude of the forskolin response
(bumetanide insensitive) in the /jejuna compared with tissue
incubated in KBR (Fig. 5).
TTX-treated preparations incubated in bilateral
HCO3-free buffer exhibited the same pattern of
bioelectric responses as did non-TTX-treated preparations in this
buffer. The magnitude of the forskolin response in
HCO3-free buffer in the +/+ preparations was
unchanged from that in KBR, whereas the / jejuna exhibited a
significantly attenuated response (Fig. 4B). The normal
preparations (0 HCO3 post-TTX) exhibited a marked
bumetanide sensitivity after forskolin (virtually 100% bumetanide
sensitive), whereas the / preparations did not respond to
bumetanide (Fig. 4C).
The forskolin responses of both genotypes were significantly attenuated
in Cl-free/HCO3-free buffer compared
with the responses exhibited by the tissues in KBR (Fig.
5A). In Cl-free/HCO3-free
buffer, the entire forskolin-stimulated Isc was
bumetanide insensitive in both groups (Fig. 5B). An
additional group of / jejunal preparations was studied in
basolateral Na+-free buffer and exhibited no significant
response to forskolin (Isc = 6 ± 2.7 µA/cm2; n = 4).
Another series of experiments was conducted on jejuna from / mice
in which only basolateral Cl or HCO3
was replaced. The data from these experiments did not differ significantly from those obtained from the bilateral solution replacement studies (data not shown).
The magnitude of the forskolin response in DIDS-pretreated tissues
(serosal) did not differ significantly from tissue not pretreated with
serosal DIDS in either genotype [+/+: Isc
with no DIDS treatment 51.7 ± 8.1 µA/cm2
(n = 10) vs. Isc with DIDS
pretreatment 50.0 ± 9.7 µA/cm2 (n = 7); /: Isc with no DIDS treatment
19.2 ± 3.8 µA/cm2 (n = 10) vs.
Isc with DIDS pretreatment 19.8 ± 4.4 µA/cm2 (n = 4)]. When DIDS was applied
after forskolin and bumetanide (in KBR) in the normal preparations, the
DIDS response did not differ significantly from zero (7.2 ± 4.4 µA/cm2; n = 4). In contrast, DIDS caused
a significant inhibition in the stimulated Isc
in the / preparations (19.4 ± 4.5 µA/cm2;
n = 4). In both control and / jejuna, the forskolin
increase in Isc was insensitive to serosal DMA
(105 M) and hydrochlorothiazide (104 M)
(data not shown; both applied after forskolin).
Cecum.
In contrast to the jejuna, the ceca of the NKCC1 / mice exhibited
basal Isc that were significantly reduced
compared with those of ceca from normal mice (Fig.
6 and Table
3). Amiloride (104 M,
mucosal) had no significant effect on the magnitude of the basal
Isc in either the normal or / tissue (data
not shown). Bumetanide caused a 34 ± 2.9% inhibition of basal
Isc in the +/+ ceca, whereas in the / ceca,
bumetanide had no significant effect on basal
Isc (5.6 ± 5.6% increase in basal
Isc). The bumetanide-insensitive Isc in the +/+ tissue (55.9 ± 15 µA/cm2, n = 4) was significantly greater
than that in the / tissue (P < 0.05).
Interestingly, unlike the jejuna, in the ceca there was no significant
difference in the transepithelial conductance between the two genotypes
(Table 3).
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 6.
Net basal Isc of ceca in the
buffers indicated (see Fig. 1 legend). Open bars, normal; solid bars,
NKCC1 /. ** P0.01 vs. normal in KBR;
* P0.05 vs. same genotype in KBR. In KBR,
n = 15 and 14 for normal and NKCC1 /, respectively;
in 0 Cl and 0 HCO3, n = 5 and 4 for normal and NKCC1 /, respectively; in 0 Cl/0 HCO3, n = 4 and 5 for normal and NKCC1 /, respectively.
|
|
Incubating normal ceca in bilateral Cl-free buffer
significantly decreased the basal Isc, whereas
this maneuver did not change the basal Isc of
the / tissue (Fig. 6). The effect of bumetanide on the basal
Isc was studied only in preparations incubated
in KBR. Incubating normal ceca in bilateral HCO3-free
buffer significantly reduced basal Isc compared
with KBR (Fig. 6). This maneuver did not significantly reduce the
magnitude of Isc in the / ceca. Removing
both Cl and HCO3 bilaterally from the
medium reduced the magnitude of basal Isc in the
normal tissue compared with tissue incubated in KBR (Fig. 6). In this
buffer, the basal Isc of the / ceca was
small (only ~9.4 µA/cm2) and was significantly
different from the basal Isc in KBR.
Interestingly, the magnitude of the forskolin response in the /
ceca did not differ significantly from that of the +/+ ceca (Fig.
7A, KBR data).
However, bumetanide inhibited most of the forskolin response in the
normal ceca but had no significant effect on the response of ceca from
the / mice (Fig. 7B, KBR data); thus the
bumetanide-insensitive Isc was significantly
elevated in the / tissue (KBR) compared with normal preparations
(Fig. 7B).
View larger version (16K):
[in this window]
[in a new window]
|
Fig. 7.
A: change in Isc in
response to forskolin in ceca incubated in buffers indicated. Open
bars, normal; solid bars, NKCC1 /. ** P0.01 vs.
respective genotype in KBR. In KBR, n = 10 for both
genotypes; in 0 Cl and 0 HCO3,
n = 5 and 4 for normal and NKCC1 /, respectively;
in 0 Cl/0 HCO3, n = 4 for both genotypes. B: effect of bumetanide on the
forskolin-stimulated Isc in ceca. Open bars,
bumetanide-sensitive component of the forskolin-stimulated
Isc in +/+ jejuna; hatched bars,
bumetanide-insensitive Isc in +/+ preparations;
solid bars, bumetanide-insensitive Isc in the
/ jejuna. **P0.01 bumetanide-sensitive forskolin
response in +/+ tissue vs. bumetanide-sensitive forskolin response (=
0) in /tissue in same buffer; ++ P 0.05
vs. bumetanide-insensitive forskolin response in +/+; § P
0.01 vs. bumetanide-insensitive response in KBR for /;
* P0.05 vs. respective genotype in KBR
(bumetanide-insensitive Isc). In KBR,
n = 10 for both genotypes; in 0 Cl,
n = 5 and 4 for normal and NKCC1 /, respectively;
in 0 HCO3, n = 4 for both genotypes;
in 0 Cl/0 HCO3, n = 4 for both genotypes.
|
|
Ion-substitution studies were again carried out in an attempt to
further characterize the response to forskolin in these tissues. Bilateral Cl-free buffer significantly reduced the
magnitude of the forskolin response in tissues of both genotypes
compared with the respective tissues incubated in KBR (Fig.
7A). Neither tissue responded to bumetanide (postforskolin)
in Cl-free buffer (Fig. 7B); therefore, the
magnitude of the bumetanide-insensitive forskolin responses was nearly
identical in this buffer (Fig. 7B).
The forskolin response was similar to control (KBR) in normal ceca in
bilateral HCO3-free buffer, and the ratio of the
bumetanide-sensitive vs. bumetanide-insensitive response was similar to
that of the +/+ tissues in KBR (Fig. 7). However, in the / ceca,
the magnitude of the forskolin response was significantly reduced in
bilateral HCO3-free buffer compared with KBR (Fig.
7A). The / ceca incubated in bilateral
HCO3 buffer again did not respond to bumetanide (Fig.
7B). The bumetanide-insensitive Isc
was significantly greater in the / preparations compared with the
bumetanide-insensitive Isc of the normal
preparation incubated in the buffer. The effect of bilateral
HCO3-free buffer on the magnitude of the forskolin
response in TTX-treated ceca was also investigated. The responses of
these preparations were similar to those of ceca not pretreated with
TTX [forskolin response, HCO3 free, TTX treated
+/+Isc, 137.4 ± 17 µA/cm2
(n = 4) vs. /, 31.9 ± 2.1 µA/cm2 (n = 4); p0.001;
compare with data in Fig. 7A].
When both Cl and HCO3 were removed
bilaterally from the medium, the forskolin response was very small and
did not differ between the two genotypes (Fig. 7A). Neither
tissue responded to bumetanide when incubated in this buffer (Fig.
7B).
In some preparations (KBR) DIDS was added after forskolin/bumetanide.
DIDS caused a significant decrease in the magnitude of the
forskolin-stimulated Isc in the / ceca only
(30.6 ± 7.9 µA/cm2; n = 4;
P < 0.05).
Removal of both HCO3 and Na+ from the
basolateral side of the / ceca (HCO3 was also
removed from the mucosal side to ensure that no HCO3
was transported across the tissue) caused no further attenuation in the
forskolin response compared with HCO3 removal alone
(Fig. 8). The forskolin response of the
/ ceca in this buffer was not sensitive to DIDS (data not shown).
However, when Cl was also removed from the basolateral
side of the tissue (in addition to the above ions), the forskolin
response was significantly decreased (Fig. 8).
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 8.
Isc response to forskolin in NKCC1
/ ceca studied in the buffers indicated. * P0.05
vs. preparations incubated in HCO3and
Na+-free buffer; n = 4 for each group.
Buffers are bilateral (m/s) or serosal (s) only.
|
|
| |
DISCUSSION |
We found that NKCC1 (encoded for by the Slc12a2 gene)
mRNA expression was very prominent in the salivary gland and the more distal regions of the intestinal tract (Fig. 1). The
Na+-K+-2Cl cotransporter is
located in the basolateral membrane of many secretory epithelial cell
types. Although the cotransporter is thought to play a role in basal
and stimulated Cl secretion by transporting
Cl across this barrier (Fig.
9), the basolateral (and apical)
membranes of many types of epithelial cells also may contain
Cl/HCO3 exchangers (anion exchanger;
AE) that can transport Cl in exchange for
HCO3 (Fig. 9).
Na+-HCO3 (NBC)-coupled entry has also
been identified on the basolateral membrane of some intestinal
epithelia, which can mediate a HCO3 contribution to
intestinal anion secretion.
View larger version (23K):
[in this window]
[in a new window]
|
Fig. 9.
Model of potential basolateral entry paths in an
intestinal epithelial cell. As both crypt and villus cells can secrete,
this model could represent either cell type. CA, carbonic anhydrase;
AE, anion exchanger; NHE, Na+/H+ exchanger;
NBC, Na+-HCO3 cotransporter; CFTR, cystic
fibrosis transmembrane conductance regulator; DMA, dimethylamiloride.
|
|
The NKCC1-deficient mutant mice do not die of intestinal obstruction
and rupture either in the early prenatal period or as adult mice
(20). They thus differ from CF mice, which have a mutation
in the CFTR gene that mediates apical membrane Cl
translocation, and typically exhibit severe intestinal pathophysiology resulting in both neonatal and adult mortality (3, 28). A study was recently published on another NKCC1-deficient mouse model
consistent with our finding that there was no increased death rate in
neonatal NKCC1 / pups (5). However, unlike our study,
Flagella et al. (5) reported a relatively high death rate
(28%) of the NKCC1 / mice in the periweaning period. Although no
definite cause of death was given, intestinal bleeding was noted in
some pups (5). Whereas CF mice, which are deficient in
apical membrane Cl channel activity, exhibited dilated
mucus-filled crypts and intestinal goblet cell hypertrophy and
hyperplasia (8), the NKCC1 / mice exhibited only very
mild intestinal histopathology in the form of scattered dilated crypts
that did not appear to contain mucus (Fig. 1, D and
E). In CF mice, the intestinal morbidity and high rate of
prenatal mortality appear to result from the inability of the
intestinal tract to secrete either Cl or
HCO3 in the basal or stimulated state (2, 7,
12, 24). Because our NKCC1 / mice exhibit little intestinal
histopathology and no obvious morbidity or mortality related to
intestinal problems, it seems likely that the intestinal tracts of
these animals are capable of sufficient liquid secretion to circumvent
the intestinal problems that characterize the CF mice. We therefore
characterized the anion transport paths that likely mediate liquid
secretion in these animals.
Jejunum.
Flux studies as well as ion-replacement studies (7, 25)
suggest that a large part of the basal Isc in
the normal mouse jejunum reflects Cl secretion or is
Cl dependent (7, 25). In our study, the
basal Isc was insensitive to bumetanide in both
the +/+ and / jejuna and was of nearly identical magnitude in these
genotypes. Blocking neurotransmitter release from the enteric nervous
system (TTX-treated tissue) did not appear to alter this relationship.
However, Cl is clearly involved in the maintenance of the
basal Isc, because removal of Cl
from the bathing medium significantly reduced the basal
Isc similarly in both genotypes.
It is well established (2, 7, 15, 24, 25) that the normal
murine intestine is capable of substantial electrogenic HCO3 secretion under both basal and stimulated
conditions. Studies have identified an NBC coupled cotransport entry
path at the basolateral membrane of enterocytes of several species
(Refs. 1, 17; Fig. 9). The NBC has been found to be electrogenic with
the usual stoichiometry being 1:2 or 1:3 (23). In
addition, the NBCs (there may be a number of isoforms, see Ref. 23 for
details) are Na+ dependent, HCO3
dependent, and usually blocked by stilbenes such as DIDS
(23). It should be pointed out that some studies have
failed to detect this basolateral cotransporter in intestine
(10).
HCO3 removal significantly decreased the magnitude of
the bumetanide-insensitive Isc in both
genotypes. In addition, in the +/+ jejuna, this protocol significantly
increased the bumetanide-sensitive component of the basal
Isc, leaving the net Isc
unchanged. This observation lends support to previous reports that
removal of one of the anions (HCO3) from the buffer
increased the secretion of the other (Cl) (17,
25). In both +/+ and / jejuna, one could speculate that in
KBR at least part of the basal bumetanide insensitive Isc may be mediated by NBC basolateral entry. In
the NKCC1 / tissue, the criteria for assigning NBC a role in basal
transport appear to be met in that the basal Isc
was electrogenic, Na+ dependent (see Table 2),
HCO3 dependent, and at least partially sensitive to
DIDS. In the / jejuna, DIDS completely eliminated the spontaneous
oscillations in the basal Isc as well as
decreasing the magnitude of the basal Isc. An
inhibitory effect of DIDS on the NBC and or AE would be compatible with
this observation. If the bumetanide-insensitive Isc in either genotype is due at least in part
to HCO3 secretion, then why is the
Isc diminished by Cl removal? An
interesting study on isolated colonic crypts reported that basolateral
Cl removal significantly attenuated the magnitude of the
stimulated rate of HCO3 secretion in these
preparations (6). In pancreatic cells, it has been
suggested that Cl secretion maintains the driving force
for basolateral NBC entry by depolarization of the membrane
(26). If this mechanism is operating in the jejuna under
the conditions of our study (either or both +/+ and /), basolateral
Cl entry would have to be via a non-NKCC1 pathway (see
Fig. 9). Another possibility is that a parallel operation of a
basolateral NBC and AE may been involved. In this scenario,
HCO3 taken up via the NBC would be recycled across
the basolateral membrane via the AE. Thus net electrogenic
Cl secretion would result. Either or both of these
possibilities would explain the need for all three ions
(Na+, Cl, and HCO3; Table
2).
Interestingly, the transepithelial conductance of the NKCC1 /
jejuna was significantly greater than that of the normal preparations in every buffer studied. The reason for this result is not known. As
>80% of the transepithelial conductance of intestinal epithelia has
been attributed to the paracellular pathway, the increased conductance
most likely reflects a change in the conductance of this pathway (see
Ref. 21 for discussion). It has been shown that when intestinal cells
are stimulated to secrete, there is up to a 50% decrease in
conductance (21). It has been suggested that this change
reflects cell swelling, which in turn reduces the width of the
intracellular space, thus decreasing conductance (21). If
NKCC1 played a role in cell volume and the volume of the NKCC1 /
cells was decreased compared with normal (unstimulated) cells, this
difference theoretically might increase the transepithelial conductance. In our normal intestine, bumetanide (before forskolin) had
no significant effect on the transepithelial conductance (data not
shown). Therefore, at least on an acute basis, blocking the NKCC1
pathway in normal jejuna does not increase transepithelial conductance.
Further study will thus be necessary to explain the difference in the
transepithelial conductance across the jejuna of the two genotypes.
Removing both Cl and HCO3 from the
medium did not further diminish the magnitude of the basal
Isc in / jejuna compared with the
Isc when either anion was present. The origin of
the Isc in bilateral
Cl/HCO3-free medium in the +/+ and
/ jejuna is not known. It would not be expected to be
amiloride-sensitive Na+ absorption because this transporter
is not present in jejunal tissue. Also, because there was no apical
glucose present, electrogenic Na+-glucose cotransport could
not explain the basal Isc.
In the jejunum of the normal mouse, forskolin induced an increase in
Isc that was Cl dependent. The
inhibition of the forskolin response by bumetanide, the marked decrease
in the magnitude of this response when tissues were incubated in
bilateral Cl-free buffer, and the failure of bilateral
HCO3-free buffer to change the magnitude of this
response support the speculation that the forskolin response reflects
Cl secretion. Therefore, NKCC1 appears to play a major
role in the forskolin-stimulated response in this tissue.
It is interesting to note that there is a significant
bumetanide-insensitive response to forskolin in the normal preparations in 0 Cl buffer (12.4 ± 4.2 µA/cm2)
and that this response did not differ in magnitude from the /
jejuna in the same buffer. It is likely that the same basolateral entry
mechanism supports the bumetanide-insensitive forskolin response in 0 Cl buffer in normal and / jejuna.
The forskolin response was significantly reduced in the jejuna of the
NKCC1 / mutant mice and was bumetanide insensitive. The /
jejunal forskolin response in TTX-treated tissues was also
significantly reduced compared with normal TTX-treated tissue (KB
TOP
ABSTRACT
INTRODUCTION
