| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Institut für Physiologische Chemie, Universitätsklinikum, D-45122 Essen, Germany
1Correspondence: Institut für Physiologische Chemie, Universitätsklinikum, Hufelandstr. 55, D-45122 Essen, Germany. E-mail: ursula.rauen{at}uni-essen.de
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: hydrogen peroxide • superoxide anion radical • hydroxyl radical • Fenton reaction • transition metal ions
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5
). Evidence for the mediation of the injury by ROS has
been provided by the demonstration of prominent lipid peroxidation
during cold incubation and by the fact that injury is strongly
inhibited by hypoxia and by a number of antioxidants such as the free
radical scavengers 5,5-dimethyl-1-pyrroline N-oxide and
4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl, the
lipophilic antioxidants 
-tocopherol, butylated hydroxytoluene (BHT)
or butylated hydroxyanisole and the iron chelators deferoxamine,
1,10-phenanthroline and 2,2'-dipyridyl or the hydroxyl radical
scavenger dimethyl sulfoxide (DMSO; refs 1
2
3
4
).
When liver cells are rewarmed after ‘sublethal’ cold incubation
periods, they present a clearly apoptotic picture during the rewarming
phase (6)
. The extent of this cold-induced apoptosis
greatly depends on the duration of the cold incubation, suggesting that
cellular alterations occurring during this period are a prerequisite
for the occurrence of the cold-induced apoptosis. Protection by hypoxia
and various antioxidants named above suggest that ROS are also a key
mediator of cold-induced apoptosis. In fact, the two phenomena of
hypothermia injury and cold-induced apoptosis are likely to be closely
related, being similar in initiation and differing only in the events
that occur downstream (where rewarming appears to be a prerequisite for
the development of a full-blown apoptotic picture).
Taken together, the previous data hardly leave any doubt that ROS are involved in the pathogenesis of hypothermia injury/cold-induced apoptosis. They also suggest that the main ROS involved in the two related processes of hypothermia injury and cold-induced apoptosis is the hydroxyl radical or a closely related ferryl species that is formed in an iron-dependent reaction from hydrogen peroxide (H2O2). However, on assessing the release of O2-/H2O2 in cultured hepatocytes and liver endothelial cells in the present study, we could find no evidence of an increase either in the release or in the levels of O2- or H2O2 during cold incubation. Similarly, there was no evidence that increased O2-/H2O2 release contributed to cold-induced apoptosis. So our next step was to study the availability of redox-active iron, and we found that after initiation of cold incubation the cellular pool of chelatable iron rapidly increases and that this increase—even when the generation of O2-/H2O2 decreases—is likely to be a key factor in the pathogenesis of the hypothermia injury/cold-induced apoptosis suffered by hepatocytes and liver endothelial cells.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell culture
Hepatocytes were isolated from male Wistar rats (250–310 g) as
described previously (7)
. The cells were seeded onto
collagen-coated 12.5 cm2 or 25
cm2 culture flasks (Falcon, Heidelberg, Germany)
or onto collagen-coated 6.2 cm2 glass coverslips
(Assistent, Sondheim/Röhn, Germany) and cultured in L-15 medium
supplemented with 5% fetal calf serum, L-glutamine (2 mM), glucose
(8.3 mM), bovine serum albumin (0.1%), NaHCO3
(14.3 mM), gentamicin (50 µg/ml), and dexamethasone (1 µM). Two
hours after seeding, adherent cells were washed three times with
Hanks’ balanced salt solution and supplied with fresh medium.
Experiments were started 20–24 h after the isolation of the cells.
Two rat liver endothelial cell lines, derived from the livers of male
Wistar rats, were used for additional experiments. The cells had been
isolated and characterized as described previously (8)
.
Cells were cultured in RPMI 1640 medium supplemented with fetal calf
serum (20%), L-glutamine (2 mM), penicillin/streptomycin (50 U/ml and
50 µg/ml, respectively), and dexamethasone (1 µM). Subcultures were
obtained by trypsinization. In this study,
6th–22nd passage cultures were used. For the
experiments, the cells were split 1:3 and seeded onto collagen-coated
12.5 cm2 or 25 cm2 culture
flasks or onto fibronectin-coated 6.2 cm2 glass
coverslips for fluorescence microscopy. The cells were used for
experiments on day 6 or 7 after subcultivation. By this time, the cells
were in the confluent state.
Experimental procedures
At the beginning of the experiments the cells were washed three
times with Hanks’ balanced salt solution (37°C) and then covered
with cell culture medium (Leibovitz L-15 medium for hepatocytes and
RPMI 1640 medium for liver endothelial cells, both supplemented as
described above) or Krebs-Henseleit buffer (KH; NaCl 115 mM,
NaHCO3 25 mM, KCl 5.9 mM,
MgCl2 1.2 mM,
NaH2PO4 1.2 mM,
Na2SO4 1.2 mM,
CaCl2 2.5 mM, HEPES 20 mM, pH 7.4) at room
temperature. The cells were then incubated either at 4°C or at
37°C. The incubations were performed in an atmosphere of 95% air/5%
CO2. For cold incubations, cell culture flasks
were placed in air-tight vessels that were flushed with the gas
mixture. In some experiments the transition metal chelator
2,2'-dipyridyl (2,2'-DPD, 100 µM) or the lipophilic antioxidant BHT
(20 µM) were added to the medium at the beginning of the cold
incubation. Some cultures were preincubated with deferoxamine (10 mM;
in cell culture medium, 30 min at 37°C) prior to the start of the
experiments. Solvent controls were included. After various periods of
cold incubation, the cells were rewarmed to 37°C in an incubator
containing a humidified atmosphere of 95% air/5%
CO2.
Determination of
O2-/H2O2
Determination of O2-
Cellular O2- formation
was determined by the nitroblue tetrazolium (NBT) reduction assay
(9
, 10)
. NBT was added to Krebs-Henseleit buffer in a
final concentration of 1 mg/ml and cells were incubated in this buffer
for 30–120 min at 37°C. To determine
O2- release during cold
incubations, cells were cooled down in Krebs-Henseleit buffer not
containing NBT, and this buffer was replaced by Krebs-Henseleit buffer
that did contain 1 mg/ml NBT 1) after different periods of
cold incubation and 2) for different periods of continuing
cold incubation (sometimes with rewarming) as described in results. At
the end of the incubation period, the extracellular solution was
aspirated (no formazan was found in the extracellular medium), cells
were carefully washed with Hanks’ balanced salt solution of the
respective temperature, and then lysed at 37°C with 5% sodium
dodecyl sulfate in phosphate buffer (80 mM, pH 7.8) containing 0.45%
gelatin. Samples were centrifuged for 5 min at 13,000 g. The
absorbance at 540 nm (formazan) and at 450 nm was determined against a
lysis buffer blank. Formazan concentration was calculated from
E540 corrected for unspecific
absorbance/turbidity (E450=0.51 x
E540+unspecific absorbance, as determined
spectrophotometrically using a NBT solution treated with solid
potassium superoxide and lysing the precipitated formazan in the lysis
buffer;
E540corr=(E540-E450)/0.49)
using 
540 = 7.2
cm2/µmol (11)
.
Determination of H2O2
H2O2 release was
determined by chemiluminescence using the luminol/peroxidase system,
and H2O2 steady-state
levels in the incubation medium were determined by the
H2O2-sensitive fluorescent
dye scopoletin as well as by the ferrous oxidation-xylenol orange assay
(FOX test). Chemiluminescence was detected with the ARGUS-50
photon-counting imaging system (Hamamatsu Photonics, Herrsching,
Germany). Cells were incubated in Krebs-Henseleit buffer in
stoppered cell culture flasks (gassed with 21%
O2/5% CO2/74%
N2 before measurements), which were maintained at
either 37°C or 4°C on a liquid-heated/cooled platform within the
system. Luminol (250 µM) and peroxidase (2.5 U/ml) were added to
generate the chemiluminescence signal (12)
. Photons were
counted over an area of 15.4 cm2 using an
integration time of 4 min. Cell-free controls using authentic hydrogen
peroxide added to Krebs-Henseleit buffer containing 250 µM luminol
and 50 mU/ml peroxidase were run at 37°C and at 4°C (integration
time: 5 s) in order to confirm the responsiveness of the system at
4°C (the total counts detected per nmol hydrogen peroxide added did
not decrease but actually increased at 4°C).
For the scopoletin assay (13)
, samples (1.4 ml) of the
incubation medium were taken at the times indicated and immediately
added to 100 µl Hanks’ balanced salt solution (pH 7.4) containing
0.1 mM DTPA and scopoletin in a concentration of 60 µM. After
recording the baseline fluorescence, 10 units/ml horseradish peroxidase
was added. The loss of fluorescence with oxidation catalyzed by
horseradish peroxidase/H2O2
was detected at 
ex = 355 nm and
em = 460 nm. The assay was standardized using
internal standards with known concentrations of
H2O2 prepared by diluting
30% (w/v) H2O2 in the
presence of 0.1 mM DTPA.
The FOX test was used in version 1 as described by Wolff
(14)
. Samples of the supernatant (750 µl) were added to
a concentrated reagent (250 µl). After incubation for 30 min at room
temperature, the absorbance was read at 560 nm. To increase the
specificity of the test, parallel samples were treated with catalase
(10 µg/ml) for 5 min at room temperature prior to the addition of the
reagent. Catalase-inhibitable color development was taken as a measure
of hydrogen peroxide and calibrated using authentic hydrogen peroxide
as a standard.
Determination of intracellular chelatable iron
Intracellular chelatable iron was determined using the
fluorescent indicator phen green SK, the fluorescence of which is
quenched by iron (and ‘dequenched’ when iron is removed from the
indicator by an excess of a second, nonfluorescent iron chelator; ref
15
). This method was recently refined for use in laser
scanning microscopy (allowing better quantification than conventional
fluorescence microscopy as differences in cellular dye loading can be
taken into account; ref 16
).
Cellular measurements
Hepatocytes cultured on glass coverslips were loaded with phen
green SK (20 µM of the diacetate) for 10 min at 37°C (as described
previously, ref 15
) or for 30 min at 4°C (loading time
was prolonged in order to achieve comparable dye loading). Liver
endothelial cells were loaded with 20–50 µM phen green SK diacetate
for 10–30 min at 37°C or for 30 min at 4°C. Measurements were
performed on a laser scanning microscope (LSM 510, Zeiss, Oberkochen,
Germany) equipped with an argon laser and a helium/neon laser.
Quantitative fluorescence measurements were performed at
ex = 488 nm and em 
505 nm as described previously (16)
. Neither phen green SK
nor laser exposure using the settings described in ref 16
had any cytotoxic effects on hepatocytes. The required temperature was
maintained either by using a thermostated microscope stage (37°C;
Zeiss) or by using a liquid-cooled aluminum microscope stage (Zeiss)
connected to a cryostat (Thermomix BM, B. Braun Biotech International,
Melsungen, Germany).
Five to 10 min after the beginning of the measurements, cellular
chelatable iron was removed from the indicator phen green by the
addition of an excess of the cell-permeable iron chelator 2,2'-DPD (5
mM) to the supernatant (15
, 16)
. At the end of the
experiments, the uptake of the vital dye propidium iodide (5 µg/ml)
was routinely determined in order to detect loss of cell viability. The
red fluorescence of propidium iodide excited at 543 nm using the
helium/neon laser was collected through a 560 nm long-pass filter.
Determination of intracellular phen green SK concentrations
The intracellular concentration of phen green SK in hepatocytes
and liver endothelial cells was determined from cellular fluorescence
(in arbitrary units) after ‘dequenching’ using 2,2'-DPD (5 mM)
compared with that of phen green SK (2–50 µM, free dye) standards
dissolved in a chelex-treated medium designed to simulate the
composition of the cytosol (15
, 16)
. To perform a
calibration curve, aliquots (100 µl) of the medium (37°C or 4°C)
containing known concentrations of phen green SK (free dye) were placed
on chelex-treated glass coverslips (the same as used to collect
cellular microfluorographs). The fluorescence measurements were
performed in a focal plane 10 µm above the surface of the coverslips,
safely within the medium, with the same laser scanning parameters used
for cellular measurements (16)
. Cellular dye
concentrations were assessed at 4°C as well as at 37°C.
Ex situ calibration of Fe2+-induced
quenching of phen green SK fluorescence
In a cell-free system the quenching effect of
Fe2+ on phen green SK fluorescence in a
‘cytosolic’ medium (37°C or 4°C) was determined using the LSM 510
imaging system as described in ref 16
.
Assessment of cellular chelatable iron using the fluorescent
indicator calcein
The procedure used was derived from the original procedure
described for K562 cells by Breuer et al. (17
, 18)
, a
modification used for cultured hepatocytes by Stäubli and
Boelsterli (19)
, and our own previous experience with
calcein (15)
. Cultured hepatocytes were loaded with
calcein (50 nM of the acetoxymethyl ester in Hanks’ balanced salt
solution) for 10 min at 37°C or for 30 min at 4°C, cultured liver
endothelial cells were loaded with 50–100 nM calcein acetoxymethyl
ester for 10 min at 37°C. Measurements were performed on the laser
scanning microscope using the same instrument settings as for phen
green measurements except that excitation intensity was set to 1%.
Dequenching was achieved by the addition of 2,2'-DPD (5 mM).
Other assays
Thiobarbituric acid-reactive substances
Thiobarbituric acid-reactive substances (TBARS) were determined
in the supernatant incubation solution after various incubation times
using the assay described in ref 20
a second with minor
modifications. The amounts of TBARS formed were expressed as
malondialdehyde equivalents using 1,1,3,3-tetramethoxy-propane as a
standard.
Lactate dehydrogenase (LDH) release
Extracellular, i.e., released, LDH activity was measured using a
standard assay. At the end of the incubation period, cellular LDH
activity was determined after lysis of the cells with the detergent
Triton X-100 (1% in Hanks’ balanced salt solution, 30 min at 37°C).
LDH values were corrected for the change in the volume of incubation
medium resulting from repetitive sampling and released LDH activity was
given as a percentage of total LDH activity.
Statistics
All experiments were performed in duplicate and repeated three
to nine times. Data are expressed as means ± SD. Data
obtained from two groups were compared by means of Student’s
t test and comparisons among multiple groups were performed
using an analysis of variance with Student-Newman-Keuls post
hoc comparisons. A P value of < 0.05 was
considered significant.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
, both hepatocytes and
liver endothelial cells showed pronounced hypothermia injury when
incubated in Krebs-Henseleit buffer or cell culture medium at 4°C
under normoxic conditions (e.g., 90±5% of cultured hepatocytes lost
viability when incubated in Krebs-Henseleit buffer at 4°C for 48 h). Rewarming the hepatocytes after several hours of cold incubation
strongly aggravated the cold-induced injury (Fig. 1
). As described previously (6)
, the injury occurring on
rewarming was dependent on the duration of the cold incubation (Fig. 1)
and controls kept at 37°C for the whole incubation time showed far
less injury (LDH release after 27 h warm incubation: 28±7%);
this stresses the importance of the hypothermic period for the
development of rewarming injury. Similarly, loss of viability of liver
endothelial cells after 24 h of cold incubation was strongly
aggravated by rewarming (LDH release 39±9% after 24 h of cold
incubation in Krebs-Henseleit buffer and 75±14% after 24 h cold
incubation/3 h rewarming). In both cell types, the injury occurring on
rewarming was accompanied by prominent morphological changes consistent
with apoptosis, such as blebbing of the plasma membrane, chromatin
condensation (hepatocytes), and nuclear condensation (data not shown);
the results in Krebs-Henseleit buffer were comparable to the previous
findings in cell culture medium (6)
. As described
previously (1
2
3
4
, 6)
, both hypothermia injury and
cold-induced apoptosis arising during rewarming were almost completely
prevented by BHT and largely (hepatocytes) or completely (liver
endothelial cells) inhibited by deferoxamine or 2,2'-dipyridyl (data
not shown). Also in accordance with previous results (1
2
3
4
, 6)
, the two related injuries were accompanied by lipid
peroxidation (data not shown).
|
Cellular formation of superoxide anion radicals during cold
incubation/rewarming
Given that increased production of the superoxide anion radical
(O2-) and/or of
H2O2 appeared to be the
most likely factor responsible for the ROS-mediated hypothermia injury
(see first page of article), we used several methods to assess the
release/level of these species. Hepatocytes formed considerable
quantities of O2- under
physiological conditions: 2.1 ± 0.6 nmol formazan
formation/106 cells/min were detected with the
NBT reduction assay (1 mg/ml NBT) during warm incubation of the cells
(at 37°C) in Krebs-Henseleit buffer (Table 1
, Fig. 2A
). The O2- detected
chiefly appeared to react intracellularly with the NBT, as neither
extracellular formation/precipitation of formazan nor attenuation of
the formazan formation by extracellular superoxide dismutase (50
µg/ml) was observed (data not shown). Formazan formation was linear
over time up to a total formation of 
200
nmol/106 cells (Fig. 2A
). Hypoxia
strongly inhibited formazan formation (Table 1)
, confirming that an
oxygen-derived reductant, i.e.,
O2-, was responsible at least
for a large part of the NBT reduction. As positive controls, cyanide
(which is known to increase the release of
O2- from the mitochondrial
respiratory chain by banking up electrons; ref 21
) and
menadione (a redox-cycling quinone giving rise to intracellular
O2- generation; ref
22
) strongly increased formazan formation.
|
|
When we measured the O2- formed
by cultured hepatocytes during cold incubation, we were surprised to
find a marked decrease rather than an increase in the signal at 4°C,
even though ROS-mediated injury occurred. During cold incubation,
detectable O2- formation was
very low and remained below the release observed at 37°C for the
whole period of cold incubation (Table 1
, Fig. 2B
). A
similar low rate of O2-
formation during cold incubation was observed when NBT was not present
during the entire period of cold incubation but added after
12 h of cold incubation for some hours (data not shown). Further
controls performed at 4°C showed that the assay system was sensitive
at this low temperature, since cyanide as well as menadione produced
easily detectable signals (Table 1)
. During rewarming, the rate of
cellular O2- generation
remained decreased for half an hour (Fig. 3A
), irrespective of the cold incubation period. After 30 min
of rewarming, O2- formation
re-increased but only reached the rates of untreated, i.e., warm
control cells during rewarming after short cold incubation periods
(i.e., 5 h). The rate of
O2- formation during the second
half hour of rewarming remained slightly below warm controls after
16 h of cold incubation (compare Fig. 3A
and Fig. 2A
) and far below control values after 24 h of cold
incubation (data not shown).
|
Cultured liver endothelial cells incubated in Krebs-Henseleit buffer at
37°C showed far less formation of
O2- than cultured hepatocytes
(Table 1)
, a finding that is in line with the largely glycolytic energy
metabolism of these cells (23)
and their comparatively low
number of mitochondria (24)
. As in hepatocytes,
O2- formation was also
substantially decreased at 4°C in liver endothelial cells (Table 1)
.
During rewarming of liver endothelial cells after 24 h of cold
incubation, O2- formation
increased and tended to exceed warm control levels; however, this
increase over warm control levels was not significant (data not shown)
and occurred when substantial endothelial cell injury had already
occurred (see above).
H2O2 release/levels during cold
incubation/rewarming
The hydrogen peroxide released by cultured hepatocytes incubated
in Krebs-Henseleit buffer at 37°C gave rise to a chemiluminescence
signal of 1010 ± 343 counts/15.4 cm2/4 min
(values corrected for background counts) using the luminol/peroxidase
system. When the cells were cooled down to 4°C, the signal decreased
practically to background levels (335±296 counts/15.4
cm2/4 min) and stayed there for the remainder of
the cold incubation (up to 24 h), suggesting a very low release of
hydrogen peroxide.
While the cellular formation of hydrogen peroxide had decreased at
4°C, H2O2 degradation
might also be impaired at the lower temperature. However, determination
of H2O2 steady-state levels
in hepatocyte cultures with the scopoletin assay showed that
H2O2 levels also had not
increased either during cold incubation or during rewarming:
H2O2 steady-state levels
ranged from 46 to 115 nM in cells incubated at 37°C and 55–96 nM
during cold incubation (means and time courses, see Table 2
). These values were even below the values found in (cell-free) aqueous
solutions/buffers exposed to air, which contained 80–120 nM
H2O2. During rewarming,
H2O2 steady-state levels
remained unchanged when cold incubation time was 5 h or 24 h
and actually dropped during the first half hour of rewarming after
16 h of cold incubation (Fig. 3B
; note that the drop in
H2O2 steady-state levels is
in line with the continuing lower release of
O2- during this period, Fig. 3A
, and occurs at the time of prominent rewarming injury,
Fig. 1
). During rewarming of cultured hepatocytes at the different time
points, H2O2 steady-state
levels never exceeded warm control levels (data not shown).
|
In line with the results of the scopoletin assay, determination of
H2O2 (or short-chain,
water-soluble hydroperoxides) in the hepatocyte supernatant with the
FOX test (a peroxidase-independent assay for
H2O2) showed that
H2O2 steady-state levels
remained below 100 nM during cold incubation (Table 2)
. For both the
scopoletin assay and the FOX test, calibrations were performed with
cold and warm standards (giving the same values).
Cultured liver endothelial cells exposed to 4°C showed largely
similar results during cold incubation. Hydrogen peroxide steady-state
levels, as determined with the scopoletin assay, remained below 100 nM
during 24 h of cold incubation and were not significantly
different from the levels measured during warm incubation (Table 2)
.
During rewarming, however, the endothelial cell results were somewhat
inconsistent with single values exceeding warm control levels although
this did not correlate with loss of viability: most of the increased
values (only in one out of four experiments exceeding 100 nM) were
observed in cells that did not die during rewarming (early
rewarming after 5 h); values after 16 h of cold incubation/1
h of rewarming amounted to 72 ± 2 nM, values after 24 h of
cold incubation/1 h rewarming to 90 ± 57 nM (only one single
value exceeded 100 nM).
Determination of the chelatable iron pool during cold
incubation/rewarming
Since it was not possible to account for the iron-dependent
hypothermia injury/cold-induced apoptosis, apparently mediated by
hydroxyl radicals, in terms of increased release or levels of
O2-/H2O2,
we looked for alterations in the cellular homeostasis of redox-active,
chelatable iron. Similar to the results described previously
(16)
, the hepatocellular chelatable iron pool consisted of
3.1 ± 2.3 µM iron when determined by the phen green method (at
37°C; Fig. 4A
, B
, C
, Fig. 5
). When cells were cooled down to 4°C, then loaded with phen green and
measured immediately (i.e., cooling time 60 min, incubation at 4°C 30
min), the cellular chelatable iron pool had increased to 7.7 ±
2.4 µM. Dye loading and cellular distribution of the dye were similar
in cells loaded at 4°C and at 37°C (Fig. 4B
, 4E
). Longer cold incubation periods did not elicit any
further significant changes in the cellular chelatable iron pool (Fig. 5
; at time points exceeding 6 h of cold incubation it was
impossible to assess the cellular chelatable iron pool, as most cells
died during the measurements; what did the damage was not the scanning
procedure or the fluorescent indicator, but the multiple washing
procedures required for loading).
|
|
When cells were rewarmed after 3 h of cold incubation, the
increase in the cellular pool of chelatable iron proved to be
reversible: 1 h after rewarming, the cellular chelatable iron pool
amounted to 1.8 ± 2.0 µM (Fig. 5)
. When cells were rewarmed
after 6 h of cold incubation, the cellular pool of chelatable iron
could not be determined as cells died before loading with phen green
was completed.
To confirm this rapid increase in the cellular pool of chelatable iron,
we used a second method, namely, the calcein method. In line with the
results obtained with phen green, the calcein-detectable iron pool of
cultured hepatocytes also increased after initiation of cold
incubation; this is evidenced by a stronger quenching of the calcein
fluorescence in cells incubated at 4°C as compared to cells incubated
at 37°C (Fig. 6
).
|
The chelatable iron pool of cultured liver endothelial cells
(untreated), as determined by the phen green method, already amounted
to 6.4 ± 3.7 µM. Relative to this fairly high chelatable iron
pool, cellular loading with phen green was weak: the intracellular dye
concentration was 22.3 ± 11.8 µM (even after loading was
optimized: 50 µM, 30 min at 37°C). This dye loading sets an upper
limit to the useful detection range at around 7 µM chelatable iron
(cf. to ref 16
), i.e., intracellular phen green
fluorescence was already almost completely quenched under
warm control conditions. Cellular dye loading thus did not allow the
detection of increases in the chelatable iron pool in the endothelial
cells with phen green. Unfortunately, in liver endothelial cells
calcein could not be used to determine the chelatable iron pool either,
as liver endothelial cells strongly compartmentalized calcein.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
2
3
4
5
6)
. The most
likely cellular source of the hydrogen peroxide required for this
process are mitochondria, which are known to release
O2- as a by-product of
oxidative phosphorylation; this subsequently dismutates, spontaneously
or catalyzed by (Mn or Cu/Zn) superoxide dismutase, to
H2O2 (25
26
27
28
29)
.
An increased mitochondrial formation of
O2-, secondary to mitochondrial
alterations such as functional defects in the mitochondrial respiratory
chain or the mitochondrial permeability transition, has been described
as a key factor in the pathogenesis of various types of cell injury and
is currently considered to play a crucial role in apoptotic processes
(25
, 27
, 29
, 30)
. Furthermore, few reports suggest that
mitochondrial alterations might occur during cold storage of organs
such as the liver (31
, 32)
. However, the data presented
here show that, in contrast to this well-known pathway,
O2- release actually
decreased during the initiation of this type of injury
(i.e., during cold incubation; Fig. 2
, Table 1
). Similarly,
H2O2 release also
decreased during cold incubation. Thus,
H2O2 steady-state levels
did not at all increase during cold incubation of the cells (although
H2O2 degradation has also
decreased at 4°C, it can obviously keep pace with
H2O2 formation;
H2O2 degradation at 4°C
is delayed by a factor of 3–5; unpublished result). Furthermore, we
found that hydrogen peroxide levels measured during cold incubation
were < 1/50 of those required to elicit hepatocellular or liver
endothelial cell injury in the absence of hypothermia, i.e., at 37°C
(unpublished results). Although pronounced cell death already occurs
during cold incubation (Fig. 1
; refs 1
, 2
, 4
), the
hypothermic period, which is associated with decreased
O2-/H2O2
release, is also essential for the development of rewarming
injury/cold-induced apoptosis (Fig. 1
; refs 3
, 6
). In
addition, O2- release and
H2O2 steady-state levels
were either decreased or unaltered during the rewarming phase of
hepatocytes too (Fig. 3)
, with a decrease predominating at the times
when prominent rewarming injury occurred (Fig. 3
, cf. Fig. 1
).
Similarly, in liver endothelial cells,
O2-
formation/H2O2 levels did
not significantly increase during rewarming; although these cells
showed somewhat inconsistent results, these occasional (and slight)
increases were obtained either under conditions in which no subsequent
cell death was observed or at late time points when they were more
likely a consequence than a trigger of the injury. Taken together, the
results obtained make it highly unlikely that an increased formation of
O2-/H2O2
is the decisive trigger of (iron-dependent) hypothermia
injury/cold-induced
apoptosis.
The iron ions catalyzing the formation of highly reactive hydroxyl
radicals and ferryl species are generally thought to belong to an
ill-defined pool of ‘redox-active’, non-heme, non-ferritin iron that
is considered to be chelated by low molecular weight components such as
ATP, phosphate, or citrate, but possibly also loosely attached to
proteins or lipids (26)
. This pool is sometimes also
called the ‘transition pool’; other terms used include ‘free iron’
and ‘low molecular weight iron’. However, for practical purposes the
pool is best characterized methodologically as a ‘chelatable iron
pool’, which is the term we have used here.
As the bulk of cellular iron is tightly bound to proteins, only a small
portion of cellular iron (0.2–3%) belongs to the chelatable iron pool
(26
, 33
34
35
36
37)
. The small size of the chelatable iron
pool—where there is also a wealth of protein-bound iron—has long
hampered its experimental accessibility and still poses challenges to
its determination. A broad range of methods have been used to determine
the cellular chelatable iron pool, most of them involving tissue
homogenization or cell lysis prior to the iron determination. As these
steps involve a high inherent danger of artifacts (e.g., due to
proteolysis)—and all the more so if they are used when cell injury has
occurred—we used two methods applicable to viable cells to determine
the cellular chelatable iron pool during cold incubation. The first
method, using the metal-sensitive fluorescent dye phen green, is fairly
specific for the reduced forms of the ions of the transition metals
iron and copper, and in cellular systems it appears to detect mainly Fe
(II) (15)
. The second method, employing the fluorescent
indicator calcein, has been used for quantitative iron measurements in
K562 cells (17
, 18)
; however the calibration used there
did not work in isolated hepatocytes (15)
, in which
calcein has only been used for semiquantitative measurements
(19)
as in the present study. Using the phen green method,
we determined a chelatable iron pool of 3.1 ± 2.3 µM in
cultured hepatocytes under physiological conditions. This pool size is
at the lower end of the range of the values previously determined for
this pool in liver tissue or cultured hepatocytes (3.5–230 µM; refs
35
, 36
, 38
39
40
); however, this appears to be reasonable as
all of these previous studies either used cell-destructive methods or
nonviable tissue with the associated danger of release of protein-bound
iron and thus a high probability of overestimation of the pool size. In
cultured liver endothelial cells, we found a surprisingly high
chelatable iron pool of 6.4 ± 3.7 µM; to our knowledge, the
chelatable iron pool has not previously been determined in endothelial
cells of any kind.
When the temperature was lowered, the hepatocellular chelatable iron
pool increased to 7.7 ± 2.4 µM (Fig. 5)
. The dye-loading
conditions chosen gave similar intracellular dye concentrations and dye
distribution for both temperatures (Fig. 4)
. Furthermore, our
calibration procedure took into account the actual dye concentration
for every single cell (for details, see ref 16
), and
stratification of the data for intracellular indicator concentration
(to exclude potential artifacts by an altered competition with cellular
constituents) also showed an increase in cellular chelatable iron
during cold incubation (unpublished results); thus, erroneous results
based on different dye loading are unlikely. Calibrations were
performed at 37°C as well as at 4°C, both temperatures giving the
same calibration curves (Fig. 4C
, 4F
), in
line with the formation of the 3:1 complex to be expected for
phenanthroline (compare ref 16
). Measurements with a
chemically quite different indicator, calcein, also showed a rapid
increase in the hepatocellular chelatable iron pool at 4°C (Fig. 6)
.
Due to the stoichiometry of the calcein:iron complex (1:1 or 1:2, ref
41
) and to potential competition with cellular
constituents (17
, 18)
, the response of calcein to iron is
less than that of phen green SK to iron (3:1 complex, no competition
with low molecular weight compounds; ref 16
). Taken
together, despite the technically demanding measurements of the
chelatable iron pool, there is strong evidence of a rapid increase in
the chelatable iron pool during cold incubation of cultured
hepatocytes. With regard to the low release of
O2-/H2O2
by cultured liver endothelial cells and the prominent protective effect
of iron chelators in this cell type, a similar pathomechanism appears
likely for liver endothelial cells, although in this cell type
increases of the pool were not amenable to measurements.
The increase in chelatable iron observed after introduction of
hypothermic conditions was fairly quick but then remained constant for
some hours (Fig. 5)
. The source of the iron responsible for the increase during cold incubation, which appeared to be reversible at
early time points, remains to be identified. Usually, a release of
iron from the highly iron-loaded iron storage protein ferritin,
provoked either by ‘reductive stress’, e.g.,
O2-, or by proteolysis, is
considered to be capable of increasing the cellular chelatable
iron pool under pathological conditions (19
, 35
, 36
, 38
, 42
, 43)
. However, given the low levels of O2- release during cold
incubation (Table 1
, Fig. 2
), the rapidity of the increase in iron
levels during cold incubation, and the decrease during
rewarming (Fig. 5)
, a release of iron by these ways does not appear
very likely. Possibly, a release of iron by macromolecules altered in
their conformation at the lower temperature contributes, but a release
from cellular compartments such as lysosomes must also be considered
(44
, 45)
.
The iron-initiated, free radical-mediated injury studied here is of
potential relevance for cold storage of a whole range of cell types: we
observed a similar iron-dependent, free radical-mediated injury in
isolated rabbit proximal tubules (5)
and in
LLC-PK1 kidney cells (G. Schulze Frenking, U.
Rauen, unpublished results). Magni et al. have described a
Desferal-inhibitable injury during cold aerobic perfusion of the rat
heart (46)
. Fuller et al. observed an increased
susceptibility of kidney homogenates to lipid peroxidation (assessed
after incubation of the homogenate for 90 min at 37°C under aerobic
conditions) after cold storage of the kidneys, which was preventable
using deferoxamine (47)
. Similarly, Vreugdenhil et al.
found an increased susceptibility of cold-stored hepatocytes and rat
livers toward oxidative stress induced by t-butyl
hydroperoxide, which was decreased by deferoxamine (48)
.
These findings might well be due to the alterations discussed here.
Using cell-disruptive methods, Healing et al. (49)
found
an increase in chelatable iron in kidneys after cold ischemic storage;
however, they attributed this increase to ischemia and the subsequent
release of O2- during
reperfusion. Taken together, there is ample evidence that cold-induced
release of iron might be a more widespread phenomenon contributing to
injury during/after cold exposure of cells and tissues. As hypothermia
is widely used for the (short-term) storage of many sorts of cells,
tissues and organs for scientific and especially for clinical purposes,
further elucidation of this injury with the aim of potently inhibiting
or indeed preventing it appears worthwhile. The type of injury
described here should especially be considered in all instances in
which oxygen is present during cold storage: the increased availability
of chelatable iron and the risk of ROS-mediated injury cast severe
doubts on the benefits of measures such as cold aerobic perfusion for a
better preservation of organs for transplantation. The use of
membrane-permeable iron chelators that trap the released iron in a
redox-inactive form would appear desirable or mandatory under these
conditions.
Besides its potential importance for practical purposes, the
injury studied here poses new questions about the nature and the
pathophysiological role of the cellular chelatable iron pool. Usually,
this pool is thought to play a merely catalytic role in
H2O2 toxicity. Increases in
redox-active iron have been suggested in certain pathological
conditions, such as ischemia or ethanol toxicity (35
, 42
, 49)
; however, these increases were accompanied or preceded by an
increased release of the reactive oxygen species
O2- and/or
H2O2, i.e., the increases
in the chelatable iron pool were considered as merely contributory
factors enhancing the toxicity of the ROS released (35
, 36
, 42
, 49
, 50)
. Here, however, we have shown that an alteration in the
cellular chelatable iron pool can elicit oxidative cell injury even
when there is no increased release, and indeed when there is actually a
decreased release of
O2-/H2O2.
Either these tiny, (sub)physiological amounts of
H2O2 are sufficient, when
increased concentrations of redox-active iron are present, to produce
lethal amounts of hydroxyl radicals via classical Fenton chemistry or,
alternatively, oxidizing species derived from the reaction of dioxygen
with iron—as recently suggested by Qian and Buettner
(51)
—represent the major injurious species. The fact that
prominent iron-dependent injury also occurs during cold incubation of
LLC-PK1 kidney cells, which produce far less
O2- even than the cells
described here (unpublished result), as well as the finding that the
cell injury described here could not be decreased by the addition of
the H2O2 degrading enzyme
catalase (unpublished result), both lend some credence to the latter
possibility, which has received far less consideration. The ratios of
[O2]/[H2O2]
and [Fe]/[H2O2] conform
to the assumptions made for iron-dioxygen chemistry made in ref
51
. However, irrespective of whether the injury is
elicited by hydroxyl radicals formed from ‘physiological’
H2O2 due to an increased
availability of redox-active iron or by iron-oxygen complexes, the
results presented here require the consideration of the cellular
chelatable iron pool as a pathogenetically decisive factor in its own
right, whose alteration can elicit cell injury in the absence of any
increase in
O2-/H2O2
release, a finding that may be also of relevance for other
oxidative cell injuries including ROS-mediated apoptotic
processes.
|
ACKNOWLEDGMENTS |
|---|
Received for publication February 22, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
