European Journal o)" Pharmacology - Em'ironmental Toxicolot,w and Pharmacology Section, 228 (1992) 77-83
77
© 1992 Elsevier Science Publishers B.V. All rights reserved 0926-6917/92/$(15.00
EJPTOX 40011
Multiple mechanisms for doxorubicin cytotoxicity on glomerular epithelial cells 'in vitro' G i a n M a r c o G h i g g e r i , R o b e r t a Bertelli, F a b r i z i o Ginevri, R o b e r t a Oleggini, P a o l a Altieri, A n t o n e l l a Trivelli a n d R o s a n n a G u s m a n o Department ~>t'Nephrology, G. Gaslini Institute, Genoa, ltal.v
Received 16 September 1991, revised MS received 3 March 1992, accepted 1(1March 1992
This study was planned to define the metabolic pathways for frec radical production by isolated glomeruli and glomerular epithelial cells in vitro after exposure to cytotoxic doses of doxorubicin. A net increment in glomerular superoxide anion (0 2) synthesis was observed at doxorubicin doses between 10 and 30 p.g/ml, a drug level which also induced a parallel increment in uric acid synthesis. Since the synthesis of O-; with production of uric acid implies an activity of xanthine oxidase, a few experiments were performed with glomeruli which had been deprived of xanthine oxidase activity. In this case doxorubicin-inducible O; and uric acid synthesis by glomeruli was practically nil. A similar stimulatory cffcct of O~ synthesis was induced by doxorubicin on glomerular epithelial cells and also in this case O.; synthesis was suppressed by pre-treating cells with deoxyconformicin, a selective inhibitor of adenosine deaminase. Finally, equimolar amounts of the drug were equally cytotoxic even when kept constantly outside the cell by a stable linkage with an agarose macroporous bed. In summary, these data demonstrate that O~ is generated by isolated glomeruli and glomerular epithelial cells 'in vitro' when exposed to cytotoxic amounts of doxorubicin and that purine degradation to uric acid furnish the metabolic pathways for glomerular O; generation. However, doxorubicin is comparably cytotoxic on glomerular epithelial cells from outside cells thus suggesting that also a membrane perturbation may activate the series of events leading to cell injury. Doxorubicin; Nephrosis (experimental); Supcroxide anion; Renal toxicity
1. Introduction
Doxorubicin is an anthracycline antibiotic used widely in antineoplastic therapy, which exerts a cytotoxic effect on several organs such as the heart and the kidney. In rats, doxorubicin is nephrotoxic and responsible for the development of a nephrotic syndrome which is the animal model of human minimal change nephropathy (Bertani et al., 1982). Several mechanisms have been identified as potential effectors of the cytotoxic effect of doxorubicin on neoplastic cell lines 'in vitro' (Bates and Winterbourn, 1982; Doroshow, 1983, 1986; Gutterridge et al., 1984; Lee et al., 1989), but little is currently known on their involvement as determinants of the renal damage. Owing to its hydrophobic nature, doxorubicin reacts with structural membranes and induced membrane perturbation, while the planar hydroxylated anthraquinone chromophore exhibits strong affinity for D N A (Oth et al., 1987; Goor-
Correspondence to: G.M. Ghiggeri, Laboratorio di Nefrologia, G. Gaslini Institute, Largo Gaslini 3, 16148 Genoa, Italy.
maghtigh et al., 1990). Moreover, doxorubicin exerts oxidative damage on cell structure through a direct reduction to a semiquinone radical or by generating the superoxide anion (O~) via enzymatic systems. With N A D P H as an electron donor, doxorubicin is in fact reduced by NADPH-cytochrome P-450 reductase to a semiquinone structure with formation of O~ (Bachur et al., 1979; Davies et al., 1983). Superoxide may also derive from purine catabolism, specifically from the conversion of hypoxanthine into uric acid via the XO system (Ginevri et al., 199(I; Ghiggeri et al., 1991). On the whole, the available data suggest that the oxidative potential of doxorubicin is the main mechanism by which the drug is cytotoxic 'in vitro' and in vivo (Myers et al., 1977). In this paper we investigated the metabolic pathways for O~ generation by doxorubicin in the presence of glomeruli and of glomerular epithelial cells 'in vitro'. For this reason we employed cellular models in which the catabolism of purines was blocked by selective inhibitors of xanthine oxidase or of adenosine deaminase. The implication of the intracellular oxidative pathways for doxorubicin cytotoxicity on glomerufar epithelial cells was also evaluated by devising a
78 simple model in which doxorubicin was chemically linked to a macroporous substance such as agarose, which prevented the intracellular entrance of the drug.
2.3 Cell culture
Doxorubicin was a gift from Farmitalia (Carlo Erba, Milan, ltaly); Roswell Park Memorial Institute medium (RPMI) 1640, penicillin, streptomycin, L-glutamine and Trypan Blue were from Flow Laboratories (Milan, Italy); bovine type I collagen (Vitrogen) was from Collagen Corporation (Palo Alto, CA, USA); 1,3-dimethyl2-thiourca and phenylmethanesulfonylfluoride were from Aldrich (Steinheim, Germany); Reacti-gel 6 × was from Pierce (Oud Beijerland, Netherlands); [ 3H]thymidine (5 C i / m m o l ) and [ 3sS]methionine ( 1000 C i / m m o l ) were from A m e r s h a m (Little Chalfont, A m e r s h a m , England); superoxide dismutase (EC 1.15.1.1), cytochrome C, deoxy-coformicin, xanthine oxidase (EC 1.1.3.22) and adenosine deaminase (EC 2.4.2.11 were from Sigma (St. Louis, MO, USA); and reverse-phase chromatography columns RP-18 were from Waters (Milford, MA, USA).
Glomerular epithelial cells were cultured starting from purified glomeruli obtained from Sprague-Dawley rats (125-150 g) which were killed with a lethal dose of sodium pentobarbital. Glomeruli isolated by the sieving technique (see above) under sterile conditions were resuspended in RPMI medium supplemented with I(1% fetal calf serum, 2 m m o l / l sodium bicarbonate, 2 m m o l / l glutamine, 100 U / m l penicillin, 0.(t1% p,g/ml streptomycin and I).18 U / m l insulin. They were then plated onto 25 cm 2 Falcon tissue culture flasks, previously coated with a thin film of collagen and mainrained at 37°C in humidified 5% COo atmosphere. Cultured cells were recognized as epithelial from: (11 typical cobblestone-like morphology and polygonal shape; (2) early outgrowth; (3) absence of factor Vlll as determined by a double immunofluorescent technique; (4) positivity for cytokeratins (Kreisberg and Karnovsky, 1983). To avoid mesangial cell contamination, only cells obtained from the primary culture (1-2 week duration) were used. Mesangial cells were recognized from epithelial cells based on the following criteria: (1) morphology; (2) negative staining by indirect immunofluorescence for cytokeratin; (3) contractile response to angiotensin lI.
2.2. ls'olation of glomertdi
2.4. Superoxide release assay
Glomeruli were isolated by a differential sieving technique according to standard methods (Salant et al., 1980) from Sprague-Dawley rats weighing 150-180 g which were anesthetized with ketamine (Ketalar). Rats were fed a normoprotein standard diet (20% casein), and in some cases they were fed a normoprotein diet supplemented with 0.7% sodium tungstate. This diet did not induce any effect on the animal growth nor on other clinical parameters in these rats. After clamping the aorta and cutting off the renal veins, the kidneys were perfused with ice-cold (I.9% NaCI containing 1 mM ethyldiaminetetraacetate (EDTA), 0.2 m m o l / l phenylmethanesulfonylfluoride (PMSF), 1 m m o l / 1 dithiothreitol (DTT), 5 / , g / m [ trypsin inhibitor and 2 U / m l kallikrein (Trasylol). The dissected cortex was then minced with a razor blade in PBS containing the same protease inhibitors and gently squeezed through a stainless steel sieve (250 #m). The resulting suspension was filtered through 150 /xm and 75 p,m sieves and washed three times in Hanks' balanced salt solution (HBSS). Glomeruli were centrifuged at 150 × g for 5 min and resuspended in a small volume of 5(/ mmol/1 Tris-HCl buffer, pH 7.5, containing protease inhibitors. After disruption of cellular structures by sonication, protein content was determined using the Coomassie assay (Read and Northcate, 1981).
Glomerular sonicates (1 m g / m l ) from normal and tungsten-fed rats and sonicated glomerular epithelial cells (105 cells) were incubated with increasing amounts of doxorubicin (3-30 /xg/ml) in sterile phosphate buffer solution (pH 7.4) at 37°C for 2 h. Treatment of cellular extracts with deoxy-coformicin ( 4 / , m o l / l ) was performed for 15 min at 37°C. Release of O~ was estimated by classical assays based on the superoxidc dismutase (SOD)-inhibitable reduction of cytochrome c (Fridovich, 1985). A reaction buffer containing 1 mmol/CaC12, 2 mmol/1 glucose and 100 /xmol/l cytochromc c dissolved in PBS (pH 7.4) was added to each well. For each p a r a m e t e r assessed, half of the wells were incubated with 100 /,tg SOD and half without. After incubation at 37°C for 80 min, reactions were terminated by addition of 2 mM N-ethylmaleimide. Reduced cytochrome c was determined spectrophotometrically at 546 nm using E = 21.2 m m o l / l cm ~ as an extinction coefficient; results were corrected for the protein content in mg. Blank samples were conducted in cytochrome c and doxorubicin without glomeruli and in cytochrome c and glomeruli without doxorubicin. The amount of O ; was calculated by subtracting the amount of reduced cytochrome c in the presence of SOD from total (SOD non inhibited) reduced cytochrome c.
2. Materials and methods 2.1. Materials'
79
2.5. Assays for adenosine deaminase and xanthine oxidase Adenosine deaminase activity in isolated glomeruli was assayed spectrophotometrically at 265 nm, using 0.1 m o l / l adenosine in 0.1 Tris-HCl buffer, pH 7.4 as a substrate. Xanthine oxidase activity was assayed spectrophotometrically at 293 nm, using 0.9 m m o l / l hypoxanthine in 0.1 M Tris-HCI buffer, pH 7.4 as a substrate. All assays were performed at 37°C in 1 ml reaction mixtures. The enzyme activity unit (EU) is the amount of enzyme required to convert 1 /xmol of xanthine or adenosine to uric acid/rain. Specific activity is defined as U / r a g of protein.
for 02 generation from purines by deoxy-coformicin (4 # t o o l / l ) which was incubated for 1 h at 37°C and the cells were then washed twice and resuspended in fresh RPMI. On completion of drug exposure, cells were washed twice and resuspended in fresh medium.
2. 9. Cell viabifity For m e a s u r e m e n t of cell viability, cultures were harvested with 0.05% trypsin-0.02% EDTA, stained with 0.05% Trypan Blue (Shanne et al., 1979) and counted in a hemocytometer. The Trypan Blue assay was performed 1 and 6 days after doxorubicin treatment.
2.6. Coupfing of doxorubicin
2.10. Thyrnidine incorporation
Doxorubicin was coupled with Reacti-gel 6 × , which consists of 6% cross-linked agarose beads ( 4 0 - 2 1 0 / x m in diameter) derivatized with 1,1'-carbonyldiimidazole. In a typical experiment, 3 × 10 s mol of doxorubicin were allowed to react per mg of activated agarose for 25-72 h in 0.1 M borate buffer at p H 8.0 and 4°C. Unreacted imidazole carbonate groups were eliminated with excess hydroxylamine (Tritton and Yee, 1982). Under these conditions about 3 0 - 4 0 % of doxorubicin was attached to the macroporous bed while the unbound fraction was removed by extensive washings (1 week) in water. The washings were continued until no free doxorubicin was found in the supernatant and the resulting immobilized doxorubicin was stored at 4°C in the dark. The stability of doxorubicin coupled with agarose was studied by measuring the rate of release of doxorubicin in phosphate and in R P M I at 37°C for several days. Doxorubicin was determined by H P L C using a reverse phase column (Bondapak C-18) and fluorescence detection (Broggini et al., 1984) and its cellular uptake was determined after incubation for 2h.
Thymidine incorporation was evaluated by incubating 104/ml cells with 5 / z C i / m l [3H]thymidine for 24 h at 37°C in a humidified 5% CO 2 atmosphere. After extensive washings in RPMI, cells were counted in a Packard /3 scintillation counter after being transferred under vacuum to filter p a p e r sheets (Whatman, glass microfibre filters) and solubilized in filter-count scintillation fluid. All analyses were performed 'in triplicate.
2. Z Effect on glomerular epithefial cells of free and agarose-bound doxorubicin
2.11. [~SS]methionine incorporation For [35S]methionine incorporation, cells (104/ml) were incubated with 2 /xCi/ml [3SS]methionine in methionine-free R P M I medium supplemented with 10% dialyzed fetal calf serum for 18 h in a humidified 5% CO 2 atmosphere. The reaction was stopped with ice cold 0.5% Triton X-100-0.05% E D T A with 2 m m o l / l PMSF and 1 retool/1 dithiothreitol (DTT). The precipitate in 10% thrichloracetic acid (TCA) (20 rain room temperature) was filtered under vacuum to filter paper (Whatman, glass microfibre filters) and was solubilized in filter-count scintillation fluid.
2.12. Other methods
Glomerular epithelial cells were transferred to multiwell dishes by plating 105 cells/ml in complete medium, maintained at 37°C in a 5% CO 2 humidified atmosphere. Cells were used for drug experiments 1 day after plating. The effect of doxorubicin on intact glomerular epithelial cells was examined by exposure with the same amounts (7.5 /xg/ml) of free or agarose-bound doxorubicin for 2 h.
The Coomassie binding assay (Read and Northcate, 1981) was employed to evaluate proteins. Uric acid was determined spectrophotometrically with an enzymatic assay based on the conversion of uric acid into allantoine with release of H 2 0 2. Enzymatically produced H 2 0 2 reacts in turn with 3,5-dichloro-2-hydroxyben z e n e sulphonic acid in the p r e s e n c e of 4aminophenazine to give a red quinonine with an absorbance maximum at 520 rim.
2.8. Inhibition experiments"
2.13. Statistical analysis
Doxorubicin cytotoxicity on glomerular epithelial ceils was tentatively inhibited by blocking the pathway
The one way analysis of variance was used in all statistical tests. Results are given as means _+ S.E.
80
3. Results
3.1. Cytotoxic effect lial cells
of doxorubicin on glomerular epithe£
[]
NORMAL
[]
TUNGSTAT E
GLOMERULI
200-
O
As shown in fig. 1, doxorubicin induced a dose related inhibition of [3H]thymidine incorporation into DNA of intact glomerular epithelia which was maximal for concentrations of the drug >_ 15 /~g/ml. Furthermore, exposure of cells to the same levels of drug inhibited protein synthesis ([~SS]methionine incorporation) and reduced cell viability (Trypan Blue).
100-
> O -
÷
--
4-
SOD
SOD
o';
o~
30
15
3.2. Glomerular O~ synthesis Doxorubicin ~glrnL
The glomerular production of 02 in the presence of doxorubicin was evaluated as the SOD-inhibitable reduction of cytochromc c, which is one of the most popular tests to detect 02 in biological fluids. The production of 02 by glomeruli in the presence of doxorubicin was evident for levels of the drug producing cytostasis ( > 15 /zg/ml); at drug levels of 30 /zg/min the SOD-inhibitable amount of reduced cytochrome c was 50 +_ 10 # m o l / m g protein (fig. 2). However, 0 2 was not the only responsible for this effect and the sum of SOD-inhibitable and non-SODinhibitable reduction of cytochrome c in the presence of glomeruli and of doxorubicin was in fact greater by a factor of 3.5 (175 +_ 10 /~mol/mg of protein). Other reactions involving doxorubicin transformation into an unstable radical by glomeru[ar extract would account for it. Since most of the O 2 produced by biological system arises from the transformation of hypoxanthine and xanthine into uric acid via the xanthine oxidase system, glomerular 0 2 synthesis would be paralleled by uric acid formation. Indeed as shown in fig. 3, in the presence of doxorubicin, the glomerular synthesis of
w°t
li00m F o
z
30
Fig. 2. Reduction of cytochrome c by doxorubicin (15-30 ~ g / m l ) incubated for 2 h at 37°C with sonicated glomeruli (l m g / m l ) isolated from normal rats (open bars) and from rats fed a tungsten enriched diet (shaded bars). For each parameter assessed, half the wells were incubated with S O D (100 >g) and half without; O; synthesis was calculated by subtracting the amount of SOD-inhibitable reduction of cytochrome c from total non-SOD-inhibitable reduction. Non-SOD-inhibitable cytochrnme e reduction accounted for about 7 0 ~ of the overall effect. Results of six experiments for each group are given as mean + S.E. * P < 0.01.
uric acid increased in a dose-related manner. A few experiments were performed with glomeruli purified from kidneys of rats which were fed a diet enriched in tungsten, a heavy metal which competes with the intestinal absorption of molybden. Since molybden is essential for xanthine oxidase activity, glomeruli purified from tungsten-enriched diet presented only minimal activity of this enzyme (XO 0.8 _+ 0.1 vs. 5.1 _+ 1.5 mEU mg L). According to the concept that doxorubicin-induced glomerular 0 2 is synthesized with the intervention of xanthine oxidase, the production of O_~ in the presence of glomeruli purified from tungsten-fed
,c_
o
o
(3.
75
75
10 za.
-50
~
~ ~
25
25
x
a < O
5
D
0
3
7
Doxorubicin ¢on¢.
15
75
Pg/ml
Fig. 1. Dose-response inhibition of [~H]thymidine incorporation into DNA, and cell viability (6 days following exposure) of glomerular epithelial cells after incubation with increasing a m o u n t s of doxorubicin. Each experiment was performed in quadruplicate. Values have been normalized to [3HJthymidine incorporation by normal cells ( 100% incorporation).
6
~ lO ~
Doxorubicin
~ 2 ~
ug Iml
0
1
2
3
4
5
e
Time (hour's)
Fig. 3. Production of uric acid from normal glomeruli (1 m g / m l ) incubated for 4 h with increasing amounts of doxorubicin (5-30 p.g/ml). Time-dependent uric acid synthesis was evaluated by incubating glomeruli with constant 15 /~g/ml doxorubicin for various times. Each experiment was performed in quadruplicate. Results are the mean +_S.E.
81
20-
[]
GEC
[]
GEC -f- COFORMICIN
TABLE 1 In vitro cytotoxicity of doxorubicin (15 /zg/ml) on intact cultured glomerular epithelial cell in the presence of deoxy-coformicin (4 IzM) or in its absence. 24 h [3H]thymidine and [35S]methionine incorporation was evaluated after a 2 h exposure of cells to doxorubicin. Cell viability was evaluated 6 days after exposure to doxorubicin by the Trypan Blue dye exclusion test.
r
oi
Treatment
[3H]thymidine (cpm/104 cells)
[35S]methionine (cpm/104 cells)
Cell viability (% of total cells)
None (n = 5) Coformicin alone (n = 4) doxorubicin alone (n = 4) Doxorubicin + coformicin (n=4)
4,000_+500
35,000_+5,000
75
_+0.5
4,100+800
33,000_+1,200
74
+ 1
O :z. -- 4SOD Doxorubicin
SOD
O~
30
~UgImL
Fig. 4. Reduction of cytochrome c by doxorubicin (30 txg/ml) in the presence of cellular extracts from glomerular epithelial cells. Incubation was carried out for 2 h at 37°C. Sonicated glomerular epithelial cells (open bars); cells pretreated with deoxy-coformicin, a specific inhibitor of adenosine deaminase (shaded bars). For 0 2 calculation see Materials and methods. Experiments were performed with six different samples of GEC. Results are the mean _+S.E. * P < 0.01 vs. untreated cells.
rats (e.g., without xanthine oxidase activity) was practically nil. Taken together these data demonstrate that doxorubicin induces the glomerular synthesis of O~ which derives from the conversion of xanthine into uric acid by the enzyme xanthine oxidase.
3.3. 0 2 synthesis by glomerular epithelial cells As shown in fig. 4, reduction of cytochrome c by doxorubicin was also elevated in the presence of glomerular epithelial cell extracts. As already reported for isolated glomeruli, a part of this effect ( = 22%) was suppressed by SOD and represented the effective O;~ synthesis (9.2 ~ m o l / m g protein). Also in the case of glomerular epithelial cells, experiments aimed at blocking the purine degradative pathway leading to hypoxanthine and uric acid were successful in producing a reduction of the oxidative potential of doxorubicin. In fact, in this case, the production of 0 2 was inhibited by deoxy-coformicin which is a specific inhibitor of adenosine deaminase (Whoowdion et al., 1974) and therefore regulates the production of substrates for xanthine oxidase. Treatment with deoxy-coformicin also reduced the cytolytic effect of doxorubicin on cells which is evident as an increase in cell viability after 6 days (table 1)
200 + 100 ~'
7,200 +
325+ 150 "
6,700_+ 210 "
52.2 + 1.2 "
67.3+3h
" P < 0.001 vs. control cells, b p < 0.05 vs. doxorubicin alone.
P-450 reductase. Both mechanisms imply the presence of the drug within the cell. To prove this point doxorubicin was chemically linked to agarose (Tritton and Yee, 1982) and incubated with glomerular epithelial cells in the same experimental conditions as free doxorubicin. Since the agarose-doxorubicin complex is too large to penetrate into cells, the intracellular levels of the drug were nil while more than 70% was detected inside the ceils when the free drug was used (5 vs. < 1 /~g/ml where 1 p~g/ml is the upper limit of detection). In spite of this, free and agarose-bound doxorubicin produced the same cytostatic effect on glomerular epithelial cells corresponding to 50% inhibition of thymidine incorporation with doxorubicin levels of 7.5 ~ g / m l (table 2). This fact suggests that other mechanisms besides O~ and direct oxidation are responsible for doxorubicin cytotoxicity on glomerular epithelial ceils.
TABLE 2 In vitro cytotoxicity of free doxorubicin and agarose-bound doxorubicin on cultured epithelial cells. 24 h [3H]thymidine incorporation was evaluated after exposure of cells to A D R (7.5 txg/ml) for 2 h. Cell viability was evaluated by the Trypan Blue assay 1 and 6 days after 2 h doxorubicin exposure. Treatment
3.4. Cytotoxicity of free and agarose-bound doxorubicin The basic assumption of the experiments so far presented is that any direct cytotoxic effect of doxorubicin by means of O 2 a n d / o r as a semiquinone molecule should require an enzymatic pathway, namely xanthine oxidase activity or conversion by cytochrome
380 "
None (n = 6) Doxorubicin (n = 4) Agarose-doxorubicin (n=4)
% [3H]thymidine incorporation
% cell excluding Trypan Blue
(vs. controls)
Days after treatment
100 43 _+5 ~ 40_+6"
~' P < 0.001 vs. control ceils.
I st
6th
99_+0.5 98 _+ 1
84_+6 51 + 2 "
99_+1
65+8
82
4. Discussion This paper was planned to focus on the mechanisms for doxorubicin cytotoxicity on glomerular epithelial cells 'in vitro', with particular emphasis on assessing a possible role of free radicals. This implication was suggested by recent works on the toxicity of doxorubicin on other cell lines (Doroshow, 1983, 1986) and in vivo (Myers et al., 1977; Doroshow et al., 1981; Ghiggeri et al., 1991), which supported the concept that superoxide, peroxide and hydroxyl radicals are responsible for the toxic action of the drug. At least two metabolic pathways could be responsible for free radical generation from doxorubicin. One involves the production of semiquinone intermediates from doxorubicin by intervention of N A D P H , N A D P H - c y t o c h r o m e P-450 reductase or cytochrome P-450 and probably other enzymes such as N A D P H dehydrogenase (Bacher et al., 1979; Davies et al., 1983). The second mechanism is based on a direct stimulatory effect of doxorubicin on O~ generation. In accordance with this idea, the cytotoxicity of several anticancer quinones on Ehrlich tumor cells ',in vitro' was reduced, if not suppressed, by several antioxidant enzymes and chemical scavengers such as catalase and SOD, dimethyl sulfoxide and dimethylurea which act on the hydroxyl free radical and by the iron chelator deferoxamine (Lee et al., 1983). 'In vivo' the cardiotoxicity of A D R was partially inhibited by the administration of cysteine (Doroshow et al., 1991), the thiol component of glutathione, which exerts a key scavenger action within the cell (Faber et al., 1990). The nephrotoxic potential of doxorubicin in SpragueDawley rats, which is responsible for the development of proteinuria similar in many aspects to that found in human minimal change nephropathy, was likewise blunted by selective inhibitors of renal xanthine oxidase (Ginevri et al., 1990), the enzymatic system responsible for the generation of O;, from hypoxanthine (Fridovich, 1978) and by dimethylthiourea, the scavenger of the hydroxyl radical (Ghiggeri et al., 1991). In general, although the data are rather convincing, conclusive evidence for a role of free radicals in the anticancer action of doxorubicin, and more specifically in the cytotoxic effect on renal cells, is lacking. The data reported here reinforce the idea that doxorubicin in the presence of glomeruli and of glomerular epithelial cells 'in culture' induces O~ generation which is partially responsible for its cytotoxicity. They also extend the knowledge on the metabolic pathways responsible for O~ generation, which involve purine catabolism and the final conversion of hypoxanthine into uric acid by the enzyme xanthine oxidase. Two central experiments involving selective blocks of the purine cascade, one of xanthine oxidase by dietary sodium tungstate and the other by the adenosine
deaminase inhibitor deoxy-coformicin, support this conclusion. Sodium tungstate was chosen as a chronic model of xanthine oxidase inhibition since it determines a selective and protracted inhibition of xanthinc oxidase without affecting the cellular uptake of purines. Tungsten competes with molybdenum for its intestinal absorption, and molybdenum-deprived rats present markedly reduced glomerular XO activity since this metal is a structural component of the hydrophobic functional core of the enzyme (Tophan et al., 1988). This model has been extensively used in recent years as an alternative to allopurinol in experimental models of functional block of renal xanthine oxidase, especially in studies on renal ischemia (McKelvey et al., 1988). The data reported here unequivocally demonstrate that O:, is produced by glomeruli and glomerular epithelial cells in the presence of doxorubicin while no O~ is spontaneously synthesized and that purine catabolism via xanthine oxidase represents the metabolic pathway for O; generation. Nevertheless, even in the case of a block of O~ production by tungsten and coformicin, doxorubicin is still able to exert an oxidative potential in the presence of glomeruli or cells. It is reasonable to suspect that, in this case, this effect is directly mediated by transformation of doxorubicin by cytochrome P-450 reductase into an unstable semiquinone. Another important point raised by this study suggests that, besides oxidative mechanisms, doxorubicin has other effects on renal cells. This conclusion is based on the observation that doxorubicin is actively cytostatic and cytotoxic on glomerular epithelia, even when its intracellular levels arc nil, owing to a stable chemical linkage with a macroporous bed which keeps the drug out of the cells during incubation. We suggest that in this case a m e m b r a n e perturbation may be responsible for the toxicity and that doxorubicin may exert its biological activity solely by interaction with the cell surface. The second implication of these data is that doxorubicin exerts a multiple toxic mechanism on glomerular epithelia 'in vitro' and that multiple sites are targets for its action. In accordance with these concepts, other works (Goormaghtigh et at., 1980: Oth et al., 1987) have already demonstrated that doxorubicin induces modifications of the membrane lipid composition, fluidity and permeability of several cell lines 'in vitro' by interaction with membrane phospholipids.
Acknowledgements The authors are indebted to Prof. Giovanni Ccrcignani for having determined glomerular adenosine deaminase and xanthine oxidase contents. Mr. Robert B. Stubinski is acknowledged for his secretarial support. These data were in part presented at the XXIV Annual Meeting of the American Society of Nephrology, Baltimore, November 17-20, 1991.
83 This work was carried out with a grant of the Italian Ministry of Health to the G. Gaslini Institute.
References Bachur, N.R., S.L. Gordon, M.V. Gee and H. Kon, 1979, NADPH cytocbrome P-450 reductase activation of quinone anticancer agents to free radicals, Proc. Natl. Acad. Sci. USA 76, 954. Bates, D.A. and C.C. Winterbourn, 1982, Deoxyribose breakdown by the adriamycin semiquinone and H202: evidence for hydroxyl radical participation, FEBS Lett. 145, 137. Bertani, T., A. Poggi, R. Pozzoni, F. Delaini, G. Sacchi, Y. Thoua, G. Mecca, G. Remuzzi and M.B. Donati, 1992, Adriamycin-induced nephrotic syndrome in rats. Sequence of pathological events, Lab. Invest. 46, 16. Broggini, M., C. Italia, T. Colombo, L. Marmonti and M.G. Donelli, 1984, Activity and distribution of i.v. and oral 4-demetboxydaunorubicin in murine experimental tumors, Cancer Treat Rep. 68. 739. Davies, K.J.A., J.H. Doroshow and P. Hochstein, 1983, Mitochondrial NADH dehydrogenase-catalyzed oxygen radical production by adriamycin and the relative inactivity of 5-iminodaunorubicin, FEBS Lett. 153, 227. Doroshow, J.H., 1983, Anthracycline antibiotic stimulated superoxide, hydrogen peroxide and hydroxyl radical production by NADH dehydrogenase, Canc. Res. 43, 4543. Doroshow, J.H., 1986, Role of hydrogen peroxide and hydroxyl radical formation in the killing of Ehrlich tumor cells by anticancer quinones. Proc. Natl. Acad. Sci USA 83, 4514. Doroshow, J.H., G.Y. l,x)cker, !. lfrim and C.E. Myers, 1981, Prevention of doxorubicin cardiac toxicity in the mouse by N-acetylcysteine, J. Clin. Invest. 68, 1053. Faber, J.L., M.E. Kyle and J.B. Coleman, 1990, Biology of disease. Mechanisms of cell injury by activated oxygen species, Lab Invest. 62, 670. Fridovich, I., 1978. The biology of oxygen radicals, Science 201, 875. Fridovich I., 1985, Cytochrome c, in: Methods for Oxygen Radical Research, ed. R.A. Greenwald (CRC Press, Boca Raton, FL) p. 121. Ghiggeri, G.M.. F. Ginevri, G. Cercignani, R. Oleggini. A. Garberi and R. Gusmano, 1990, Effect of dietary protein restriction on renal purines and related metabolizing enzymes in adriamycin nephrosis in rats. A mechanism for protection against proteinuria involving xanthine oxidase inhibition, Clin. Sci. 79, 647. Ginevri, F.. R. Gusmano, R. Oleggini, S. Acerbo, R. Bertelli, F. Perfumo, G. Cercignani, S. Allegrini, F. D'Allegri and G.M.
Ghiggeri, 1990, Renal purine efflux and xanthine oxidase activity during experimental nephrosis in rats: difference between puromycin aminonucleoside and adriamycin nephrosis, Clin. Sci. 78, 283. Goormaghtigh, E., P. Chatelain, J. Caspers and J.M. Ruysschaert, 1980, Evidence of a specific complex between adriamycin and negatively charged phospholipids, Biochim. Biophys. Acta 597, 1. Gutteridge, J.M.C., G.J. Quinlan and S. Wilkins, 1984, Mitomycin C-induced deoxyribose degradation inhibited by superoxide dismutase, FEBS Lett. 167, 37. Kreisberg, S.I. and M.J. Karnovsky, 1983, Glomerular cells in culture, Kidney Int. 23, 439. Lee, F.Y,F., D.W. Siemann and R.M. Sutherland, 1989, Changes in cellular glutathione content during adriamycin treatment in human ovarian cancer - a possible indicator of chemosensitivity, Br. J. Cancer, 60, 291. McKelvey, T.G., M.E. Hollwarth, D.N. Granger, T.D. Engerson, U, Landler and H.P. Jones, 1988, Mechanisms of conversion of xantbine oxidase in ischemic rat liver and kidney, Am. J, Physiol, 254, G753. Myers, C.E., W.P, McGuire, R.H. Liss. I. llfrim, K. Grotzinger and R.C. Young, 1977, Adriamycin: the role of lipid peroxidation in cardiac toxicity and tumor response, Science 197, 165. Oth, D., M. Begin, P. Bischoff, J.Y. Leroux, G. Mercier and C. Bruneau, 1987, Induction by adriamycin and mitomycin C of modifications in lipid composition, size distribution, membrane fluidity and permeability of cultured RDM4 lymphoma cells. Biochim. Biophys. Acta 900, 198. Read S.M. and D.M. Northcate, 1981, Minimization of variation in the response of different proteins to the Coomassie-Blue dye binding assay for protein, Anal. Biochem. 116, 53. Salant, D,J., C. Darby, C. Couser and W.G. Couser, 1980, Experimental membranous glomerulonephritis in rats. Quantitative studies of glumerular immune deposit formation in isolated glomeruli and whole animals, J. Clin. Invest. 66, 71. Schanne, F.A., A.B. Kane, E.E. Young and J.C. Farler, 1979, Calcium dependence of toxic cells death: a final common pathway, Science 206, 700. Tophan, R.W., M.C. Walker, M.P. Lalish and W. Williams, 1982, Evidence for the participation of intestinal xanthine oxidase in the mucosal processing of iron, Biochemistry 21, 4529. Tritton, T.R. and G. Yee, 1982, The anticancer agent adriamycin can be actively cytotoxic without entering cells. Science 217, 248. Woohwdion, P.W.K., S.M. Lange, L.S. Dohl and L.J. Durhan, 1974, A novel adenosine and ara-A deaminase inhibition, (R)-3-(2-deoxy-/3-D-erythro-pentafuranosyl)-3,6,7,8-tetrahydroimidazo[4,5-6][1,3]diazepin-8-AL, J. Antibiotics 11,641.
77
© 1992 Elsevier Science Publishers B.V. All rights reserved 0926-6917/92/$(15.00
EJPTOX 40011
Multiple mechanisms for doxorubicin cytotoxicity on glomerular epithelial cells 'in vitro' G i a n M a r c o G h i g g e r i , R o b e r t a Bertelli, F a b r i z i o Ginevri, R o b e r t a Oleggini, P a o l a Altieri, A n t o n e l l a Trivelli a n d R o s a n n a G u s m a n o Department ~>t'Nephrology, G. Gaslini Institute, Genoa, ltal.v
Received 16 September 1991, revised MS received 3 March 1992, accepted 1(1March 1992
This study was planned to define the metabolic pathways for frec radical production by isolated glomeruli and glomerular epithelial cells in vitro after exposure to cytotoxic doses of doxorubicin. A net increment in glomerular superoxide anion (0 2) synthesis was observed at doxorubicin doses between 10 and 30 p.g/ml, a drug level which also induced a parallel increment in uric acid synthesis. Since the synthesis of O-; with production of uric acid implies an activity of xanthine oxidase, a few experiments were performed with glomeruli which had been deprived of xanthine oxidase activity. In this case doxorubicin-inducible O; and uric acid synthesis by glomeruli was practically nil. A similar stimulatory cffcct of O~ synthesis was induced by doxorubicin on glomerular epithelial cells and also in this case O.; synthesis was suppressed by pre-treating cells with deoxyconformicin, a selective inhibitor of adenosine deaminase. Finally, equimolar amounts of the drug were equally cytotoxic even when kept constantly outside the cell by a stable linkage with an agarose macroporous bed. In summary, these data demonstrate that O~ is generated by isolated glomeruli and glomerular epithelial cells 'in vitro' when exposed to cytotoxic amounts of doxorubicin and that purine degradation to uric acid furnish the metabolic pathways for glomerular O; generation. However, doxorubicin is comparably cytotoxic on glomerular epithelial cells from outside cells thus suggesting that also a membrane perturbation may activate the series of events leading to cell injury. Doxorubicin; Nephrosis (experimental); Supcroxide anion; Renal toxicity
1. Introduction
Doxorubicin is an anthracycline antibiotic used widely in antineoplastic therapy, which exerts a cytotoxic effect on several organs such as the heart and the kidney. In rats, doxorubicin is nephrotoxic and responsible for the development of a nephrotic syndrome which is the animal model of human minimal change nephropathy (Bertani et al., 1982). Several mechanisms have been identified as potential effectors of the cytotoxic effect of doxorubicin on neoplastic cell lines 'in vitro' (Bates and Winterbourn, 1982; Doroshow, 1983, 1986; Gutterridge et al., 1984; Lee et al., 1989), but little is currently known on their involvement as determinants of the renal damage. Owing to its hydrophobic nature, doxorubicin reacts with structural membranes and induced membrane perturbation, while the planar hydroxylated anthraquinone chromophore exhibits strong affinity for D N A (Oth et al., 1987; Goor-
Correspondence to: G.M. Ghiggeri, Laboratorio di Nefrologia, G. Gaslini Institute, Largo Gaslini 3, 16148 Genoa, Italy.
maghtigh et al., 1990). Moreover, doxorubicin exerts oxidative damage on cell structure through a direct reduction to a semiquinone radical or by generating the superoxide anion (O~) via enzymatic systems. With N A D P H as an electron donor, doxorubicin is in fact reduced by NADPH-cytochrome P-450 reductase to a semiquinone structure with formation of O~ (Bachur et al., 1979; Davies et al., 1983). Superoxide may also derive from purine catabolism, specifically from the conversion of hypoxanthine into uric acid via the XO system (Ginevri et al., 199(I; Ghiggeri et al., 1991). On the whole, the available data suggest that the oxidative potential of doxorubicin is the main mechanism by which the drug is cytotoxic 'in vitro' and in vivo (Myers et al., 1977). In this paper we investigated the metabolic pathways for O~ generation by doxorubicin in the presence of glomeruli and of glomerular epithelial cells 'in vitro'. For this reason we employed cellular models in which the catabolism of purines was blocked by selective inhibitors of xanthine oxidase or of adenosine deaminase. The implication of the intracellular oxidative pathways for doxorubicin cytotoxicity on glomerufar epithelial cells was also evaluated by devising a
78 simple model in which doxorubicin was chemically linked to a macroporous substance such as agarose, which prevented the intracellular entrance of the drug.
2.3 Cell culture
Doxorubicin was a gift from Farmitalia (Carlo Erba, Milan, ltaly); Roswell Park Memorial Institute medium (RPMI) 1640, penicillin, streptomycin, L-glutamine and Trypan Blue were from Flow Laboratories (Milan, Italy); bovine type I collagen (Vitrogen) was from Collagen Corporation (Palo Alto, CA, USA); 1,3-dimethyl2-thiourca and phenylmethanesulfonylfluoride were from Aldrich (Steinheim, Germany); Reacti-gel 6 × was from Pierce (Oud Beijerland, Netherlands); [ 3H]thymidine (5 C i / m m o l ) and [ 3sS]methionine ( 1000 C i / m m o l ) were from A m e r s h a m (Little Chalfont, A m e r s h a m , England); superoxide dismutase (EC 1.15.1.1), cytochrome C, deoxy-coformicin, xanthine oxidase (EC 1.1.3.22) and adenosine deaminase (EC 2.4.2.11 were from Sigma (St. Louis, MO, USA); and reverse-phase chromatography columns RP-18 were from Waters (Milford, MA, USA).
Glomerular epithelial cells were cultured starting from purified glomeruli obtained from Sprague-Dawley rats (125-150 g) which were killed with a lethal dose of sodium pentobarbital. Glomeruli isolated by the sieving technique (see above) under sterile conditions were resuspended in RPMI medium supplemented with I(1% fetal calf serum, 2 m m o l / l sodium bicarbonate, 2 m m o l / l glutamine, 100 U / m l penicillin, 0.(t1% p,g/ml streptomycin and I).18 U / m l insulin. They were then plated onto 25 cm 2 Falcon tissue culture flasks, previously coated with a thin film of collagen and mainrained at 37°C in humidified 5% COo atmosphere. Cultured cells were recognized as epithelial from: (11 typical cobblestone-like morphology and polygonal shape; (2) early outgrowth; (3) absence of factor Vlll as determined by a double immunofluorescent technique; (4) positivity for cytokeratins (Kreisberg and Karnovsky, 1983). To avoid mesangial cell contamination, only cells obtained from the primary culture (1-2 week duration) were used. Mesangial cells were recognized from epithelial cells based on the following criteria: (1) morphology; (2) negative staining by indirect immunofluorescence for cytokeratin; (3) contractile response to angiotensin lI.
2.2. ls'olation of glomertdi
2.4. Superoxide release assay
Glomeruli were isolated by a differential sieving technique according to standard methods (Salant et al., 1980) from Sprague-Dawley rats weighing 150-180 g which were anesthetized with ketamine (Ketalar). Rats were fed a normoprotein standard diet (20% casein), and in some cases they were fed a normoprotein diet supplemented with 0.7% sodium tungstate. This diet did not induce any effect on the animal growth nor on other clinical parameters in these rats. After clamping the aorta and cutting off the renal veins, the kidneys were perfused with ice-cold (I.9% NaCI containing 1 mM ethyldiaminetetraacetate (EDTA), 0.2 m m o l / l phenylmethanesulfonylfluoride (PMSF), 1 m m o l / 1 dithiothreitol (DTT), 5 / , g / m [ trypsin inhibitor and 2 U / m l kallikrein (Trasylol). The dissected cortex was then minced with a razor blade in PBS containing the same protease inhibitors and gently squeezed through a stainless steel sieve (250 #m). The resulting suspension was filtered through 150 /xm and 75 p,m sieves and washed three times in Hanks' balanced salt solution (HBSS). Glomeruli were centrifuged at 150 × g for 5 min and resuspended in a small volume of 5(/ mmol/1 Tris-HCl buffer, pH 7.5, containing protease inhibitors. After disruption of cellular structures by sonication, protein content was determined using the Coomassie assay (Read and Northcate, 1981).
Glomerular sonicates (1 m g / m l ) from normal and tungsten-fed rats and sonicated glomerular epithelial cells (105 cells) were incubated with increasing amounts of doxorubicin (3-30 /xg/ml) in sterile phosphate buffer solution (pH 7.4) at 37°C for 2 h. Treatment of cellular extracts with deoxy-coformicin ( 4 / , m o l / l ) was performed for 15 min at 37°C. Release of O~ was estimated by classical assays based on the superoxidc dismutase (SOD)-inhibitable reduction of cytochrome c (Fridovich, 1985). A reaction buffer containing 1 mmol/CaC12, 2 mmol/1 glucose and 100 /xmol/l cytochromc c dissolved in PBS (pH 7.4) was added to each well. For each p a r a m e t e r assessed, half of the wells were incubated with 100 /,tg SOD and half without. After incubation at 37°C for 80 min, reactions were terminated by addition of 2 mM N-ethylmaleimide. Reduced cytochrome c was determined spectrophotometrically at 546 nm using E = 21.2 m m o l / l cm ~ as an extinction coefficient; results were corrected for the protein content in mg. Blank samples were conducted in cytochrome c and doxorubicin without glomeruli and in cytochrome c and glomeruli without doxorubicin. The amount of O ; was calculated by subtracting the amount of reduced cytochrome c in the presence of SOD from total (SOD non inhibited) reduced cytochrome c.
2. Materials and methods 2.1. Materials'
79
2.5. Assays for adenosine deaminase and xanthine oxidase Adenosine deaminase activity in isolated glomeruli was assayed spectrophotometrically at 265 nm, using 0.1 m o l / l adenosine in 0.1 Tris-HCl buffer, pH 7.4 as a substrate. Xanthine oxidase activity was assayed spectrophotometrically at 293 nm, using 0.9 m m o l / l hypoxanthine in 0.1 M Tris-HCI buffer, pH 7.4 as a substrate. All assays were performed at 37°C in 1 ml reaction mixtures. The enzyme activity unit (EU) is the amount of enzyme required to convert 1 /xmol of xanthine or adenosine to uric acid/rain. Specific activity is defined as U / r a g of protein.
for 02 generation from purines by deoxy-coformicin (4 # t o o l / l ) which was incubated for 1 h at 37°C and the cells were then washed twice and resuspended in fresh RPMI. On completion of drug exposure, cells were washed twice and resuspended in fresh medium.
2. 9. Cell viabifity For m e a s u r e m e n t of cell viability, cultures were harvested with 0.05% trypsin-0.02% EDTA, stained with 0.05% Trypan Blue (Shanne et al., 1979) and counted in a hemocytometer. The Trypan Blue assay was performed 1 and 6 days after doxorubicin treatment.
2.6. Coupfing of doxorubicin
2.10. Thyrnidine incorporation
Doxorubicin was coupled with Reacti-gel 6 × , which consists of 6% cross-linked agarose beads ( 4 0 - 2 1 0 / x m in diameter) derivatized with 1,1'-carbonyldiimidazole. In a typical experiment, 3 × 10 s mol of doxorubicin were allowed to react per mg of activated agarose for 25-72 h in 0.1 M borate buffer at p H 8.0 and 4°C. Unreacted imidazole carbonate groups were eliminated with excess hydroxylamine (Tritton and Yee, 1982). Under these conditions about 3 0 - 4 0 % of doxorubicin was attached to the macroporous bed while the unbound fraction was removed by extensive washings (1 week) in water. The washings were continued until no free doxorubicin was found in the supernatant and the resulting immobilized doxorubicin was stored at 4°C in the dark. The stability of doxorubicin coupled with agarose was studied by measuring the rate of release of doxorubicin in phosphate and in R P M I at 37°C for several days. Doxorubicin was determined by H P L C using a reverse phase column (Bondapak C-18) and fluorescence detection (Broggini et al., 1984) and its cellular uptake was determined after incubation for 2h.
Thymidine incorporation was evaluated by incubating 104/ml cells with 5 / z C i / m l [3H]thymidine for 24 h at 37°C in a humidified 5% CO 2 atmosphere. After extensive washings in RPMI, cells were counted in a Packard /3 scintillation counter after being transferred under vacuum to filter p a p e r sheets (Whatman, glass microfibre filters) and solubilized in filter-count scintillation fluid. All analyses were performed 'in triplicate.
2. Z Effect on glomerular epithefial cells of free and agarose-bound doxorubicin
2.11. [~SS]methionine incorporation For [35S]methionine incorporation, cells (104/ml) were incubated with 2 /xCi/ml [3SS]methionine in methionine-free R P M I medium supplemented with 10% dialyzed fetal calf serum for 18 h in a humidified 5% CO 2 atmosphere. The reaction was stopped with ice cold 0.5% Triton X-100-0.05% E D T A with 2 m m o l / l PMSF and 1 retool/1 dithiothreitol (DTT). The precipitate in 10% thrichloracetic acid (TCA) (20 rain room temperature) was filtered under vacuum to filter paper (Whatman, glass microfibre filters) and was solubilized in filter-count scintillation fluid.
2.12. Other methods
Glomerular epithelial cells were transferred to multiwell dishes by plating 105 cells/ml in complete medium, maintained at 37°C in a 5% CO 2 humidified atmosphere. Cells were used for drug experiments 1 day after plating. The effect of doxorubicin on intact glomerular epithelial cells was examined by exposure with the same amounts (7.5 /xg/ml) of free or agarose-bound doxorubicin for 2 h.
The Coomassie binding assay (Read and Northcate, 1981) was employed to evaluate proteins. Uric acid was determined spectrophotometrically with an enzymatic assay based on the conversion of uric acid into allantoine with release of H 2 0 2. Enzymatically produced H 2 0 2 reacts in turn with 3,5-dichloro-2-hydroxyben z e n e sulphonic acid in the p r e s e n c e of 4aminophenazine to give a red quinonine with an absorbance maximum at 520 rim.
2.8. Inhibition experiments"
2.13. Statistical analysis
Doxorubicin cytotoxicity on glomerular epithelial ceils was tentatively inhibited by blocking the pathway
The one way analysis of variance was used in all statistical tests. Results are given as means _+ S.E.
80
3. Results
3.1. Cytotoxic effect lial cells
of doxorubicin on glomerular epithe£
[]
NORMAL
[]
TUNGSTAT E
GLOMERULI
200-
O
As shown in fig. 1, doxorubicin induced a dose related inhibition of [3H]thymidine incorporation into DNA of intact glomerular epithelia which was maximal for concentrations of the drug >_ 15 /~g/ml. Furthermore, exposure of cells to the same levels of drug inhibited protein synthesis ([~SS]methionine incorporation) and reduced cell viability (Trypan Blue).
100-
> O -
÷
--
4-
SOD
SOD
o';
o~
30
15
3.2. Glomerular O~ synthesis Doxorubicin ~glrnL
The glomerular production of 02 in the presence of doxorubicin was evaluated as the SOD-inhibitable reduction of cytochromc c, which is one of the most popular tests to detect 02 in biological fluids. The production of 02 by glomeruli in the presence of doxorubicin was evident for levels of the drug producing cytostasis ( > 15 /zg/ml); at drug levels of 30 /zg/min the SOD-inhibitable amount of reduced cytochrome c was 50 +_ 10 # m o l / m g protein (fig. 2). However, 0 2 was not the only responsible for this effect and the sum of SOD-inhibitable and non-SODinhibitable reduction of cytochrome c in the presence of glomeruli and of doxorubicin was in fact greater by a factor of 3.5 (175 +_ 10 /~mol/mg of protein). Other reactions involving doxorubicin transformation into an unstable radical by glomeru[ar extract would account for it. Since most of the O 2 produced by biological system arises from the transformation of hypoxanthine and xanthine into uric acid via the xanthine oxidase system, glomerular 0 2 synthesis would be paralleled by uric acid formation. Indeed as shown in fig. 3, in the presence of doxorubicin, the glomerular synthesis of
w°t
li00m F o
z
30
Fig. 2. Reduction of cytochrome c by doxorubicin (15-30 ~ g / m l ) incubated for 2 h at 37°C with sonicated glomeruli (l m g / m l ) isolated from normal rats (open bars) and from rats fed a tungsten enriched diet (shaded bars). For each parameter assessed, half the wells were incubated with S O D (100 >g) and half without; O; synthesis was calculated by subtracting the amount of SOD-inhibitable reduction of cytochrome c from total non-SOD-inhibitable reduction. Non-SOD-inhibitable cytochrnme e reduction accounted for about 7 0 ~ of the overall effect. Results of six experiments for each group are given as mean + S.E. * P < 0.01.
uric acid increased in a dose-related manner. A few experiments were performed with glomeruli purified from kidneys of rats which were fed a diet enriched in tungsten, a heavy metal which competes with the intestinal absorption of molybden. Since molybden is essential for xanthine oxidase activity, glomeruli purified from tungsten-enriched diet presented only minimal activity of this enzyme (XO 0.8 _+ 0.1 vs. 5.1 _+ 1.5 mEU mg L). According to the concept that doxorubicin-induced glomerular 0 2 is synthesized with the intervention of xanthine oxidase, the production of O_~ in the presence of glomeruli purified from tungsten-fed
,c_
o
o
(3.
75
75
10 za.
-50
~
~ ~
25
25
x
a < O
5
D
0
3
7
Doxorubicin ¢on¢.
15
75
Pg/ml
Fig. 1. Dose-response inhibition of [~H]thymidine incorporation into DNA, and cell viability (6 days following exposure) of glomerular epithelial cells after incubation with increasing a m o u n t s of doxorubicin. Each experiment was performed in quadruplicate. Values have been normalized to [3HJthymidine incorporation by normal cells ( 100% incorporation).
6
~ lO ~
Doxorubicin
~ 2 ~
ug Iml
0
1
2
3
4
5
e
Time (hour's)
Fig. 3. Production of uric acid from normal glomeruli (1 m g / m l ) incubated for 4 h with increasing amounts of doxorubicin (5-30 p.g/ml). Time-dependent uric acid synthesis was evaluated by incubating glomeruli with constant 15 /~g/ml doxorubicin for various times. Each experiment was performed in quadruplicate. Results are the mean +_S.E.
81
20-
[]
GEC
[]
GEC -f- COFORMICIN
TABLE 1 In vitro cytotoxicity of doxorubicin (15 /zg/ml) on intact cultured glomerular epithelial cell in the presence of deoxy-coformicin (4 IzM) or in its absence. 24 h [3H]thymidine and [35S]methionine incorporation was evaluated after a 2 h exposure of cells to doxorubicin. Cell viability was evaluated 6 days after exposure to doxorubicin by the Trypan Blue dye exclusion test.
r
oi
Treatment
[3H]thymidine (cpm/104 cells)
[35S]methionine (cpm/104 cells)
Cell viability (% of total cells)
None (n = 5) Coformicin alone (n = 4) doxorubicin alone (n = 4) Doxorubicin + coformicin (n=4)
4,000_+500
35,000_+5,000
75
_+0.5
4,100+800
33,000_+1,200
74
+ 1
O :z. -- 4SOD Doxorubicin
SOD
O~
30
~UgImL
Fig. 4. Reduction of cytochrome c by doxorubicin (30 txg/ml) in the presence of cellular extracts from glomerular epithelial cells. Incubation was carried out for 2 h at 37°C. Sonicated glomerular epithelial cells (open bars); cells pretreated with deoxy-coformicin, a specific inhibitor of adenosine deaminase (shaded bars). For 0 2 calculation see Materials and methods. Experiments were performed with six different samples of GEC. Results are the mean _+S.E. * P < 0.01 vs. untreated cells.
rats (e.g., without xanthine oxidase activity) was practically nil. Taken together these data demonstrate that doxorubicin induces the glomerular synthesis of O~ which derives from the conversion of xanthine into uric acid by the enzyme xanthine oxidase.
3.3. 0 2 synthesis by glomerular epithelial cells As shown in fig. 4, reduction of cytochrome c by doxorubicin was also elevated in the presence of glomerular epithelial cell extracts. As already reported for isolated glomeruli, a part of this effect ( = 22%) was suppressed by SOD and represented the effective O;~ synthesis (9.2 ~ m o l / m g protein). Also in the case of glomerular epithelial cells, experiments aimed at blocking the purine degradative pathway leading to hypoxanthine and uric acid were successful in producing a reduction of the oxidative potential of doxorubicin. In fact, in this case, the production of 0 2 was inhibited by deoxy-coformicin which is a specific inhibitor of adenosine deaminase (Whoowdion et al., 1974) and therefore regulates the production of substrates for xanthine oxidase. Treatment with deoxy-coformicin also reduced the cytolytic effect of doxorubicin on cells which is evident as an increase in cell viability after 6 days (table 1)
200 + 100 ~'
7,200 +
325+ 150 "
6,700_+ 210 "
52.2 + 1.2 "
67.3+3h
" P < 0.001 vs. control cells, b p < 0.05 vs. doxorubicin alone.
P-450 reductase. Both mechanisms imply the presence of the drug within the cell. To prove this point doxorubicin was chemically linked to agarose (Tritton and Yee, 1982) and incubated with glomerular epithelial cells in the same experimental conditions as free doxorubicin. Since the agarose-doxorubicin complex is too large to penetrate into cells, the intracellular levels of the drug were nil while more than 70% was detected inside the ceils when the free drug was used (5 vs. < 1 /~g/ml where 1 p~g/ml is the upper limit of detection). In spite of this, free and agarose-bound doxorubicin produced the same cytostatic effect on glomerular epithelial cells corresponding to 50% inhibition of thymidine incorporation with doxorubicin levels of 7.5 ~ g / m l (table 2). This fact suggests that other mechanisms besides O~ and direct oxidation are responsible for doxorubicin cytotoxicity on glomerular epithelial ceils.
TABLE 2 In vitro cytotoxicity of free doxorubicin and agarose-bound doxorubicin on cultured epithelial cells. 24 h [3H]thymidine incorporation was evaluated after exposure of cells to A D R (7.5 txg/ml) for 2 h. Cell viability was evaluated by the Trypan Blue assay 1 and 6 days after 2 h doxorubicin exposure. Treatment
3.4. Cytotoxicity of free and agarose-bound doxorubicin The basic assumption of the experiments so far presented is that any direct cytotoxic effect of doxorubicin by means of O 2 a n d / o r as a semiquinone molecule should require an enzymatic pathway, namely xanthine oxidase activity or conversion by cytochrome
380 "
None (n = 6) Doxorubicin (n = 4) Agarose-doxorubicin (n=4)
% [3H]thymidine incorporation
% cell excluding Trypan Blue
(vs. controls)
Days after treatment
100 43 _+5 ~ 40_+6"
~' P < 0.001 vs. control ceils.
I st
6th
99_+0.5 98 _+ 1
84_+6 51 + 2 "
99_+1
65+8
82
4. Discussion This paper was planned to focus on the mechanisms for doxorubicin cytotoxicity on glomerular epithelial cells 'in vitro', with particular emphasis on assessing a possible role of free radicals. This implication was suggested by recent works on the toxicity of doxorubicin on other cell lines (Doroshow, 1983, 1986) and in vivo (Myers et al., 1977; Doroshow et al., 1981; Ghiggeri et al., 1991), which supported the concept that superoxide, peroxide and hydroxyl radicals are responsible for the toxic action of the drug. At least two metabolic pathways could be responsible for free radical generation from doxorubicin. One involves the production of semiquinone intermediates from doxorubicin by intervention of N A D P H , N A D P H - c y t o c h r o m e P-450 reductase or cytochrome P-450 and probably other enzymes such as N A D P H dehydrogenase (Bacher et al., 1979; Davies et al., 1983). The second mechanism is based on a direct stimulatory effect of doxorubicin on O~ generation. In accordance with this idea, the cytotoxicity of several anticancer quinones on Ehrlich tumor cells ',in vitro' was reduced, if not suppressed, by several antioxidant enzymes and chemical scavengers such as catalase and SOD, dimethyl sulfoxide and dimethylurea which act on the hydroxyl free radical and by the iron chelator deferoxamine (Lee et al., 1983). 'In vivo' the cardiotoxicity of A D R was partially inhibited by the administration of cysteine (Doroshow et al., 1991), the thiol component of glutathione, which exerts a key scavenger action within the cell (Faber et al., 1990). The nephrotoxic potential of doxorubicin in SpragueDawley rats, which is responsible for the development of proteinuria similar in many aspects to that found in human minimal change nephropathy, was likewise blunted by selective inhibitors of renal xanthine oxidase (Ginevri et al., 1990), the enzymatic system responsible for the generation of O;, from hypoxanthine (Fridovich, 1978) and by dimethylthiourea, the scavenger of the hydroxyl radical (Ghiggeri et al., 1991). In general, although the data are rather convincing, conclusive evidence for a role of free radicals in the anticancer action of doxorubicin, and more specifically in the cytotoxic effect on renal cells, is lacking. The data reported here reinforce the idea that doxorubicin in the presence of glomeruli and of glomerular epithelial cells 'in culture' induces O~ generation which is partially responsible for its cytotoxicity. They also extend the knowledge on the metabolic pathways responsible for O~ generation, which involve purine catabolism and the final conversion of hypoxanthine into uric acid by the enzyme xanthine oxidase. Two central experiments involving selective blocks of the purine cascade, one of xanthine oxidase by dietary sodium tungstate and the other by the adenosine
deaminase inhibitor deoxy-coformicin, support this conclusion. Sodium tungstate was chosen as a chronic model of xanthine oxidase inhibition since it determines a selective and protracted inhibition of xanthinc oxidase without affecting the cellular uptake of purines. Tungsten competes with molybdenum for its intestinal absorption, and molybdenum-deprived rats present markedly reduced glomerular XO activity since this metal is a structural component of the hydrophobic functional core of the enzyme (Tophan et al., 1988). This model has been extensively used in recent years as an alternative to allopurinol in experimental models of functional block of renal xanthine oxidase, especially in studies on renal ischemia (McKelvey et al., 1988). The data reported here unequivocally demonstrate that O:, is produced by glomeruli and glomerular epithelial cells in the presence of doxorubicin while no O~ is spontaneously synthesized and that purine catabolism via xanthine oxidase represents the metabolic pathway for O; generation. Nevertheless, even in the case of a block of O~ production by tungsten and coformicin, doxorubicin is still able to exert an oxidative potential in the presence of glomeruli or cells. It is reasonable to suspect that, in this case, this effect is directly mediated by transformation of doxorubicin by cytochrome P-450 reductase into an unstable semiquinone. Another important point raised by this study suggests that, besides oxidative mechanisms, doxorubicin has other effects on renal cells. This conclusion is based on the observation that doxorubicin is actively cytostatic and cytotoxic on glomerular epithelia, even when its intracellular levels arc nil, owing to a stable chemical linkage with a macroporous bed which keeps the drug out of the cells during incubation. We suggest that in this case a m e m b r a n e perturbation may be responsible for the toxicity and that doxorubicin may exert its biological activity solely by interaction with the cell surface. The second implication of these data is that doxorubicin exerts a multiple toxic mechanism on glomerular epithelia 'in vitro' and that multiple sites are targets for its action. In accordance with these concepts, other works (Goormaghtigh et at., 1980: Oth et al., 1987) have already demonstrated that doxorubicin induces modifications of the membrane lipid composition, fluidity and permeability of several cell lines 'in vitro' by interaction with membrane phospholipids.
Acknowledgements The authors are indebted to Prof. Giovanni Ccrcignani for having determined glomerular adenosine deaminase and xanthine oxidase contents. Mr. Robert B. Stubinski is acknowledged for his secretarial support. These data were in part presented at the XXIV Annual Meeting of the American Society of Nephrology, Baltimore, November 17-20, 1991.
83 This work was carried out with a grant of the Italian Ministry of Health to the G. Gaslini Institute.
References Bachur, N.R., S.L. Gordon, M.V. Gee and H. Kon, 1979, NADPH cytocbrome P-450 reductase activation of quinone anticancer agents to free radicals, Proc. Natl. Acad. Sci. USA 76, 954. Bates, D.A. and C.C. Winterbourn, 1982, Deoxyribose breakdown by the adriamycin semiquinone and H202: evidence for hydroxyl radical participation, FEBS Lett. 145, 137. Bertani, T., A. Poggi, R. Pozzoni, F. Delaini, G. Sacchi, Y. Thoua, G. Mecca, G. Remuzzi and M.B. Donati, 1992, Adriamycin-induced nephrotic syndrome in rats. Sequence of pathological events, Lab. Invest. 46, 16. Broggini, M., C. Italia, T. Colombo, L. Marmonti and M.G. Donelli, 1984, Activity and distribution of i.v. and oral 4-demetboxydaunorubicin in murine experimental tumors, Cancer Treat Rep. 68. 739. Davies, K.J.A., J.H. Doroshow and P. Hochstein, 1983, Mitochondrial NADH dehydrogenase-catalyzed oxygen radical production by adriamycin and the relative inactivity of 5-iminodaunorubicin, FEBS Lett. 153, 227. Doroshow, J.H., 1983, Anthracycline antibiotic stimulated superoxide, hydrogen peroxide and hydroxyl radical production by NADH dehydrogenase, Canc. Res. 43, 4543. Doroshow, J.H., 1986, Role of hydrogen peroxide and hydroxyl radical formation in the killing of Ehrlich tumor cells by anticancer quinones. Proc. Natl. Acad. Sci USA 83, 4514. Doroshow, J.H., G.Y. l,x)cker, !. lfrim and C.E. Myers, 1981, Prevention of doxorubicin cardiac toxicity in the mouse by N-acetylcysteine, J. Clin. Invest. 68, 1053. Faber, J.L., M.E. Kyle and J.B. Coleman, 1990, Biology of disease. Mechanisms of cell injury by activated oxygen species, Lab Invest. 62, 670. Fridovich, I., 1978. The biology of oxygen radicals, Science 201, 875. Fridovich I., 1985, Cytochrome c, in: Methods for Oxygen Radical Research, ed. R.A. Greenwald (CRC Press, Boca Raton, FL) p. 121. Ghiggeri, G.M.. F. Ginevri, G. Cercignani, R. Oleggini. A. Garberi and R. Gusmano, 1990, Effect of dietary protein restriction on renal purines and related metabolizing enzymes in adriamycin nephrosis in rats. A mechanism for protection against proteinuria involving xanthine oxidase inhibition, Clin. Sci. 79, 647. Ginevri, F.. R. Gusmano, R. Oleggini, S. Acerbo, R. Bertelli, F. Perfumo, G. Cercignani, S. Allegrini, F. D'Allegri and G.M.
Ghiggeri, 1990, Renal purine efflux and xanthine oxidase activity during experimental nephrosis in rats: difference between puromycin aminonucleoside and adriamycin nephrosis, Clin. Sci. 78, 283. Goormaghtigh, E., P. Chatelain, J. Caspers and J.M. Ruysschaert, 1980, Evidence of a specific complex between adriamycin and negatively charged phospholipids, Biochim. Biophys. Acta 597, 1. Gutteridge, J.M.C., G.J. Quinlan and S. Wilkins, 1984, Mitomycin C-induced deoxyribose degradation inhibited by superoxide dismutase, FEBS Lett. 167, 37. Kreisberg, S.I. and M.J. Karnovsky, 1983, Glomerular cells in culture, Kidney Int. 23, 439. Lee, F.Y,F., D.W. Siemann and R.M. Sutherland, 1989, Changes in cellular glutathione content during adriamycin treatment in human ovarian cancer - a possible indicator of chemosensitivity, Br. J. Cancer, 60, 291. McKelvey, T.G., M.E. Hollwarth, D.N. Granger, T.D. Engerson, U, Landler and H.P. Jones, 1988, Mechanisms of conversion of xantbine oxidase in ischemic rat liver and kidney, Am. J, Physiol, 254, G753. Myers, C.E., W.P, McGuire, R.H. Liss. I. llfrim, K. Grotzinger and R.C. Young, 1977, Adriamycin: the role of lipid peroxidation in cardiac toxicity and tumor response, Science 197, 165. Oth, D., M. Begin, P. Bischoff, J.Y. Leroux, G. Mercier and C. Bruneau, 1987, Induction by adriamycin and mitomycin C of modifications in lipid composition, size distribution, membrane fluidity and permeability of cultured RDM4 lymphoma cells. Biochim. Biophys. Acta 900, 198. Read S.M. and D.M. Northcate, 1981, Minimization of variation in the response of different proteins to the Coomassie-Blue dye binding assay for protein, Anal. Biochem. 116, 53. Salant, D,J., C. Darby, C. Couser and W.G. Couser, 1980, Experimental membranous glomerulonephritis in rats. Quantitative studies of glumerular immune deposit formation in isolated glomeruli and whole animals, J. Clin. Invest. 66, 71. Schanne, F.A., A.B. Kane, E.E. Young and J.C. Farler, 1979, Calcium dependence of toxic cells death: a final common pathway, Science 206, 700. Tophan, R.W., M.C. Walker, M.P. Lalish and W. Williams, 1982, Evidence for the participation of intestinal xanthine oxidase in the mucosal processing of iron, Biochemistry 21, 4529. Tritton, T.R. and G. Yee, 1982, The anticancer agent adriamycin can be actively cytotoxic without entering cells. Science 217, 248. Woohwdion, P.W.K., S.M. Lange, L.S. Dohl and L.J. Durhan, 1974, A novel adenosine and ara-A deaminase inhibition, (R)-3-(2-deoxy-/3-D-erythro-pentafuranosyl)-3,6,7,8-tetrahydroimidazo[4,5-6][1,3]diazepin-8-AL, J. Antibiotics 11,641.
