[CANCER RESEARCH 49, 1671-1675, April 1, 1989]
Intracellular Hydroxyl Radical Production Induced by Recombinant Human Tumor
Necrosis Factor and Its Implication in the Killing of Tumor Cells in Vitro
Naofumi Yamauchi, Hiroshi Kuriyama, Naoki Watanabe, Hiroshi Neda, Masahiro Maeda, and Yoshiro Niitsu
Department of Internal Medicine (Section 4), Sapporo Medical College, South-1, West-16, Chuo-ku, Sapporo, 060, Japan
ABSTRACT
This study investigated the effect of recombinant human tumor necrosis
factor (rh I NT') on hydroxyl radical production by established cell lines
in vitro, and its implication in the killing of tumor cells by rhTNF. During
incubation of TNF sensitive mouse tumorigenic fibroblast L-M cells (2
x 10"cells) in the presence of rhTNF (100 U), hydroxyl radical produc
tion as detected by the evolution of methane gas from dimethyl sulfoxide
increased gradually, at 18 h reaching 1.8 times that in the absence of
rhTNF. This increase was dependent on the concentration of rhTNF and
was effectively prevented by the simultaneous addition of anti-rhTNF
monoclonal antibody III 2F3, which inhibited both the binding of rhTNF
to its receptor and the cytotoxic activity of rhTNF. The addition of iron
chelator 2,2'-bipyridine, which inhibits iron-catalized Fenton reaction
and so inhibits hydroxyl radical generation, suppressed both the increase
of hydroxyl radical production and the cytotoxicity induced by rhTNF.
A similar increase in hydroxyl radical production in the presence of
rhTNF was also detected with TNF-sensitive human myosarcoma-derived
KYM cells, but no such increase was detected with TNF insensitive
human embryonic lung fibroblast HEL cells.
The results show that rhTNF induces increased hydroxyl radical
production in TNF-sensitive cells, and suggest that this plays an impor
tant role in the mechanism of tumor cell killing by rhTNF.
INTRODUCTION
TNF1 is a macrophage/monocyte-derived
anticancer cytokine
(1-3) with strong cytotoxicity to tumor cells in vitro (4-10).
The mechanism of its lethal effect on tumor cells, however,
remains largely unexplained, although some studies have sug
gested the involvement of lysosomal enzyme (11, 12) and hy
droxyl radical (12-14), based on the observation of suppressed
TNF cytotoxicity in the presence of various inhibitors.
Various studies have shown that TNF cytotoxicity is sup
pressed in the presence of the hydroxyl radical scavengers
dimethyl sulfoxide (DMSO) (12-14) and promethazine (14),
and also in the presence of an iron chelator (desferal) (13, 14),
which is known to block hydroxyl radical production. Such
studies are methodologically limited, however, as cytotoxic
suppression in itself does not provide a means of determining
the kinetics of hydroxyl radical generation in TNF-sensitive
cells, nor of detecting any changes in hydroxyl radical quantities
in cells insensitive to TNF. It has thus not been possible to
demonstrate a direct relationship between the cytotoxic effect
of TNF and its stimulation of hydroxyl radical production.
We therefore attempted to measure the amount of hydroxyl
radical produced in several cell lines under various conditions
following their exposure to rhTNF and investigated the relation
between changes in this production and the cytotoxic effects of
rhTNF.
MATERIALS
AND METHODS
Materials. SOD, 2,2'-bipyridine,
and DETAPAC were purchased
from Sigma (St. Louis, MO); catalase from Boeringer Mannheim (W.
Germany); DMSO and HEPES from Wako Chem. Co., Ltd. (Osaka).
rhTNF and Monoclonal Antibody to rhTNF. rhTNF, produced in
Escherichia coli and purified (99.9%) (15), and mouse anti-rhTNF
monoclonal IgG antibody III 2F3 which neutralized the rhTNF activity
(16) were generously provided by Asahi Chemical Ind. Co., Ltd. (To
kyo). The rhTNF had a specific activity of 2.3 x 10" U/mg protein as
determined by its cytotoxic activity against mouse L-M cells (16) and
a molecular mass of 51,000 as trimer.
Radioiodination of rhTNF was performed by the method of Bolton
and Hunter (17) using Bolton and Hunter reagent (4000 Ci/mmol, 148
TBq/mmol, NEN Research Products, Boston, MA) to obtain 125Ilabeled rhTNF with specific activity of 250 cpm/fmol and no significant
loss in bioactivity (18).
Cell Culture. L-M (mouse tumorigenic fibroblast) cells and HEL
(human embryonic fibroblast) cells were obtained from the American
Type Culture Collection, Rockville, MD. These cells were cultured in
EMEM (Flow Lab., Australia) with 5% or 10% heat-inactivated FBS
(Flow Lab.), 100 units/ml of penicillin, 100 Mg/ml of streptomycin and
10 HIMHEPES (pH 7.2) at 37'C in a 5% CO2 incubator. KYM cells
derived from human myosarcoma were kindly provided by Dr. M.
Sekiguchi (Cancer Research Institute, Tokyo University) and were
cultured in DM-160 medium (Kyokuto Pharmaceutical Ind., Tokyo)
with 10% FBS, antibiotics, HEPES and other conditions as above.
In Vitro Assay for Cytotoxic Action. L-M cells, (1 x 10" cells/200
ii\) were added to well of 96-well microculture plate (Sumitomo Bake
lite, Tokyo), and incubated at 37"C for 4 h in 5% CÛ2. Medium was
aspirated and both rhTNF (200 U/ml, 100 ¿/I)
and one of the agents
indicated in Table 1 (100 ¿il)in EMEM were added, followed by
incubation at 37°Cfor 24 h in 5% CO2. The cytotoxic activity was then
assessed by the dye uptake method using méthylène
blue (16).
Binding Assay. Binding assays were performed in triplicate on monolayers of L-M cells (5 x IO5 cells/well) in 12-well culture clusters
(Costar, Cambridge, MA). The cells were incubated with 5 nM (586 U/
ml)['"l]rhTNFfor
120 min at 4°Cin EMEM containing 0.5% gelatin
as binding medium. The cells were then washed three times with chilled
binding medium and lysed with 1 ml of 0.5 M NaOH, and 0.5 ml of
the lysate was counted for radioactivity. Nonspecific binding was deter
mined in the presence of a 1000-fold excess of unlabeled rhTNF protein.
Quantification of Hydroxyl Radical Formation. Hydroxyl radical pro
duction was measured by the formation of methane from DMSO, in
accordance with the method of Repine et al. (19). Each 2-ml reaction
mixture was prepared in a siliconized 3.5-ml glass tube (Pierce, Rockford, IL) by the sequential addition of 1 ml of cell suspension (2 x IO7/
ml), 0.4 ml of DMSO solution (2 M) and 0.4 ml of solution containing
bipyridine, III 2F3 or mouse IgG as applicable, each of these in the
culture medium described above but containing 20 mM HEPES (pH
7.2), followed by sealing of the tube with a silicon teflon septum and
open-top screw cap, and injection of 0.2 ml of the same 20 mM HEPES
medium containing rhTNF through the septum. The mixture was
vigorously mixed and then incubated up to 24 h at 37°Cin a shaking
water bath; the reaction was terminated by placing the tube on melted
ice.
Received 7/12/88; revised 10/27/88; accepted 12/13/88.
A 0.2-ml sample of the headspace gas from each tube was withdrawn
The costs of publication of this article were defrayed in part by the payment
into
a gas-tight syringe (Hamilton, Reno, NV), after its vigorous mixing
of page charges. This article must therefore be hereby marked advertisement in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
by depression and withdrawal of the syringe plunger at least 10 times.
1The abbreviations used are: TNF, tumor necrosis factor; DMSO, dimethyl
The methane concentration in the 0.2-ml gas sample was determined
sulfoxide; rhTNF, recombinant human tumor necrosis factor; DETAPAC, dion a Shimazu GC-9A gas Chromatograph (Shimazu, Tokyo) equipped
ethylenetriaminopentaacetic acid; HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; EMEM. Eagle's minimal essential medium; SOD, Superoxide
with a flame-ionization detector and a 3-mm x 2-m stainless steel
dismutase; FBS, fetal bovine serum.
column packed with 60/80 mesh active carbon (Gasukuro Kogyo,
1671
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
HVDROXYL RADICAL PRODUCTION
Tokyo). Nitrogen gas was used as carrier, at a flow rate of 50 ml/min;
injector, column, and detector temperatures were 200, 150, and 200°C,
respectively. A calibration curve, linear over the range of 0.01-35 nmol
methane, was constructed with a 99.7% methane standard (Gasukuro
Kogyo). The retention time for authentic methane under these condi
tions was 1 min. The methane content, expressed as pmol of methane
per 1.5 ml of headspace gas, was determined from the intersection on
the calibration curve of the peak area after subtraction of the area
corresponding to content of the laboratory air (14.2-18.2 pmol/0.2
ml). Because of the known tendency for variations to occur in the results
of biological assays, each experiment was accompanied by an assay, by
the method described above, for hydroxyl radical production by a
control containing no rhTNF.
RESULTS
Suppression of rhTNF Cytotoxicity. The effects of various
hydroxyl radical scavengers, iron chelators and antioxidant
enzymes on rhTNF cytotoxicity, as determined for L-M cells,
are shown in Table 1. In the presence of DMSO (400 HIM)as
hydroxyl radical scavenger and bipyridine (60 pM) as iron
chelator and thus suppressor of hydroxyl radical formation
(20), the cytotoxic effect of rhTNF was decreased to 24 and
63%, respectively, ofthat observed with rhTNF alone. In con
trast, no significant reduction in cytotoxic effect was observed
in the presence of the SOD or catalase antioxidant enzyme or
the DETAPAC iron chelator, for all of which the cell membrane
is impermeable (21-23). DETAPAC in fact apparently served
to enhance the rhTNF cytotoxicity for reasons which are as yet
unknown.
Time Course of Cytotoxic Activity of rhTNF on L-M Cells.
Following incubations at 37°Cfor 0-24 h with or without
rhTNF but in the absence of any hydroxyl radical scavenger, as
shown in Fig. 1, cytotoxic assay by the dye uptake method
showed no significant cell death at 0-15 h of incubation with
rhTNF, but increasingly high levels of cell death at 18, 21, and
24 h of incubation periods.
Influence of DMSO Concentration on Methane Formation.
The amount of methane produced in 18 h of L-M cell incuba
tion after addition of DMSO at concentrations of 0-600 mM is
shown in Fig. 2. In the absence of rhTNF, methane production
reached a plateau at 100 mM DMSO, with no further increase
at higher concentrations. Additional methane production ap
parently induced by the presence of rhTNF (100 U/ml), as
represented by the shaded area in Fig. 2, reached a similar
plateau at 400 mM DMSO, which was therefore the DMSO
concentration utilized for all subsequent determinations of
methane production.
BY TNF
Time Course of rhTNF-induced Methane Production. The
time course of methane production in L-M cell incubation with
and without rhTNF is shown in Fig. 3. In the absence of rhTNF,
the cumulative methane production (O) increased during the
first 21 h of incubation, and apparently reached a plateau at
that time. Methane production attributable to the presence of
rhTNF in similar incubation, represented by the shaded portion
in Fig. 3 as the difference between the amounts found in
incubation with (•)and without rhTNF, increased throughout
the entire investigated incubation time of 24 h. A significant
difference of 15.5 ±5.3 pmol/2 x IO7 cells was observed at 18
h, with further increases at 21 and 24 h.
Influence of rhTNF Concentration on Methane Production.
Methane production amounts as measured after 18 h of L-M
cell incubation were clearly dependent on the rhTNF concen
tration throughout the range investigated, as shown in Fig. 4.
The values with rhTNF at 1, 10, 100, and 1000 U/ml were,
respectively, 2.0, 2.4,2.8, and 3.2 times that found after similar
incubation with no rhTNF.
Effect of Bipyridine on Methane Production. As shown in Fig.
5, the presence of bipyridine during L-M cell incubation clearly
resulted in a decrease in the difference between the methane
production levels with and without rhTNF (shaded bars). In the
absence of bipyridine, methane production with rhTNF was
higher than that without by 37.0 ±9.1 pmol/2 x IO7 cells. In
its presence (60 /¿M),
this difference was effectively reduced (P
< 0.05) to 17.5 ±2.9 pmol/2 x IO7 cells. This effect, by a
known inhibitor of hydroxyl radical production, clearly indi
cates that the increased level of methane production observed
in the presence of rhTNF throughout this study is actually
attributable to an increased level of hydroxyl radical production.
Effect of III 2F3 on Hydroxyl Radical Production. The inhib
iting effect of the monoclonal IgG antibody HI 2F3 on rhTNF
receptor binding and cytotoxicity to L-M cells is shown in
Table 2. Its effect on hydroxyl radical production by L-M cells
in the presence of rhTNF is indicated in Fig. 6. Without III
2F3, the addition of rhTNF at 100 U/ml resulted in an apparent
doubling of methane production during L-M cell incubation,
from 22.0 ±4.4 (bar a) to 48.0 ±3.8 (bar b) pmol/2 x IO7
cells. With HI 2F3 (10 Mg/ml), this increase was completely
suppressed (bar c). In contrast, as shown by bar d, mouse IgG
(10 iig/ml) apparently had no effect on the heightening of
methane production by the presence of rhTNF.
It was further found that methane production in incubation
of the cell-free medium and DMSO with rhTNF (100 U/ml)
was no higher than that without rhTNF, as shown by bars fand
Table I Effect of reactive oxygen scavengers on L-M cell susceptibility to rhTNF
L-M cells (1 x IO4cells/200 jil) were added to wells of 96-well microculture plate, and incubated at 37°Cfor 4 h. Medium was aspirated and both rhTNF (200 U/
ml, 100 i¡\)and one reactive oxygen scavenger (100 ^1) in EMEM were added, followed by incubation at 37*C for 24 h in 5% CO2. Cytotoxicity was assessed by dye
uptake method as described previously.
cyto
(%)70.4
toxicity''
AgentNone
sensitivity*
to rhTNF1.00
±1.87C
SOD
Catalase
SOD
catalaseDMSO
+
aniónscavenging
Hydrogen peroxide scavenging
Inhibition of hydroxyl radical
generation
Hydroxyl radical scavenging
Bipyridine
Iron chelating
DETAPACFunctionSuperoxideIron chelatingConcentration1.25
ng/ml
750 units/ml
0.63 Mg/ml +
325 units/ml
400 mM
MM60
60
MMrhTNF
66.3 ±0.70
68.4 ±0.93
66.0
1.3616.6±
±
0.94
0.97
0.940.24
1.24''
44.0 ±0.19'
86.5 ±1.48'Relative
0.631.23
of rhTNF treated cells\
' Calculated as cytotoxicity = 100 x I 1 - OP
OD of rhTNF free cells /'
rhTNF cytotoxicity with agent
Calculated as relative sensitivity =
.
rhTNF cytotoxicity without agent
' Mean ±SE of three separate experiments.
* ' Values significantly (ä,
P < 0.001; ',P< 0.005) different from value with no agent added.
1672
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
HYDROXYL RADICAL PRODUCTION BY TNF
È 60
O
X
e 40
O
Ü 20
0
Q
O
CE
12
O
INCUBATION
1l
24
0.
10-
TIME (hours)
Fig. 1. Time course of cytotoxic activity of rhTNF on L-M cells. L-M cells (1
x IO4 cells/200 i/1) were added to wells of 96 well microculture plate, and
incubated at 37*C for 4 h. Medium was aspirated and 200 «IEMEM containing
rhTNF (100 U/ml) was added, followed by incubation at 37'C for the indicated
periods in 5% CO... Cytotoxicity was assessed by dye uptake method as described
previously. Percentage cytotoxicity was calculated by the formula shown in Table
1. Values are means ±SD of triplicate wells. *, **, values significantly (*, P <
0.005, **, P < 0.001) different from value without rhTNF.
I
H
Ul
INCUBATION
TIME (hours)
Fig. 3. Time course of the production methane by L-M cells incubated with
rhTNF. L-M cells (2 x IO7 cells/2 ml) were incubated at 37'C in medium
containing 400 mM DMSO with (•)or without (O) the addition of rhTNF (100
U/ml). At indicated times, produced methane was assessed by gas chromatogra
phy as described in Fig. 1. No methane was detected when medium containing
only 400 mM DMSO was incubated (•).Shaded area, produced methane induced
by the presence of rhTNF. Values are means ±SE of eight separate preparations.
*, **, values significantly (*,/»<0.05; **,/"< 0.001 ) different from value without
rhTNF.
60
5°
100
200
300
400
DMSO
CONCENTRATION
500
600
(mM)
40
O
Fig. 2. Effect of DMSO on the production of methane by L-M cells incubated
with rhTNF. Each 3.5-ml tube contained 2 X IO7 L-M cells and the indicated
concentration of DMSO with (•)or without (O) 100 U/ml of rhTNF and was
incubated at37'Cforl8h.
The concentration of methane in 200 nI of headspace
o:
o.
gas was measured by gas chromatography and total produced methane was
calculated as described in "Materials and Methods.11 •¿
produced methane when
U
medium containing only DMSO was incubated. Shaded area, produced methane
induced by the presence of rhTNF. Values are mean ±SE of six separate
preparations. *, **, ***, values significantly (*, P < 0.05; **, P < 0.005; ***, P <
0.001) different from value without rhTNF.
III
O
Q
O
30
Z
01
e. respectively, thus tending to preclude the possibility that
rhTNF itself functions enzymatically or catalytically in the
production of hydroxyl radical.
Hydroxyl Radical Production with KYM and HEL Cells. The
effect of rhTNF addition on two other lines of different TNF
sensitivity are shown in Table 3. The results with rhTNFsensitive KYM cells were similar to those described above for
L-M cells. Methane production amounts in the presence and
absence of rhTNF were, respectively, 39.3 ±5.6 and 15.7 ±4.7
pmol/2 x IO7 KYM cells, a difference of approximately 2.5
times. With rhTNF-insensitive HEL cells, on the other hand,
no significant difference in methane production was observed
between incubations with and without rhTNF, which yielded
values of 13.3 ± 1.7 and 13.2 ±3.5 pmol/2 x IO7 cells,
10
100
rhTNF CONCENTRATION
1000
(U/ml)
Fig. 4. Production of methane by L-M cells as a function of rhTNF concen
tration. L-M cells (2 x IO7cells/2 ml) were indicated in medium containing 400
mM DMSO with indicated concentration of rhTNF at 37"C for 18 h. Produced
methane was assessed by gas chromatography, as described in Fig. 1. Values are
mean ±SE of eight separate preparations.
TNF, on the other hand, reportedly stimulates the expulsion
of Superoxide aniónby neutrophils (28), and hydrogen peroxide
by macrophages (29). Its cytotoxic effect also reportedly is
suppressed by hydroxyl radical scavengers (12-14), and by iron
chelator, an inhibitor of hydroxyl radical production (13, 14),
thus suggesting that hydroxyl radical production is involved in
the mechanism of TNF cytotoxicity.
The present study provides the first quantitative evidence of
respectively.
a correspondence between hydroxyl radical production and
TNF concentration, cell sensitivity, and cytotoxic effect, by the
measurement of methane production from DMSO during in
DISCUSSION
cubation of three cell lines under various degrees of hydroxyl
The hydroxyl radical is a highly potent reactive oxygen known
radical suppression and rhTNF concentration.
to function as a direct mediator in the cytolytic effects of
We first applied this method to investigation of the kinetics
anticancer quinone (24), NK cells (25), neutrophils (26), and of hydroxyl radical production in incubation of TNF sensitive
L-M cells. The results show that the increse in hydroxyl radical
radiological exposure (27).
1673
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
HYDROXYL RADICAL PRODUCTION BY TNF
Table 3 rhTNF-induced methane production by KYM and HEL cells
8^ 0£2g£
T--T-•.--TT•--_
70ì„"
Cytotoxicity was determined by the dye uptake method previously described.
Methane production per 2x10' cells in 18 h was assessed by gas chromatography
as described in Fig. 1.
methane
(pmol/2 x IO7
of
CellL-MKYMHELrhTNF(U/ml)010001000100Cytotoxicity"(%)070.4076.000Number
experimentsg8688gProduced
h)22.0cells/18
±4.4*-,P<
0.00148.0
3.8—15.7
±
60C-si^
50^Q1
4°Og
±4.8—,P<
0.00139.3
5.6—13.2
±
30gLUK
±3.5^NSC13.3±
1.7—
" Calculated from the formula shown in Table 1.
* Mean ±SE of indicated number of experiments.
c NS, not significant.
10LU0'
ab
cd
Fig. 5. Effect of iron chelator bipyridine on the production of methane by I .M cells incubated with rhTNF. L-M cells (2 x 10' cells/2 ml) were incubated for
18 h at 37'C in medium containing 400 mM DMSO with (b,d) or without (a,c)
rhTNF (100 U/ml), in the presence (c,d) or absence (a,b) of bipyridine (60 JIM).
Produced methane was assessed by gas chromatography as described in Fig. 1.
Values are mean ±SE of six separate preparations.
Table 2 Effect ofanti-rhTNF monoclonal antibody III 2F3 on cytotoxicity and
cell surface binding of rhTNF
Cytotoxicity was determined by the dye uptake method previously described.
Total and nonspecific binding of [125I]rhTNF were assessed as follows. L-M cells
(5 x 10s cells/500 n\ binding medium) were incubated for 2 h at 4'C with [12!I]
TNF (5 nin, 586 U/ml) in the presence or absence of unlabeled rhTNF (5 JIM).
The cells were then washed, followed by elution with 0.5 M NaOH and determi
nation of the surface bound radioactivity.
bindingCytotoxicity"
TreatmentNone
(mean ±SE,
cpm)78.2
%)
rhTNF
(mean ±SE,
cpm)
Nonspecific binding
of [125I]rhTNF
(mean ±SE,
±2.5
2898 103.2
266 ±1.7
III 2F3*
1.2 ±4.3
255 ±5.3
252 ±4.2
IgG*"
Mouse 3.5:d
76.5 ±3.1
2686±
±87.6
261 ±
Calcula« from the formula shown in Table 1.
* 10 ng/lC «U
rhTNF-TTTTrhTNF of IgG solution was added simultaneously with
or l'"I]TNF.60
n
50gì408S
1
ÕÎ
30u
11^
20IUS
,00Total
Fig. 6. Effect of anti-rhTNF monoclonal antibody III 2F3 on the production
of methane by L-M cells incubated with rhTNF. L-M cells (2 x IO7 cells/2 ml)
were incubated at 37*C for 18 h in medium containing 400 mM DMSO alone (a),
with rhTNF at 100 U/ml (¿>),
with rhTNF at 100 U/ml and III 2F3 at 10 ^g/ml
(c), or with rhTNF at 100 U/ml and mouse IgG at 10 jig/ml (d). L-M cell-free
medium (2 ml) containing 400 mM DMSO and no rhTNF (e) or rhTNF (/) at
100 U/ml was incubated at 37°Cfor 18 h. Produced methane was assessed by
gas chromatography as described in Fig. 1.
production effected by the presence of rhTNF, in terms of the
difference between its production with and without rhTNF, is
dependent on both incubation time and rhTNF concentration.
The substantial hydroxyl radical production by cells in the
absence of rhTNF, as observed in this and the subsequent
experiments of the present study, may be attributed to mito
chondria! respiration and the other redox known to occur in
cells and thus generate hydroxyl radicals, and is in accord with
similar findings by other investigators for phagocytes and Erlich
tumor cells in vitro in the presence of DMSO alone (19, 24).
The next series of experiments were performed to determine
whether TNF is directly involved in the reaction yielding hy
droxyl radical in the manner of anticancer quinone and menadione, which are known to function as electron acceptors and
thus produce hydroxyl radicals by reduction of oxygen mole
cules (24, 30). The results tended to preclude this possibility,
and show rather that TNF promotes hydroxyl radical produc
tion through stimulation of intracellular production, as: (a) no
suppression of rhTNF cytotoxicity occurred in L-M cell incu
bations with extracellular reactive oxygen suppressed by SOD,
catalase, or DETAPAC; (b) no observable increase of hydroxyl
radical production during incubation of cell-free medium was
effected by the addition of rhTNF; and (c) the rhTNF stimula
tion of hydroxyl radical production was apparently suppressed
completely by the addition of the monoclonal antibody IH 2F3,
which was shown to block rhTNF receptor binding and com
pletely neutralize rhTNF cytotoxicity.
The correlation between TNF cytotoxic effect and hydroxyl
radical production was further shown by the L-M cell incuba
tion in the presence of rhTNF and bipyridine. In the presence
of bipyridine, with the resulting suppression of rhTNF-stimulated hydroxyl radical production, the cytotoxic effect of rhTNF
on the L-M cells was inhibited significantly, suggesting that the
stimulation of hydroxyl radical production is an important part
of the mechanism of rhTNF cytotoxicity.
Comparison of the observed time course of rhTNF cytotox
icity in the absence of hydroxyl radical scavenger with that of
the hydroxyl radical production induced by rhTNF indicates
that the cytotoxic effect and the increased hydroxyl radical
production are approximately concurrent. The observed inhi
bition of rhTNF cytotoxicity by DMSO and bipyridine, fur
thermore, suggests that hydroxyl radical production is rather a
cause than a result of cell death.
The investigation with KYM and HEL cell lines also showed
a correlation between rhTNF sensitivity and hydroxyl radical
production. In incubation of TNF-sensitive KYM cells, the
addition of rhTNF resulted in an increase in hydroxyl radical
1674
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
HYDROXYL RADICAL PRODUCTION
production levels similar to that found for L-M cells. In the
incubation of TNF-insensitive HEL cells under the same con
ditions, however, no observable increase in hydroxyl radical
production resulted from the addition of rhTNF.
The results of this study thus clearly indicate that rhTNF
stimulates production of the hydroxyl radical by rhTNF-sensitive cells, and that this radical plays an important role in the
overall mechanism of rhTNF cytotoxicity. It is not yet possible
to provide a definitive description of the mechanism of the
stimulated production or the precise role of the radical in the
overall mechanism of rhTNF cytotoxicity, but a partial expla
nation may be attempted in the light of the present study and
related findings.
Various recent reports have shown that TNF cytotoxicity is
effectively suppressed by a mitochondrial inhibitor (11, 12) and
by inhibitors of arachide nie acid metabolism (13, 14). More
over, both the mitochondrial electron-transport chain (31) and
the arachidonic acid cascade (32) are known to generate the
hydroxyl radical. It is therefore possible that the rhTNF-stimulated increase in hydroxyl radical production as observed in
the present study may be caused by activation of either or both
of these metabolic systems.
The hydroxyl radical, on the other hand, is known to induce
lipid peroxidation (33) and to cause activation of lysosomal
enzyme by its induction of lipid peroxidation in lysosomal
membrane (34), and also to directly effect DNA chain fragmen
tation (35). Lipid peroxidation (13), lysosomal enzyme activa
tion (12), and DNA fragmentation (36) have all been reported
to occur in cells under stimulation by TNF. It would thus
appear reasonable to suggest that the hydroxyl radical induced
by TNF could be the actual mediator of these intracellular
reactions.
ACKNOWLEDGMENTS
We wish to thank Orville M. Stever for his help in preparation of
this manuscript.
REFERENCES
1. Carswell, E. A., Old, L. J., Rüssel,R. L., Green, M., and Williamson, B. An
endotoxin induced serum factor that causes necrosis of tumors. Proc. Nati.
Acad. Sci. USA, 72: 3600-3670, 1975.
2. Matthews, N. Tumor necrosis factor from the rabbit. II. Production by
monocytes. Br. J. Cancer, 38: 310-315, 1978.
3. Niitsu, Y., Watanabe, N., Sone, H., Yamauchi, N., and Urushizaki, I.
Mechanism of the cytotoxic effect of tumor necrosis factor. Gann, 76:11931197,1985.
4. Niitsu, Y., Watanabe, N., Umeno, H., Sone, H., Yamauchi, N., Maeda, M.,
and Urushizaki, I. Synergistic effects of recombinant human tumor necrosis
factor and hyperthermia on in vitro cytotoxicity and artificial metastasis.
Cancer Res., 48:654-657, 1988.
5. Watanabe, N., Niitsu, Y., Umeno, H., Sone, H., Neda, H., Yamauchi, N.,
Maeda, M., and Urushizaki, I. Synergistic cytotoxic and anti-tumor effects
of recombinant human tumor necrosis factor and hyperthermia. Cancer Res.,
48:650-653, 1988.
6. Watanabe, N., Niitsu, Y., Umeno, H., Kuriyama, H., Neda, H., Yamauchi,
N., Maeda, M., and Urushizaki, I. Toxic effect of tumor necrosis factor on
tumor vasculature in mice. Cancer Res., 48: 2179-2183, 1988.
7. Watanabe, N., Niitsu, Y., Sone, H., Neda, H., Yamauchi, N., Umetsu, T.,
and Urushizaki, I. Anti tumor effect of tumor necrosis factor against primarily
cultured human cancer cells. Gann, 76: 1115-1119, 1985.
8. Helson, L., Green, S., Carswell, E. A., and Old, L. J. Effect of tumor necrosis
factor on cultured human melanoma cells. Nature (Lond.), 25«:791-732,
1975.
9. Matthews, N., and Watkins, J. F. Tumor necrosis factor from the rabbit I.
Mode of action, specificity and physiochemical properties. Br. J. Cancer, 38:
302-319, 1978.
BY TNF
10. Ruff, M. R., and Gifford, G. E. Tumor necrosis factor. In: E. Pick (ed.),
Lymphokines, Vol. 2, pp. 235-302. New York: Academic Press, Inc., 1981.
11. Kull, F. C., Jr., and Cuatrecasas, P. Possible requirement of internalization
in the mechanism of in vitro cytotoxicity in tumor necrosis serum. Cancer
Res., 41:4885-4890, 1981.
12. Watanabe, N., Niitsu, Y., Neda, H., Sone, H., Yamauchi, N., Maeda, M.,
and Urushizaki, I. Cytocidal mechanisms of TNF: effects of lysosomal
enzyme and hydroxyl radical inhibition on cytotoxicity. Immunopharmacol.
Immunotoxicol., 10: 109-116, 1988.
13. Suffys, F., Beyaert, R., Van Roy, F., and Fiers, W. Reduced tumor necrosis
factor-induced cytotoxicity by inhibitors of the arachidonic acid metabolism.
Biochem. Biophys. Res. Commun., 149: 735-743, 1987.
14. Matthews, N., Neale, M. L., Jackson, S. K., and Stark, J. M. Tumor cell
killing by tumor necrosis factor: inhibition by anaerobic conditions, free
radical scavengers and inhibitors of arachidonate metabolism. Immunology,
62:153-155, 1987.
15. Shirai, T., Yamaguchi, H., Ito, H., Toda, C. W., and Wallace, R. B. Cloning
and expression in Escherichia coli of the gene for human tumor necrosis
factor. Nature (Lond.), 313: 803-806, 1985.
16. Yamazaki, S., Onishi, E., Enami, K., Natori, K., Kohase, N., Sakamoto, H.,
Tanouchi, M., and Hayashi, H. Proposal of standardized methods and
reference for assaying recombinant human tumor necrosis factor. Jpn. J.
Med. Sci. Biol., 39: 105-118, 1986.
17. Bolton, A. E., and Hunter, W. M. The labeling of proteins to high specific
radioactivities by conjugation of a '"I-containing acylating agent: application
to the radioimmunoassay. Biochem. J., 133: 529-539, 1973.
18. Watanabe, N., Kuriyama, H., Neda, H., Yamauchi, N., Maeda, M., and
Niitsu, Y. Continuous internalization of TNF receptors in a human myosarcoma cell line. J. Biol. Chem., 263: 10262-10266, 1988.
19. Repine, J. E., Ealon, J. V., Anders, N. W., Hoidal, J. R., and Fox, R. B.
Generation of hydroxyl radical by enzymes, chemicals and human phagocytes
in vitro: detection with the anti-inflammatory agent, dimethyl sulfoxide. J.
Clin. Invest., 64: 1642-1651, 1979.
20. Mello-Filho, A. C., Hoffman, E., and Meneghini, R. Cell killing and DNA
damage by hydrogen peroxide are mediated by intracellular iron. Biochem.
J., 218: 273-275, 1984.
21. Michels«in.A. M., and Fugel, K. Cell penetration by exogenous Superoxide
dismutase. Acta Physiol. Scand. Suppl., 492: 67-81, 1980.
22. Petkau, A., Kelly, K., Chelack, W., and Barefoot, C. Protective effect of
Superoxide dismutase on erythrocytes of \ irradiated mice. Biochem. Bio
phys. Res. Commun., 70:452-458, 1976.
23. Arslan, P., Divirgilio, F., Beltrame, M., Toien, R., and Pozzan, T. Cytosolic
(.V' homeostasis in Ehrlich and Yoshida carcinomas. J. Biol. Chem., 260:
2719-2727, 1985.
24. Doroshow, J. H. Role of hydrogen peroxide and hydroxyl radical formation
in the killing of Ehrlich tumor cells by anticancer quiñones.Proc. Nati. Acad.
Sci. USA, 83:4515-4518, 1985.
25. Duwe, A. K., Werkmeist, J., Roder, J. C., Lauzon, R., and Payne, U. Natural
killer cell-mediated lysis involves an hydroxyl radical-dependent step. J.
Immunol., 134: 2367-2644,1985.
26. Fantone, J. C., and Ward, P. A. A review: role of oxygen-derived free radicals
and metabolites in leucocyte-dependent inflammatory reactions. Am. J. Fa
thol., 107: 395-418, 1982.
27. Repine, J. B., Pfenninger, O. W., Talmage, D. W., Berger, E. M., and
Pettijohn, D. E. Dimethyl sulfoxide prevents DNA nicking mediated by
ionizing radiation or iron/hydrogen peroxide-generated hydroxyl radical.
Proc. Nati. Acad. Sci. USA, 78:1001-1003, 1981.
28. Tsujimoto, M., Yokota, S., Vilcek, J., and Weissmann, G. Tumor necrosis
factor provokes Superoxide anión generation from neutrophils. Biochem.
Biophys. Res. Commun., 137: 1094-1100, 1986.
29. Hoffman, M., and Weinberg, J. B. Tumor necrosis factor-a induced increased
hydrogen peroxide production and Fc receptor expression, but not increased
la antigen expression by peritoneal macrophages. J. Leukocyte Biol., 42:
704-707, 1987.
30. Thor, H., Smith, M. T., Hartzel, P., Belluino. G., Jewel, S. A., and Orrenius,
S. The metabolism of menadione (2-methyl-l,4-naphtholquinone)by
isolated
hepatocytes. J. Biol. Chem., 257:12419-12425, 1982.
31. Cadenas, E., and Boveris, A. Enhancement of hydrogen formation by protophores and ionophores in antimycin-supplemented mitochondria. Biochem.
J., i««:31-37,1980.
32. Kuehl, F. A., Jr., Humes, J. L., Ham, E. A., Eagen, R. W., and Dougherty,
H. W. Inflammation: the role of peroxide-derived product. Adv. Prostaglandin Thromboxane Res., 6: 77-86, 1980.
33. Girotti, A. W., and Thomas, J. P. Damaging effect of oxygen radical on
reseated erythrocyte ghosts. J. Biol. Chem., 259: 1744-1752, 1984.
34. Tong-Mak, I., Misra, H. P., and Weglicki, W. B. Temporal relationship of
free radical-induced lipid peroxidation and loss of latent enzyme activity in
highly enriched hepatic lysosomes. J. Biol. Chem., 25«:13733-13737,1983.
35. Sugioka, K., Nakano, M., Tero-Kubota, S., and Ikegami, Y. Generation of
hydroxyl radicals during the enzymatic reductions of the Fé3*
ADP-EDTA
systems. Biochem. Biophys. Acta, 75J.-411-421, 1983.
36. Dealtry, G. B., Naylor, M. S., Fiers, W., and Balkwill, F. H. DNA fragmen
tation and cytotoxicity caused by tumor necrosis factor is enhanced by
interferon-gamma. Eur. J. Immunol., 17: 689-693, 1987.
1675
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.
Intracellular Hydroxyl Radical Production Induced by
Recombinant Human Tumor Necrosis Factor and Its Implication
in the Killing of Tumor Cells in Vitro
Naofumi Yamauchi, Hiroshi Kuriyama, Naoki Watanabe, et al.
Cancer Res 1989;49:1671-1675.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/49/7/1671
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at [email protected].
To request permission to re-use all or part of this article, contact the AACR Publications
Department at [email protected].
Downloaded from cancerres.aacrjournals.org on June 17, 2017. © 1989 American Association for Cancer Research.