Glitazones regulate glutamine metabolism by inducing a
cellular acidosis in MDCK cells
Greg
Coates1,
Itzhak
Nissim2,
Harold
Battarbee1, and
Tomas
Welbourne1
1 Departments of Molecular and Cellular Physiology,
Louisiana State University Health Science Center, Shreveport,
Louisiana 71130; and 2 Division of Child Development
and Rehabilitation, Department of Pediatrics, University of
Pennsylvania School of Medicine, Philadelphia, Pennsylvania 19104
 |
ABSTRACT |
We studied the effect of the
antihyperglycemic glitazones, ciglitazone, troglitazone, and
rosiglitazone, on glutamine metabolism in renal tubule-derived
Madin-Darby canine kidney (MDCK) cells. Troglitazone (25 µM) enhanced
glucose uptake and lactate production by 108 and 92% (both
P < 0.001). Glutamine utilization was not inhibited,
but alanine formation decreased and ammonium formation increased (both
P < 0.005). The decrease in net alanine formation occurred with a change in alanine aminotransferase (ALT)
reactants, from close to equilibrium to away from equilibrium,
consistent with inhibition of ALT activity. A shift of glutamine's
amino nitrogen from alanine into ammonium was confirmed by using
L-[2-15N]glutamine and measuring the
[15N]alanine and [15N]ammonium production.
The glitazone-induced shift from alanine to ammonium in glutamate
metabolism was dose dependent, with troglitazone being twofold more
potent than rosiglitazone and ciglitazone. All three glitazones induced
a spontaneous cellular acidosis, reflecting impaired acid extrusion in
responding to both an exogenous (NH
) and an
endogenous (lactic acid) load. Our findings are consistent with
glitazones inducing a spontaneous cellular acidosis associated with a
shift in glutamine amino nitrogen metabolism from predominantly
anabolic into a catabolic pathway.
rosiglitazone; troglitazone; ciglitazone; L-[2-15N]glutamine; 2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein; glutamate; alanine; ammonium; Madin-Darby canine kidney cells; intracellular pH
 |
INTRODUCTION |
THE
GLITAZONES troglitazone, ciglitazone, and rosiglitazone are
proxisome proliferator-activated receptor-
(PPAR
) agonists (9) that exhibit antihyperglycemic activity
(11, 13, 18), halt mesangium expansion in experimental
models of type 2 diabetes mellitus (20, 26), and act
directly on mesangial cells to limit extracellular matrix formation
(33). In cultured mesangial cells, troglitazone inhibits
glutamine utilization (34) and, specifically, glutamate transamination and alanine formation (Fig. 1, R2). In addition to the decreased
alanine formation, there is an increased ammonium formation without
increased glutamine breakdown (Fig. 1, R1), consistent with a
shift of the glutamate flux from the transamination (Fig. 1, R2)
into the deamination pathway (Fig. 1, R3). As shown in Fig. 1, one
way for troglitazone to affect increased deamination (R3) at the
expense of transamination would be to shift glutamate from the
transamination pathway (R2) into the mitochondrion, where the
deamination pathway, glutamate dehydrogenase (GDH; R3) is located.
If the transamination pathway is predominantly cytosolic, then this
response would reflex a shift of glutamate from the cytosol into the
mitochondrial compartment. Thus an important point to establish from
the perspective of this model is the intracellular localization of the
alanine aminotransferase. Furthermore, if the functional glutaminase
activity is indeed within the cytosolic compartment (23,
32), glutamate formed within the cytosol would be effectively
disrupted from subserving functions localized to this compartment,
including matrix synthesis, and be directed into the mitochondrial
pathway. Noteworthy is that not all investigators view the
mitochondrial glutaminase as being functional within the cytosolic
compartment and instead assign it to the mitochondrial matrix
(21) or matrix surface of the inner mitochondrial membrane
(12). Kvamme et al. (23), on the other hand,
found the matrix space glutaminase inactive and the inner membrane
glutaminase population to be functionally active on the cytosolic
surface. Because assayable GDH activity remained unchanged in our
previous studies (34, 39) whereas assayable alanine
aminotransferase (ALT) activity decreased, this shift in glutamate
metabolism is likely to result from either a decreased cytosolic
transamination flux or an increased transport of glutamate from the
cytosol and into the mitochondrial compartment or both. The transport
of glutamate coupled to deamination is pH dependent; at reduced
cytosolic pH, the uptake of cytosolic glutamate and flux through the
GDH pathway accelerate (35). On the basis of the view that
alanine formation occurs predominantly within the cytosol
(14), a shift of glutamate into the mitochondria points to
a developing cellular (cytosolic) acidosis after the administration of
troglitazone.

View larger version (13K):
[in this window]
[in a new window]
|
Fig. 1.
Putative glutamine and glucose metabolism in relation to
intracellular acidosis in Madin-Darby canine kidney (MDCK) cells.
Glucose (Gluc) uptake is shown coupled to lactic acid production, with
lactate (Lac ) released to the medium and H+
pumped out by the Na+/H+ exchanger (NHE).
Glutamine (Gln) is converted to ammonium (NH ) and
glutamate (Glu ) by mitochondrial glutaminase (R1)
shown present on the outer surface of inner membrane (22,
30). Note that glutaminase may also be functional within the
matrix space (10, 19), in which case the glutamate flux
would be reversed. Cytosolic conversion of glutamate to alanine (ALA)
is catalyzed by alanine aminotransferase (R2) and may be inhibited
by acidosis. Transport of glutamate into the mitochondrial glutamate
dehydrogenase (GDH) pathway (R3) and ammonium formation are
facilitated by acidosis. Glitazones may shift glutamate from alanine to
ammonium by producing a cellular acidosis, either by stimulating acid
production or inhibiting acid extrusion.
|
|
Cell intracellular pH (pHi) depends on the balance
between metabolic acid production, largely lactic acid production, and acid extrusion mediated by the sodium/hydrogen ion exchanger (NHE) activity (4, 8). Glitazones such as troglitazone may
affect cellular acidosis either by increasing the acid production or by
decreasing acid extrusion and thereby affecting a shift in amino
nitrogen metabolism. In this regard, it has been shown that troglitazone enhances glucose uptake (13, 18, 19);
however, unlike the antihyperglycemic effect of insulin, glitazones may increase lactic acid production, at least in vitro (13, 18, 19), providing an increased endogenous acid load. If this rate of acid loading exceeds the acid extrusion rate, cytosolic pH would
decline, shifting glutamate into the mitochondria, and cause either
ammonium formation (27, 29, 34) or alanine
(29) depending on the distribution of the alanine
aminotransferase. On the other hand, if glitazones were to
inhibit acid extrusion, then metabolically retained acid would
spontaneously lower the intracellular pH. In either case, a developing
cellular acidosis would have physiological consequences for glutamine
metabolism and dependent cellular processes (39).
Cytosolic glutamate represents a pivotal point in the metabolism of
glutamine as depicted in Fig. 1. If glutamate undergoes transamination
or conversion to proline, largely cytosolic reactions (2,
25), then the amino nitrogen as well as carbon skeleton may be
distributed into amino acids, which serve as building blocks for
protein synthesis and particularly collagens, as in the case of
glutamate conversion to proline (2, 24). On the other hand, if glutamate is deaminated by GDH, a mitochondrial matrix space
activity (35), then the amino nitrogen is released as ammonium, and the carbon skeleton,
-ketoglutarate, is oxidized and
therefore would not contribute to biosynthetic processes. From this
perspective, our purpose was to determine whether glitazones effect
this putative shift in glutamine metabolism in a well-established epithelial cell line, Madin-Darby canine kidney (MDCK) cells, which
actively produces basement membrane components, e.g., collagens (16). Our second goal was to determine whether a shift of
glutamate from transamination into deamination was associated with a
cellular acidosis, and, if so, whether this was attributable to
impaired acid extrusion or increased acid production.
 |
METHODS |
MDCK cells obtained from American Type Culture Collection (ATCC
catalog no. CCL-34; Manassas, VA) were grown to confluence in T150
flasks in DMEM plus 10% fetal calf serum containing (in mM) 28 sodium
bicarbonate, 10 sodium pyruvate, 5 D-glucose, and 2 L-glutamine at 37°C and 5% CO2 (pH 7.4).
Confluent cells were subcultured by detachment with the use of
trypsin-EDTA (GIBCO-BRL, Rockville, MD) and reseeded onto 6-well
culture plates (Corning Cell Wells, Corning, NY) for metabolic studies
or in specially designed 30-mm chambers (Bioptechs; Biological Optical
Technologies, Butler, PA) equipped with a heating element and cap port
for O2-to-CO2 aeration. The chambers were
placed uncapped inside a 60-mm covered tissue culture dish and
incubated at 37°C and 5% CO2. The cells were allowed to
gain confluence, usually 3-4 days for 6-well plates and 2 days for
cells in chambers.
Metabolic studies.
Studies were performed on confluent MDCK cells grown in the 6-well
plates over 16 h in DMEM containing DMSO (vehicle) or DMEM plus
troglitazone (kindly supplied by Dr. Tagata, Sankyo, Tokyo, Japan) or
ciglitazone (Cayman Chemical, Ann Arbor, MI). Media samples were
promptly treated with an equal volume of ice-cold 5% trichloroacetic
acid, left on ice for 10 min, and then centrifuged at 10,000 g for 10 min. Aliquots of the protein-cleared supernatant containing free amino acids were then treated with
O-phthalaldehyde (OPA; FLUKA, Buchs, Switzerland) for
precolumn derivatization and injected onto a C18 4.6 × 250-mm column (Microsorb; Varian, Walnut Creek, CA) for separation
of the derivatized amino acids. The column effluent was passed through
a fluorescence detector with peaks for the major amino acids at
characteristic retention times for standards (in min): glutamate, 11.2;
glutamine, 16.4; homoserine (internal standard), 17.8; and alanine,
20.5 (34). The medium concentration of the
glutamate, glutamine, and alanine was obtained from the peak areas
divided by the peak areas for their respective standards. Utilization
or production rates for the respective amino acids were obtained from
the concentration differences multiplied by the medium volume of 2 ml.
Ammonium concentration was determined by the microdiffusion method
(27), and formation rate was determined as above, by
subtracting the medium blank, and expressed on the basis of milligrams
protein. Glucose and lactate medium concentrations were measured
enzymatically (Sigma, St. Louis, MO). Utilization of glucose and
production of lactate were calculated as for the amino acids and
expressed on the basis of milligrams protein.
For analysis of alanine and ammonium formed from
L-[2-15N]glutamine, medium glutamine was
replaced with L-[2-15N]glutamine (99 atom % excess; Cambridge Isotope Laboratories, Andover, MA). After
16 h of incubation in the prescribed media above, medium samples
were taken and treated with ice-cold 40% perchloric acid. The
concentration of ammonium and alanine and their 15N
enrichment were determined on the neutralized supernatants. Briefly,
the amino acids underwent precolumn derivatization with OPA and
separation by HPLC and fluorescence detection as described above.
Analysis of 15N in the amino acids was done by GC-MS, as
previously described (6, 29). Formation of
15NH
was determined after conversion of
ammonium to norvaline (6). To calculate the conversion of [15N]glutamine to ammonium and alanine, the isotopic
enrichment (atom % excess) of 15N in the particular
metabolite was multiplied with the amount present and expressed as
nanomoles per milligram of protein. Ammonium concentration was measured
by the previously described microdiffusion method (27).
The distribution of alanine aminotransferase between the cytosolic and
mitochondrial compartments was estimated as previously described for
liver homogenates (14) in homogenates obtained from
confluent monolayers of control cells. Briefly, the cells were scraped
into 1 ml of a mannitol-sucrose buffer (0.225 M mannitol, 0.075 M
sucrose, 0.01 M Tris, pH 7.8, and 0.05 M EDTA) and gently homogenized
on a motor-driven Potter-Elvejhem homogenizer using 10 passes of a
loose-fitting pestle. This homogenate was then centrifuged at 14,000 g for 45 min at 10°C with the resulting supernatant
and pellet fractions separated; the pellet was resuspended in 1 ml of a
KHPO4-glycerol buffer (0.05 M KPO4, pH 7.8, 0.025 M alanine, 0.005 M cysteine, and 50% glycerol), and both
fractions were assayed for ALT and LDH activity under optimal
conditions. The distribution of ALT between the particulate and soluble
fractions was corrected for by the fraction of the cytosolic lactate
dehydrogenase (LDH) present in the particulate fraction. Monolayer
contents of pyruvate and
-ketoglutarate were determined on TCA
extracts after neutralization to pH 7.4 using 5 M
K2CO3. Determinations were performed by
modifications of the enzymatic analyses, using LDH or GDH, and coupled
to the oxidation of NADH + H+ monitored at 340 nm as otherwise described (3, 7). Recoveries for pyruvate
and
-ketoglutarate (25 nmol) standards were 91 ± 8 (mean ± SD, n = 4) and 101 ± 13% (mean ± SD,
n = 6), respectively.
Cell pHi measurements.
After confluence, chambers were transferred to the stage of an Olympus
IMT-2 microscope equipped with a heated stage insert (model no.
TIS04201501; Kent Scientific). Fluorescent measurements were made at
37°C through an inverted epifluorescence microscope with a UV-F ×40
objective with use of a Photon Technology International (Brunswick, NJ)
RM-D microspectrofluorometer outfitted for photometric ratio
fluorescence studies. After an autofluorescence measurement made in
Krebs-Henseleit medium containing 10 mM D-glucose, the cells were loaded with the pH-sensitive fluorescent dye, the
acetoxymethyl ester of
2',7'-bis(2-carboxyethyl)-5(6)-carboxyfluorescein
(BCECF-AM, 5 mM in DMSO stock; Molecular Probes, Eugene, OR), dissolved
in Krebs-Henseleit medium to 5 µM and added to the chambers for 25 min at 37°C. The chamber was then washed three times with
Krebs-Henseleit medium, and fluorescence measurements were obtained
after equilibration of the cells in Krebs-Henseleit medium with HEPES
substituted for bicarbonate (KHHEPES). Light emitted from a 75-W xenon
arc lamp alternately exposed cells to wavelengths of 490 and 440 nm. Excitation wavelengths were chopped (10 Hz) at 490 and 440 nm, and
emissions from a minimum of three to four aggregated cells were
monitored at 535 ± 25 nm by use of a low-pass optical filter. Instrument components and data acquisition and analysis were computer controlled using fluorescence software (FeLiX; Photon Technology International, Brunswick, NJ). Changes in the emission ratio (490/440 nm) were taken as an index of changes in the intracellular pH. The
recording periods at the various data acquisition intervals were
minimized to avoid BCECF photobleaching. Only preparations with 20-fold
greater fluorescence intensity than that of the autofluorescence were
used. The high K+-nigericin technique (38) was
used to clamp the intracellular pH to medium standards of known pH
(confirmed on a Corning 240 pH meter at 37°C after withdrawal of the
sample from the chamber), obtaining a pH calibration of the 490-to-440
signal ratio (r2 =0.99 for 490/440 ratio vs.
pHi; n = 28).
To monitor the spontaneous pHi response to glitazones, the
medium was replaced with fresh Krebs-Henseleit medium with 24 mM HEPES
buffer replacing bicarbonate, and the 490-to-440 signal ratio was
followed continuously for 8 min. The medium was exchanged a
second time with the above KHHEPES containing 25 µM troglitazone, followed by another 8 min of continuous recording. The control and
glitazone media were collected and analyzed for lactate concentration as above. Comparisons were then made between control and
troglitazone-treatment differences in pHi and acid
production with the use of the Student t-test. To test for
acid extrusion capability, the cells were incubated in HEPES-buffered
Krebs-Henseleit medium (pH 7.40) and then acid loaded with a 4-min
exposure to Krebs-Henseleit medium in which 20 mM NaCl had been
replaced with 20 mM NH4Cl. After returning the
Krebs-Henseleit medium, we continuously monitored the pHi
response for 4 min. The recovery response was taken as the change in
pHi per time interval (
pHi/min). Time
control experiments for a repetitive acid load were also performed,
establishing that the recovery rates were not different for the loading
periods. Differences within the 4-min recovery period between control
and glitazone
pHi/min responses were determined by use
of the paired Student's t-test. For repeated measurements,
differences between control and test groups were assessed with the use
of ANOVA and a corrected t-test (Dunnett's).
 |
RESULTS |
Table 1 shows the effect of 25 µM
troglitazone on glutamine and glucose metabolism. A hallmark of the
therapeutic effectiveness of glitazones is their glucose-lowering
action (11, 13), largely the result of increased
utilization (13, 18, 20). In MDCK cells, 25 µM
troglitazone increased glucose uptake by 108% (P < 0.0001) and lactate formation by 92% (P < 0.0005).
Despite this large increase in glucose uptake, there was not a
reciprocal decline in glutamine utilization (1,091 ± 199 vs.
811 ± 196 nmol/mg protein for control and troglitazone treatment,
respectively), as might be expected if these substrates were competing
as oxidative fuel (31). Despite the maintained glutamine
uptake, the products formed from subsequent glutamate metabolism were
markedly altered. In the presence of troglitazone, ammonium formation
increased by 43% (1,614 ± 60 vs. 1,130 ± 49 nmol/mg
protein; P < 0.0001); this large increase occurs
without an increase in glutamine utilization (Fig. 1, R1),
consistent with an increased glutamate flux via the deamination pathway
(Fig. 1, R3). Note that alanine production decreased also by 43%
(557 ± 90 vs. 977 ± 125 nmol/mg protein; P < 0.02). The ratio of total ammonium to alanine produced rises from
1.27 ± 0.20 to 3.40 ± 0.68 (P < 0.01) as a
result of both the increase in ammonium formed and the decrease in
alanine produced.
Because the ALT catalyzes a near-equilibrium reaction (14,
36), we measured the cellular reactants after 4 h in
control and troglitazone-treated monolayers to assess whether this
reaction remains close to equilibrium. As shown in Table
2, the ALT reaction was close to
equilibrium in the control monolayers, given the equilibrium constant
of 1.6 (36) and a measured mass action ratio of 1.85 ± 0.36. In monolayers exposed to 25 µM troglitazone for 4 h,
the mass action ratio fell 59% to 0.77 ± 0.06 (P < 0.06), indicating that the reaction has moved away from equilibrium, consistent with reduced ALT activity. This movement away from the
near-equilibrium condition of the controls reflected a 66% decrease in
cellular alanine content (from 45 ± 9 to 15 ± 2 nmol/mg protein; P < 0.02). In line with the decrease in
cellular alanine content, the net alanine flux measured for 4 h
slowed by 76% (from 397 ± 108 to 96 ± 12 nmol · 4 h
1 · mg protein
1; P < 0.03). The
-ketoglutarate content did not change, but there was a
small increase (P < 0.05) in pyruvate content (from
13 ± 1 to 16 ± 1 nmol/mg protein). The other substrate,
glutamate, decreased, whereas ammonium production increased 1.8-fold
(from 235 ± 12 to 412 ± 22 nmol · 4 h
1 · mg protein
1; P < 0.002), consistent with a shift of glutamate metabolism from the ALT
into the GDH pathway (Fig. 1). We also measured the assayable ALT
activity and the soluble vs. particulate distribution within these
cells. Although troglitazone exerts a dose-dependent inhibition of both
assayable ALT activity and net alanine flux in the proximal tubule-like
cell line LLC-PK1-F+ (39), there
was no reduction in assayable ALT in this cell line (21 ± 3 vs.
22 ± 2 U/mg protein for control and troglitazone treatment,
respectively; n = 7 pairs). The subcellular
distribution of ALT was predominantly within the soluble fraction
(59%; Table 3), and when corrected for
the activity trapped within the particular fraction, amounts to almost
76% associated with the cytosolic compartment. This distribution is
similar to that observed within enteroctyes (25). Although
the decrease in alanine production indicates a decreased flux through
the ALT, the increase in ammonium produced could represent the
contribution of both the amide and the amino nitrogen of glutamine and,
therefore, by itself would not necessarily indicate an increased
deamination of glutamate.
To confirm that the rise in the ammonium-to-alanine production ratio
does indeed reflect the putative shift in the utilization of the amino
nitrogen of glutamine from the transamination into the deamination
pathway as shown in Fig. 1, we measured the formation of
[15N]ammonium and [15N]alanine formed from
L-[2-15N]glutamine. With this tracer
approach, the formation of labeled ammonium quantitates the deamination
flux and also localizes it to the mitochondrial matrix space. As shown
in Fig. 2, 25 µM troglitazone increases
the ammonium formation from the amino nitrogen of glutamine by
2.74-fold (from 179 ± 3 to 491 ± 18 nmol/mg protein;
P < 0.0001), confirming that the increase in ammonium
formation shown in the balance study above is largely (312/484 × 100 = 65%) derived from the deamination of labeled glutamate. Alanine
formation from the amino nitrogen of glutamine decreased
(P < 0.001) from 491 ± 21 to 118 ± 12 nmol/mg, accounting for 93% of the fall in alanine formed. As a
consequence, there is a rise in the ratio of ammonium to alanine formed
from the amino nitrogen of glutamine from 0.36 ± 0.03 to
4.26 ± 0.48 (P < 0.001).

View larger version (10K):
[in this window]
[in a new window]
|
Fig. 2.
Proof that troglitazone shifts the amino nitrogen of
glutamine from transamination to deamination pathway. Troglitazone (30 µM) increases [15N]ammonia (NH )
formation and decreases [15N]alanine (ALA) formation from
L- [2-15N]glutamine. Results are means ± SE (in nmol/mg protein) from 3 monolayers/group incubated for
16 h.
|
|
This shift in glutamate into the deamination pathway at the expense of
the transamination pathway is dose dependent, as shown in Fig.
3A for troglitazone, Fig.
3B for ciglitazone, and Fig. 3C for
rosiglitazone. Both troglitazone and ciglitazone produced a decrease
(P < 0.05) in alanine production at 25, 50, and 100 µM (Fig. 3, A and B), whereas rosiglitazone
(Fig. 3C) lowered alanine only at 100 µM. In contrast, all
three glitazones increased ammonium production over the concentration
range of 10-100 µM. The ratio of ammonium produced to alanine
produced increases in the ciglitazone-treated monolayers from 0.56 ± 0.08 to 0.96 ± 0.28, 0.72 ± 0.10, 0.83 ± 0.17, 1.36 ± 0.34, and 2.23 ± 0.18 at 5, 10, 25, 50, and 100 µM, respectively (P < 0.05 at 10, 25, 50, and 100 µM). For rosiglitazone, the ratios increased from 0.70 ± 0.05 to 0.84 ± 0.05, 0.99 ± 0.07, 1.20 ± 0.10, 2.16 ± 0.31, and 4.66 ± 0.6, respectively (P < 0.05 at 10, 25, 50, and 100 µM). In the troglitazone-treated monolayers,
this ratio increases from 0.53 ± 0.02 to 0.72 ± 0.11, 0.98 ± 0.13, 2.31 ± 0.71, 2.80 ± 0.62, and 4.02 ± 0.17 at 5, 10, 25, 50, and 100 µM, respectively (P < 0.05 at 10, 25, 50, and 100 µM). Although all three glitazones increase this ratio, troglitazone appears more potent than ciglitazone, whereas rosiglitazone appears almost as potent as troglitazone at 25 µM and above.

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 3.
A: dose response for ammonium and alanine
production to troglitazone. B: dose response of ammonium and
alanine production to ciglitazone. C: dose response of
ammonium and alanine to rosiglitazone. A-C: results are
means ± SE from 4 wells/dose. * Different from control
(P < 0.05) by ANOVA and Dunnett's corrected
t-test.
|
|
One testable mechanism by which glitazones might shift glutamate from
the transamination into the deamination pathway would be to lower the
cytosolic pH and thereby accelerate glutamate transport across the
mitochondrial inner membrane and into the matrix space (Fig. 1). To
determine whether troglitazone decreases cytosolic pH, MDCK cells were
loaded with BCECF, and the spontaneous pHi was monitored
over 8 min after addition of KHHEPES or KHHEPES containing 25 µM
troglitazone. A representative recording is shown in Fig.
4A, with
results from 20 experiments presented in Table 4. The steady-state
pHi for MDCK cells did not change with the addition of the
KHHEPES (
pHi/min = 0.0006 ± 0.001, measured
over the first 4 min) and averages 7.13 ± 0.10 (measured at 8 min). In marked contrast, addition of KHHEPES plus 25 µM troglitazone produced a prompt decline in the spontaneous pHi
(
0.08 ± 0.01; P < 0.0001 vs. KHHEPES alone
over first 4 min). The resting pHi after 8 min of exposure
to troglitazone drops to 6.51 ± 0.05 (P < 0.0001 vs. control). These results show that troglitazone induces a
spontaneous cellular acidosis. Note that lactic acid production did not
acutely increase (21 ± 3 and 23 ± 3 nmol/min; Table 4), indicating that this large rise in intracellular hydrogen ion concentration was not the result of an increased endogenous acid production. To determine whether ciglitazone also induces a spontaneous cellular acidosis, MDCK cells were exposed to KHHEPES and to KHHEPES plus 25 µM ciglitazone, and the spontaneous pHi was
continuously monitored over the following 8 min. As shown in the
representative experiment in Fig. 4B, pHi
decreased with 25 µM ciglitazone in contrast to a steady or slightly
rising pHi for the KHHEPES alone; results from four
additional experiments show that ciglitazone acidifies the cell
compared with control (6.85 ± 0.05 vs. 7.13 ± 0.10;
P < 0.05) at a rate of pHi decline that is
less than that exerted by troglitazone (
0.04 ± 0.02 vs.
0.08 ± 0.01
pHi/min). Figure 4C shows
that rosiglitazone at 25 µM also produces an acute cellular acidosis,
with the steady-state pHi falling from 7.29 ± 0.07 to
6.76 ± 0.06 (P < 0.01, n = 4).
Therefore, all three glitazones induce a spontaneous cellular acidosis
in the MDCK cell line, but with the drop in pHi being about
twice as great for troglitazone (0.62 ± 0.06) and rosiglitazone
(0.53 ± 0.02) as for ciglitazone (0.28 ± 0.04).

View larger version (13K):
[in this window]
[in a new window]
|
Fig. 4.
A: representative recording of MDCK cell
intracellular pH (pHi) in response to KHHEPES medium
(control) and to KHHEPES plus 25 µM troglitazone (see
METHODS for KHHEPES medium details). Calibration standards
are high K+-nigericin (see METHODS).
B: representative recording of MDCK cell pHi in
response to KHHEPES medium (control) and to KHHEPES plus 25 µM
ciglitazone. Calibration is as in A. C:
representative response of MDCK cell pHi in response to
KHHEPES medium (control) and to KHHEPES plus 25 µM rosiglitazone.
Calibration is as in A.
|
|
To assess the duration of the intracellular acidosis elicited by
troglitazone, monolayers were incubated in either KHHEPES (Fig.
5A) or DMEM (Fig.
5B) and aerated with 5% CO2 in the presence of
vehicle or 25 µM troglitazone. As shown for the KHHEPES medium (Fig.
5A), the acidosis induced by troglitazone was maintained for
at least 45 min. Because the metabolic studies were performed in
HCO
-CO2-buffered DMEM, we wished to
determine whether an acidosis would occur in this medium as well. As
shown in Fig. 5B, troglitazone elicits an intracellular acidosis that is maintained for at least 20 min. Taken together, these
findings demonstrate that troglitazone elicits a spontaneous cellular
acidosis that is maintained for at least 20-45 min and is not
dependent on increased acid production.

View larger version (18K):
[in this window]
[in a new window]
|
Fig. 5.
A: effect of troglitazone (25 µM) on
pHi monitored over 45 min in monolayers incubated in
KHHEPES medium. Results are means ± SE from 4 control and 4 troglitazone-treated monolayers. * Different from time controls
(P < 0.05). B: effect of troglitazone (25 µM) on pHi monitored over 20 min in monolayers incubated
in DMEM with HCO -CO2 buffer. Results are
means ± SE from 3 control and 4 troglitazone-treated monolayers.
* Different from time control (P < 0.05).
|
|
To assess whether troglitazone impairs cellular acid extrusion, cells
were exposed to a 20 mM NH
acid load for 4 min and
then allowed to recover in KHHEPES or KHHEPES plus 25 µM troglitazone
(Fig. 6). Exposure to KHHEPES containing
20 mM NH
results in an initial alkalinization
(NH
in Fig. 6 recorded discontinuously) followed by a
sharp fall in the pHi on removal of the
NH
-containing KHHEPES and replacement with standard
KHHEPES; this sharp fall off in pHi reflects the dissociation of intracellular NH
and diffusion of
NH3 leaving H+ to acidify the cell interior.
Note that because of the strict dependency of the major acid extruder,
the NHE, on the prevailing pHi, it is important that the
recovery response begin with the same degree of acidosis. In the
representative example shown in Fig. 6, cell pHi indeed
shows a similar drop in both the control (pHi = 6.55)
and the troglitazone (pHi = 6.50) trials. Despite the
same degree of initial acidosis, the recovery response is slowed in the
presence of 25 µM troglitazone (0.12 ± 0.03 vs. 0.040 ± 0.02
pHi/min; P < 0.01) and is
incomplete (pHi = 6.66 ± 0.08 vs. 6.98 ± 0.08; P < 0.05). Figure
7 shows that 25 µM rosiglitazone also
inhibits the acid extrusion after the standard acid load (0.18 ± 0.07 vs. 0.08 ± 0.04
pHi/min; P < 0.01). These findings are therefore consistent with glitazone
inhibition of acid extrusion as the mechanism for the spontaneous drop
in cell pHi.

View larger version (28K):
[in this window]
[in a new window]
|
Fig. 6.
Representative pHi response to exogenous 20 mM NH load in MDCK cells. Control recovery in KHHEPES
vs. recovery in KHHEPES plus 25 µM troglitazone.
|
|

View larger version (16K):
[in this window]
[in a new window]
|
Fig. 7.
Representative pHi response to exogenous 20 mM NH load in MDCK cells. Control recovery in KHHEPES
vs. KHHEPES plus 25 µM rosiglitazone.
|
|
 |
DISCUSSION |
The expressed purpose of these studies was to determine
1) whether glitazones would alter glutamine metabolism, and
2) if so, whether this effect would be associated with a
cellular acidosis. We chose to perform the study on a well-established
renal tubule epithelial cell line, MDCK, because these cells readily
elaborate extracellular matrix proteins (16) and therefore
offer an available cellular model for the thickening of the tubular
basement membrane that occurs in diabetes mellitus (16, 20,
26). It is noteworthy that in an animal model of type 2 diabetes
mellitus, troglitazone inhibits this thickening (26). We
know that production of extracellular matrix proteins is glutamine
dependent (2, 10, 24, 34, 37), and, therefore,
troglitazone might act to halt matrix expansion at least in part
through glutamine metabolism.
The three glitazones studied clearly exert a major and previously
unrecognized effect on glutamine metabolism in this well-established renal epithelial cell line. Specifically, troglitazone shifts glutamate
from the transamination pathway(s) and into the deamination pathway
(Fig. 1, from R2 into R3). The evidence for this is an increased ammonium production and decreased alanine formation (Table 1)
without increased glutamine utilization. The proof for this effect on
the fate of the amino nitrogen of glutamine is the rise in
[15N]ammonium from [2-15N]glutamine and the
decline in [15N]alanine formation (Fig. 2). The
observation that the rise in ammonium and fall in alanine are related
(Table 1 and Fig. 3, A and B, at 25 µM for
troglitazone and at 25 µM and above for ciglitazone) is consistent
with a coupled action on both the transamination and deamination
pathways. The observation that the former is predominantly, but not
exclusively, a cytoplasmic pathway (Table 3) may explain the apparent
lower sensitivity for inhibition of alanine formation if the
mitochondrial ALT reaction compensates for a falling contribution from
the cytosolic ALT. The increased ammonium formation from [2-15N]glutamine is consistent with a shift of glutamate
from the predominant cytosolic transamination pathway into the
mitochondria, since GDH is exclusively an intramitochondrial activity
(35). This shift in pathways could reflect inhibition of
the dominant cytosolic ALT reaction, activation of the mitochondrial
glutamate transporter, or both. The observed apparent near-equilibrium
state for the ALT activity in control cells and the shift away from
this near equilibrium (Table 2) are consistent with inhibition of the
ALT activity; the finding of unchanged assayable ALT activity indicates that some factor external to the ALT enzyme is inhibitory. If both the
cytosolic ALT activity and the putative mitochondrial glutamate uptake
(35) have similar pH sensitivity, then much of what is
observed can be explained by a developing cytosolic acidosis.
The pH-sensitive fluorescent probe BCECF is distributed predominantly
within the cytosol (28), providing a measure of the acid-base balance in this compartment. Using this approach, we observed
a pHi (7.13 ± 0.10) for MDCK cells in KHHEPES buffer that approximated the values found by others in MDCK cells
(pHi = 7.20 ± 0.01; Ref. 13). The
addition of each glitazone induced a prompt decrease in the
pHi (Fig. 4, A-C), occurring within 1 min
and then declining at least 0.28 units within 4 min (representing almost a doubling of the cytosolic hydrogen ion concentration). The
potency of these three glitazones in inducing the cellular acidosis
(see Fig. 4, A-C) roughly parallels their effect on
glutamine metabolism (see Fig. 3, A-C). Whether the
spontaneous acidosis is maintained or in turn involves activation of a
pH-responsive signaling system (40), or both, remains to
be determined.
The mechanism responsible for the spontaneous decline in the
pHi was shown to be a reduction in the acid extrusion
rather than an increase in acid production. Of course, acid produced from lactate formation did increase (Table 1), but this was not observed in the acute study (Table 4) nor
would an increase in acid production alone be sufficient to produce the
degree of acidosis observed in the present study; this is because the
NHEs are activated at their internal pH-sensitive site to accelerate
the rate of acid extrusion as pHi falls (1).
Therefore, a spontaneous cellular acidosis of the severity observed
should not develop as the result of increased acid production alone.
Troglitazone clearly impaired the ability of these cells to respond to
an exogenous NH
acid load, a response that is
attributed to direct inactivation of the NHE activity by troglitazone
in an endothelial cell line (15); it is noteworthy that
these investigators did not find an inhibition of the NHE activity with
rosiglitazone, suggesting that either different cell lines or unique
NHE isoforms may account for this apparent difference. Our findings are
more consistent with an indirect action of glitazones on potential
membrane signaling systems (20) capable of modulating NHE
activity (22). There is another potential acid extruder in
these cells, the sodium-independent proton secretion (17),
but this extruder is far less active than the NHEs under
NH
loading conditions (17). Further
studies are required to better characterize the responses of other cell
lines, particularly cell lines that have a single, well-characterized
NHE; the intracellular acid-base balance and cell signaling events with
long-term exposure to glitazones; and the resulting consequences for
matrix protein turnover.
 |
ACKNOWLEDGEMENTS |
This work was supported in part by the Southern Arizona Foundation
(to T. Welbourne) and National Institute of Diabetes and Digestive and
Kidney Diseases Grant DK-53761 (to I. Nissim).
 |
FOOTNOTES |
Address for reprint requests and other correspondence: T. Welbourne, Dept. of Molecular and Cellular Physiology, LSUHSC,
Shreveport, LA 71130 (E-mail
twelbo{at}lsuhsc.edu).
The costs of publication of this
article were defrayed in part by the
payment of page charges. The article
must therefore be hereby marked
"advertisement"
in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
April 2, 2002;10.1152/ajpendo.00485.2001
Received 31 October 2001; accepted in final form 31 May 2002.
 |
REFERENCES |
1.
Aronson, PS,
Nee J,
and
Suhm MA.
Modifier role of internal H+ in activating the Na+-H+ exchanger in renal microvillus membrane vesicles.
Nature
299:
161-163,
1992.
2.
Bellon, G,
Monboisse JC,
Randoux A,
and
Borel J-P.
Effects of preformed proline and proline amino acid precursors (including glutamine) on collagen synthesis in human fibroblast cultures.
Biochim Biophys Acta
930:
39-47,
1987[ISI][Medline].
3.
Bergmeyer, HU,
and
Bernt E.
-Oxoglutarate.
In: Methods of Enzymatic Analysis (2nd ed.), edited by Bergmeyer H-U.. New York: Academic, 1965, p. 324-327.
4.
Boron, WF.
Intracellular pH regulation in epithelial cells.
Annu Rev Physiol
48:
377-388,
1986[ISI][Medline].
5.
Bradford, MM.
A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding.
Anal Biochem
72:
248-254,
1976[ISI][Medline].
6.
Brosnan, JT,
Brosnan ME,
Yudkoff M,
Nissim I,
Daikhin Y,
Lazarow A,
Horyn O,
and
Nissim I.
Alanine metabolism in the perfused rat liver. Studies with 15N.
J Biol Chem
276:
31876-31882,
2001[Abstract/Free Full Text].
7.
Bucher, T,
Czok R,
Lamprecht W,
and
Latzko E.
Pyruvate.
In: Methods of Enzymatic Analysis (2nd ed.), edited by Bergmeyer H-U.. New York: Academic, 1965, p. 253-259.
8.
Busa, W,
and
Nuccitelli R.
Metabolic regulation via intracellular pH.
Am J Physiol Regul Integr Comp Physiol
246:
R409-R438,
1984[Abstract/Free Full Text].
9.
Camp, HS,
Ou L,
Wise SC,
Hong YH,
Frankowski CL,
Shen X-Q,
Vanbogelen R,
and
Leff T.
Differential activation of peroxisome proliferator-activated receptor-
by troglitazone and rosiglitazone.
Diabetes
49:
539-547,
2000[Abstract].
10.
Clark, CC,
Richards CF,
Pacifici M,
and
Iozzo RV.
The effects of 6-diazo-5-oxo-L-norleucine, a glutamine analog, on the structure of the major cartilage proteoglycan synthesized by cultured chondrocytes.
J Biol Chem
262:
10229-10238,
1987[Abstract/Free Full Text].
11.
Cortes, P,
Riser BL,
Asano K,
Rodriguez-Barbero A,
Narins RG,
and
Yee J.
Effects of oral antihyperglycemic agents on extracellular matrix synthesis by mesangial cells.
Kidney Int
54:
1985-1998,
1998[ISI][Medline].
12.
Curthoys, NP,
and
Weiss RF.
Regulation of renal ammoniagenesis. Subcellular localization of rat kidney glutaminase isoenzymes.
J Biol Chem
249:
3261-3266,
1974[Abstract/Free Full Text].
13.
Day, C.
Thiazolidinediones: a new class of antidiabetic drugs.
Diabet Med
16:
179-192,
1999[ISI][Medline].
14.
DeRosa, G,
and
Swick RW.
Metabolic implications of the distribution of the alanine aminotransferase isoenzymes.
J Biol Chem
250:
7961-7967,
1975[Abstract].
15.
De Dios, ST,
Hannan KM,
Dilley RJ,
Hill MA,
and
Little PJ.
Troglitazone, but not rosiglitazone, inhibits Na/H exchange activity and proliferation of macrovascular endothelial cells.
J Diabetes Complications
15:
120-127,
2001[ISI][Medline].
16.
Erickson, AC,
and
Couchman JR.
Basement membrane and interstitial proteoglycans produced by MDCK cells correspond to those expressed in the kidney cortex.
Matrix Biol
19:
769-778,
2001[ISI][Medline].
17.
Fernandez, R,
Oliveira-Souza M,
and
Malnic G.
Na+-independent proton secretion in MDCK-C11 cells.
Eur J Physiol
441:
287-293,
2000[ISI][Medline].
18.
Fujiwara, T,
and
Horikoshi H.
Troglitazone and related compounds. Therapeutic potential beyond diabetes.
Life Sci
67:
2405-2416,
2000[ISI][Medline].
19.
Furnsinn, C,
Neschen S,
Noe C,
Bisschop M,
Roden M,
Vogl C,
Schneider B,
and
Waldhausl W.
Acute non-insulin-like stimulation of rat muscle glucose metabolism by troglitazone in vitro.
Br J Pharmacol
112:
1367-1374,
1997.
20.
Isshiki, K,
Haneda M,
Koya D,
Maeda S,
Sugimoto T,
and
Kikkawa R.
Thiazolidinedione compounds ameliorate glomerular dysfunction independent of their insulin-sensitizing action in diabetic rats.
Diabetes
49:
1022-1032,
2000[Abstract].
21.
Kalra, J,
and
Brosnan JT.
The subcellular localization of glutaminase isoenzymes in rat kidney cortex.
J Biol Chem
249:
3255-3260,
1974[Abstract/Free Full Text].
22.
Karim, ZG,
Chambrey R,
Chalumeau C,
Defontaine N,
Warnock DG,
Pallard M,
and
Poggioli J.
Regulation by PKC isoforms of Na+/H+ exchanger in luminal membrane vesicles isolated from cortical tubules.
Am J Physiol Renal Physiol
277:
F773-F778,
1999[Abstract/Free Full Text].
23.
Kvamme, E,
Torgner IA,
and
Roberg B.
Evidence indicating that pig renal phosphate glutaminase has a functional predominant external localization in the inner mitochondrial membrane.
J Biol Chem
266:
12185-12192,
1991[Abstract/Free Full Text].
24.
Lehtinen, P,
Takala I,
and
Kulonen E.
Dependence of collagen synthesis by embryonic chick tendon cells on the extracellular concentrations of glutamine.
Connect Tissue Res
6:
155-159,
1978[ISI][Medline].
25.
Masola, B,
Peters TJ,
and
Evered DF.
Transamination pathways influencing L-glutamine and L-glutamate oxidation by rat enterocyte mitochondria and subcellular localization of L-alanine aminotransferase and L-aspartate aminotransferase.
Biochim Biophys Acta
843:
137-143,
1985[ISI][Medline].
26.
McCarthy, KJ,
Routh RE,
Shaw W,
Walsh K,
Welbourne TC,
and
Johnson JH.
Troglitazone halts diabetic glomerulosclerosis by blockade of mesangial expansion.
Kidney Int
58:
2341-2350,
2000[ISI][Medline].
27.
Mu, X,
and
Welbourne T.
Response of LLC-PK1_F+ cells to metabolic acidosis.
Am J Physiol Cell Physiol
270:
C920-C925,
1996[Abstract/Free Full Text].
28.
Negulescu, PA,
and
Machen TE.
Intracellular ion activities and membrane transport in parietal cells measured with fluorescent dyes.
Methods Enzymol
192:
38-81,
1990[Medline].
29.
Nissim, I,
Sahai A,
Sandler R,
and
Tannen RL.
The intensity of acidosis differentially alters the pathways of ammoniagenesis in LLC-PK1 cells.
Kidney Int
45:
1014-1019,
1994[ISI][Medline].
30.
Peterkofsky, A.
Involvement of glutamic acid and glutamine in protein synthesis and maturation.
In: Enzymes of Glutamine Metabolism, edited by Prusiner S,
and Stadtman E.. New York: Academic, 1973, p. 331-342.
31.
Reitzer, LJ,
Wice BM,
and
Kennel D.
Evidence that glutamine, and not sugar, is the major energy source for cultured HeLa cells.
J Biol Chem
254:
2669-2676,
1979[ISI][Medline].
32.
Roberg, B,
Torgner IA,
Laake J,
Takumi Y,
Otterson OP,
and
Kvamme E.
Properties and submitochondrial localization of pig and rat renal phosphate-activated glutaminase.
Am J Physiol Cell Physiol
279:
C648-C657,
2000[Abstract/Free Full Text].
33.
Routh, RE,
Johnson JH,
and
McCarthy KJ.
Direct action of troglitazone on mesangial cell behavior: troglitazone suppresses the secretion of type 1 collagen by mesangial cells in vitro.
Kidney Int.
61:
1365-1376,
2002[ISI][Medline].
34.
Routh, RE,
McCarthy KJ,
and
Welbourne TC.
Troglitazone inhibits glutamine metabolism in rat mesangial cells.
Am J Physiol Endocrinol Metab
282:
E231-E238,
2002[Abstract/Free Full Text].
35.
Schoolwerth, AC,
and
LaNoue KF.
Transport of metabolic substrates in renal mitochondria.
Annu Rev Physiol
47:
143-171,
1985[ISI][Medline].
36.
Segal, HL,
Beattie DS,
and
Hopper S.
Purification and properties of liver glutamic-alanine transaminase from normal and corticoid-treated rats.
J Biol Chem
237:
1914-1920,
1962[Free Full Text].
37.
Spreight, G,
Handley CJ,
and
Lowther DA.
Role of glutamine in glycosaminoglycan synthesis in vitro by chondrocytes.
Biochim Biophys Acta
540:
238-245,
1978[ISI][Medline].
38.
Thomas, JA,
Buchsbaum RN,
Zimnia A,
and
Racker E.
Intracellular pH measurements in Ehrlich ascites tumor cells utilizing spectroscopic probes generated in situ.
Biochemistry
81:
2210-2218,
1979.
39.
Welbourne, TC,
Routh RE,
Yudkoff M,
and
Nissim I.
The glutamine/glutamate couplet and cellular function.
News Physiol Sci
16:
157-160,
2001[Abstract/Free Full Text].
40.
Yamaji, Y,
Tsuganewa H,
Moe OW,
and
Alpern RJ.
Intracellular acidosis activates c-Src.
Am J Physiol Cell Physiol
272:
C886-C893,
1997[Abstract/Free Full Text].
Am J Physiol Endocrinol Metab 283(4):E729-E737
0193-1849/02 $5.00
Copyright © 2002 the American Physiological Society