Interactions between N-acetyl-p-benzoquinone imine and fluorescent calcium probes: Implications for mechanistic toxicology

ANALYTICAL

BIOCHEMISTRY

19 1,253-261

(1990)

Interactions between A/-Acetyl-p-benzoquinone and Fluorescent Calcium Probes: Implications Mechanistic Toxicology’ Robert J. Riley,2 J. Steven Leeder, H.-Michael

lmine for

Dosch, and Stephen P. Spielberg

Divisions of Clinical PharmacologylToxicology and ImmunologylAUergy, Research Institute, Departments of Pediatrics and Pharmacology, Centre for Drug Safety Research, University of Toronto, Toronto, Ontario, Canada M56 1X8 -

Received

June

The Hospital

for Sick Children,

and

19,199O

Intracellular free calcium ([Ca”],) homeostasis has been implicated as an early target in both cellular necrosis and apoptosis. In this study, we have used peripheral blood mononuclear cells (PBMC) as target cells to investigate the effects of several reactive metabolites associated with drug toxicity on [Ca2+l, in order to delineate further early events in cytotoxicity. Compounds implicated in both drug-induced necrosis (N-acetyl-pbenzoquinone imine; NAPQI) and drug hypersensitivity (sulfamethoxazole hydroxylamine; SMX-HA) were examined and their effects on [Ca2+l, compared with those of the T cell mitogen phytohemagglutinin (PHA; 1.5 pg/ml) and the calcium ionophore ionomycin (2.5 FM). PHA and ionomycin produced characteristic elevations in [Ca2+li as monitored by an increase in the fluorescence of fluo-34oaded cells. SMX-HA did not significantly affect [Ca2’l, at concentrations previously shown to be cytotoxic to PBMC (100 and 500 PM), suggesting that Ca2+ homeostasis is not an early target for SMXHA toxicity. Addition of NAPQI (250 pM) to fluo-3loaded cells produced a marked decrease in fluorescence which was not reversed by ionomycin. Conversely, addition of NAPQI to cells loaded with indo-l resulted in a rapid increase in fluorescence. This effect, however, was found to be attributable to NAPQI addition per se rather than to an increase in [Ca2+],. HPLC and fluorescence analysis of samples generated

i Supported by the Medical Research Council of Canada and the National Cancer Institute of Canada, and a matching grant from the University Research Incentive Fund of the Ontario Ministry of Colleges and Universities and Baxter Canada. ’ To whom correspondence should be addressed at Department of Pediatrics, Division of Clinical Pharmacology/Toxicology, The Hospital for Sick Children, 555 University Avenue, Toronto, Ontario, Canada M5G 1X8. 0003-2697/90 $3.00 Copyright 0 1990 by Academic Press, Inc. All rights of reproduction in any form reserved.

from the decomposition of NAPQI revealed the presence of several products which fluoresced intensely at the excitation/emission wavelength pairs of a number of fluorescent probes commonly used to monitor [Ca”], . Our results show that none of these calcium indicators can be used to monitor changes in [Ca”], produced by NAPQI and strongly suggest that the previously observed early, sustained elevation in [Ca”], may have reflected the formation of fluorescent NAPQI decomposition products rather than a disruption in [Ca2’], homeostasis. 0 1990 Academic Press, Inc.

Although the metabolic activation of drugs andchemicals to reactive intermediates is recognized as a prerequisite for many of their adverse effects (l-3), the precise pathogenesis of idiosyncratic drug reactions remains poorly defined. Recent in vitro studies with isolated hepatocytes have suggested that intracellular free calcium ( [Ca2’]i)3 homeostasis is an important early cellular target for chemically induced cell injury produced by a variety of compounds, including carbon tetrachloride (4), tert.-butyl hydroperoxide (5), menadione (6), and the chemically reactive metabolite of acetaminophen, Nacetyl-p-benzoquinone imine (NAPQI) (7). Indeed, it could be argued that NAPQI has become the paradigm for reactive drug metabolites which are presumed to mediate cellular necrosis via this mechanism. An elevation in [Ca2+]i has also been shown to be an early event in “apoptosis” or programed cell death (8). 3 Abbreviations used: [Ca”],, intracellular free calcium concentration; NAPQI, N-acetyl-p-benzoquinone imine; SMX-HA, sulfamethoxazole hydroxylamine; PBMC, human peripheral blood mononuclear cells; PHA, phytohemagglutinin; PBS, phosphate-buffered saline; FU, arbitrary fluorescence units. 253

254

RILEY

Moreover, although this pattern of cell death was originally thought to have purely a physiological role (9), it is becoming increasingly clear that apoptosis may occur in tissues subjected to injury by a number of nonphysiological stimuli, including reactive chemicals (10,ll). It has been proposed that the chronology and magnitude of the rise in cytoplasmic calcium may determine whether a cell undergoes apoptosis or dies via an alternative mechanism, i.e., necrosis (12). We have previously shown that human peripheral blood mononuclear cells (PBMC) may be employed as target cells in the study of the biochemical mechanisms of idiosyncratic drug reactions (13-15). Indeed, it would appear from studies with acetaminophen that similar phases of toxicity occur in PBMC (13,15) and isolated hepatocytes (16) exposed to the chemically reactive metabolite(s) of this analgesic agent, namely, interaction of the resultant toxic intermediate with cellular components, followed by progressive damage and, ultimately, cell death. However, the adverse effects of several compounds which have been shown to be activated to cytotoxic metabolites in vitro are presumed to be hypersensitivity reactions rather than the exclusive result of direct necrosis (17,18). Therefore, the initial aim of the present study was to examine whether events similar to those observed in isolated hepatocytes exposed to NAPQI occur in PBMC challenged with this chemically reactive intermediate and to investigate if similar phenomena occur with reactive metabolites implicated in drug hypersensitivity. During the course of these experiments, it became apparent that NAPQI itself interacted chemically with the fluorescent probe chosen to monitor [Ca’+],, fluo-3 (19), and readily decomposed to products which fluoresced intensely at the excitation/ emission wavelength pairs used for a number of commonly used calcium indicators. Hence, this communication reports attempts to delineate the mechanisms involved in these fluorescent probe-metabolite interactions and suggests that the early, sustained rises in [Ca2+], previously described (20) may have, at least in part, reflected the rapid decomposition of NAPQI to products which precluded an accurate assessment of changes in the levels of cytosolic free calcium. EXPERIMENTAL

PROCEDURES

Materials. Fluo-3, its acetoxymethyl ester (fluo-31 AM), and indo-l/AM were purchased from Molecular Probes, Inc. (Eugene, OR). Anhydrous (dry; water ~0.005%) dimethyl sulfoxide (Me,SO) and ionomycin were obtained from Aldrich Chemical Co. (Milwaukee, WI) and Calbiochem Corp. (San Diego, CA), respectively. NAPQI and the hydroxylamine of sulfamethoxazole (SMX-HA) were custom synthesized by Dalton Chemical Laboratories (Toronto, ONT). The purity of both compounds was >99% (as determined by NMR).

ET

AL.

All other chemicals cially available.

were of the highest

grade commer-

Isolation of human PBMC. PBMC were isolated from peripheral blood of healthy volunteers as previously described (14,21). Briefly, heparinized blood was diluted 1:l with medium ((u-MEM; University of Toronto Media Services, Toronto, ONT) and layered over Histopaque-1077 (Sigma Chemical Co., St. Louis, MO). The tubes were then centrifuged (650g for 15 min) and the resultant interface layer (PBMC) harvested with a Pasteur pipet. PBMC were suspended in medium (RPM1 1640 containing 10% v/v fetal calf serum) and centrifuged at slower speeds (120g for 8 min) to reduce platelet contamination. Finally, the cells were resuspended in fresh medium to give the required cell density. Measurement of cytosolic calcium concentrations in PBMC. PBMC were loaded with fluo-3/AM (1 pM) for 30 min at room temperature. A nonionic surfactant, Pluronic F-127 (1~1 of a 25% w/v solution per 50 rg dye; BASF Wyandotte Corp., Wyandotte, MI), was used to aid solubilization of the AM ester into aqueous solution and to reduce compartmentalization of the dye within the cytoplasm. Following loading of the dye, the cells were pelleted and washed twice in RPM1 1640 (containing 10% fetal calf serum) to remove extracellular fluo-3/ AM. Initially, the effects of NAPQI and SMX-HA were compared to those of a calcium ionophore as well as to the T-cell activator phytohemagglutinin (PHA). Studies were conducted with a Hitachi F2000 fluorescence spectrophotometer, and fluorescence was monitored continuously at 37°C with excitation and emission wavelengths of 485 and 535 nm, respectively. PBMC (1 X 106, in 2 ml 0.1 M phosphate-buffered saline, pH 7.4; PBS) were stimulated with PHA (1.5 pg/ml). Ionomycin (2.5 pM) was then added to rapidly induce a large influx of calcium and to determine the maximum fluorescence signal for a given suspension of fluo-3-loaded cells V’,,). Once F,, had been established, EGTA (4.0 IIIM) was added to obtain Fmi, (the addition of MnCl, to phosphate-buffered saline resulted in precipitation and, hence, precluded the determination of F,,). Alternatively, incubations were conducted in phosphate-free Hepes-buffered medium (13), which permitted the determination of F,, . Cytosolic calcium for a given signal F was then calibrated by standard procedures (19), using the relationship [Ca2+], = K,(F - F,,)l(F,,, - F), where Kd is 450 nM. The effects of NAPQI and SMXHA were examined in a manner similar to that described for the effects of PHA. These compounds were added in anhydrous Me,SO to give a final solvent concentration of ~2% (which was not associated with any toxicity). In some experiments, the effect of NAPQI on intracellular calcium was monitored using another fluores-

FLUORESCENT

PROBE-METABOLITE

cent probe, indo-1. PBMC were loaded with indo-l/AM (1 pM at 37°C for 30 min). The cells were washed free of extracellular dye and 1 X lo6 were incubated at 37°C in the fluorescence spectrophotometer. NAPQI (250 pM) was then added and its effects on cytosolic calcium levels were monitored by the change in fluorescence observed at an excitation/emission wavelength setting of 331/410. Concentrationand time-dependence of interaction between NAPQI and @o-3. PBMC (2.5 X 104) loaded with fluo-3/AM were exposed to varying concentrations of NAPQI (O-500 PM). Incubations were conducted for 5,10, or 20 min at 37°C in a 96-well unidirectional, vacuum filtration plate (Baxter Healthcare Corp., Pandex Division, Mundelein, IL). The final concentration of anhydrous Me,SO in the system was 2%. After drug challenge, inert polystyrene microspheres (10 ~1 0.2%, 8.0 pm, Baxter-Pandex) were added to each well. Using an automated protocol (22), the filtration plates were washed and vacuumed, and the fluorescence was read by front surface fluorimetry in a Screen Machine (Baxter-Pandex) with excitation and emission wavelengths of 485 and 535 nm, respectively. In addition, the interaction of NAPQI with fluo-3 free acid and/or the fluo-3-Ca2+ complex was studied by monitoring the change in fluorescence with time observed following the addition of NAPQI (250 PM) to fluo-3 (10 PM) in PBS or Ca2+/Mg2+-free PBS containing EGTA (4.0 mM). Investigation of the decomposition of NAPQI. NAPQI (7.0 mg) was dissolved in anhydrous Me,SO to give stock concentrations of 25 and 2.5 mM. Aliquots (20 ~1) of these solutions were then immediately added to PBS (980 ~1) to give final concentrations of 500 and 50 pM and incubated at 37°C in a shaking water bath for varying lengths of time. Reactions were terminated by the addition of 1 ml ice-cold methanol:water (9O:lO) containing 2.0 mM ascorbic acid and placing the incubation mixtures on ice (23). Samples were then maintained at -20°C until analyzed by high-performance liquid chromatography. Characterization of the products of NAPQI degradation. The products of NAPQI degradation were investigated by HPLC in a manner similar to that described by Potter et al. (23), using a Waters 501 HPLC pump and a Lambda-Max Model 481 detector (Waters Associates, Milford, MA). Aliquots (lo-25 ~1) were injected and separation was achieved on an Ultrasphere 5-pm ODS Cl8 column (HPLC Technology, Cheshire, UK) with a mobile phase consisting of methanol:water:acetic acidethyl acetate (17.9:80.0:2.0:0.1). The flow rate was 1 ml/min and detection was at 254 nm. In order to investigate any possible interference of the NAPQI degradation products with the fluorescence of the commonly used calcium probes, the relative fluores-

INTERACTIONS

255

cence of the samples generated in the degradation studies was monitored at the excitation and emission wavelength settings used for quin-2, fura-2, indo-l, and fluo-3. The excitation/emission pairings (nm) used for the respective calcium-free and calcium-bound forms of these probes were 365/492 and 3391492 (quin-2); 3621 512 and 3351505 (fura-2); and 3491485 and 3311410 (indo-1), as described by Cobbold and Rink (24). RESULTS Comparative effects of mitogen stimulation, ionophore addition, and reactive metabolite challenge on cytosolic After a brief lag period, stimucalcium levels in PBMC. lation of PBMC with the T-cell mitogen phytohemagglutinin (1.5 pg/ml) produced a characteristic, slow increase in [Ca2+], from the resting level of around 160 to 230 nM (Fig. 1A). Addition of the calcium ionophore ionomycin (2.5 PM) resulted in a rapid increase in [Ca2+li to in excess of micromolar concentrations (>lO PM), which was generally maintained for 10 min or more. The detergent Triton X-100 (0.2%) did not significantly increase [Ca2+], beyond that observed with ionomycin (data not shown) and, hence, it was concluded that ionomycin had induced F,, and that significant compartmentalization of the dye had not occurred. In contrast, the addition of NAPQI (250 PM) to PBMC loaded with fluo-3 resulted in a marked decrease in fluorescence (Fig. 1B). Furthermore, subsequent addition of ionomycin did not result in a rapid increase in the fluorescence signal to F,,,. This effect appeared to be relatively specific to NAPQI (rather than a property of reactive chemicals per se) as neither SMX-HA nor its spontaneous two-electron oxidation product, nitrosoSMX, exerted a similar effect (Fig. 1B). Moreover, the apparent small rise in [Ca2+li observed with SMX-HA was a consequence of leakage of fluo-3 out of the cells, as it was also observed in control incubations (1 X lo6 and 1% Me,SO; data not shown). Interaction between NAPQI and fEuo-3. Further characterization of the effects of NAPQI on the fluorescence in fluo-3-loaded PBMC confirmed that this highly reactive intermediate produced a concentrationand time-dependent decrease in the fluorescent signal compared with that monitored in vehicle-treated cells (Fig. 2). However, an attempt to define and quantitate this inhibition by transforming the data into a double-reciprocal plot proved unsuccessful as the relationship between 1/V and 1 /S was nonlinear and, hence, precluded the determination Of Ki for this reaction (inset to Fig. 2). Addition of NAPQI (250 PM) to the free acid of fluo-3 in either PBS or Ca2+/Mg 2+-free PBS also resulted in a time-dependent decrease in the fluorescent signal (Fig. 3). Furthermore, fluorescence excitation spectra (emission at 436 nm) of the reaction mixtures showed that NAPQI (or degradation products) exhibited maximum

RILEY

0

2

4

6

Time

8

10

12

(mln)

ET

AL.

0

10

5

Tlme

15

(min)

FIG. 1. Stimulation of calcium fluxes in PBMC. Changes in [Ca’+], were measured with the fluorescent Ca2+ indicator fluo-3, as described under Experimental Procedures. Abbreviations and concentrations used are: P, phytohemagglutinin (1,5 fig/ml); I, ionomycin (2.5 PM); D, either 250 pM N-acetyl-p-benzoquinone imine (B, trace a) or 500 pM sulfamethoxazole hydroxylamine (B, trace b). The fluorescence intensity during the stable baseline was equivalent to 59.4 + 6.0 FU (mean -C SE for six experiments). Each trace is representative of at least three separate cell preparations.

excitation at a wavelength of 323 nm and that interaction of NAPQI with fluo-3 in Ca2+/Mg2+-free PBS resulted in a decrease in this fluorescence and a shift in the excitation maximum towards the blue (insets to Fig. 3). This effect appeared to be independent of the age of the NAPQI stock solution (data not shown). Effects of NAPQI on the fiuorescence in indo-l-loaded PBMC. Figure 4A shows the effect of NAPQI on the fluorescence in PBMC loaded with indo-1. Immediately after the introduction of NAPQI, the fluorescence increased rapidly to approximately twice the resting level (from 679 to 1393 FU). The signal then decreased to 1136 FU and steadily increased linearly with time to around Fm, for this dye. Neither MnCl, (2.0 mM) nor EGTA (4.0 mM) was able to inhibit the rate of increase in fluorescence. Interestingly, the addition of NAPQI to buffer (containing no cells or dye) produced a fluorescence-time profile almost identical to that observed in the presence of dye loaded cells (Fig. 4B). During the course of these experiments, it became apparent that the stock solutions of NAPQI (generally 25 mM in dry Me,SO) were decomposing, as indicated by a change in the color of the solutions from bright yellow through pale green to brown. Purging the Me,SO with nitrogen prior to dissolving the NAPQI failed to prevent this rapid degradation. Further studies therefore focused on determining whether the effects observed with NAPQI could be attributed to degradation products. Characterization of the products of NAPQI degradation. Reversed-phase HPLC analysis of NAPQI solutions which had been allowed to decompose in PBS re-

vealed the presence of five major peaks with retention times of 2.8, 7.8,11.2,17.6, and 20.4 min (a typical chromatogram is shown in Fig. 5). Furthermore, Fig. 6 shows that there was a time-dependent decrease in peak I, coupled with a concomitant increase in peaks II, III, IV, and V. The lack of authentic standards precluded identification of the products by cochromatography. However, from a consideration of the physicochemical properties of the compounds and a comparison of the HPLC profile with that obtained by Potter et al. (23), it appears likely that peaks II, III, IV, and V were acetaminophen polymers (dimer, trimer, N-dimer, and tetramer, respectively). Under the HPLC conditions employed in the present study, acetaminophen was eluted with a retention time of 2.9 min and therefore peak I presumably represents this compound, which may have been produced via disproportionation of two semiquinone imine radicals (25) and/or from the reduction of NAPQI and semiquinone imine radicals, following the addition of ascorbic acid (23). Indeed, at early time points (less than 1 min), the addition of ascorbic acid virtually quenched the formation of all polymers due to the rapid reduction of NAPQI and the semiquinone radicals to acetaminophen (data not shown). Ascorbic acid (2.0 mM) was therefore subsequently added to reaction mixtures to terminate polymerization at the end of incubation as described previously (23). Injection of authentic hydroquinone confirmed that this compound was not a major product of NAPQI degradation, although it may in theory have been produced from the hydrolysis product, benzoquinone. This time-dependent decomposition of NAPQI was also dependent on the concentration of the

FLUORESCENT

PROBE-METABOLITE

0.201 ./---

100 0.06

0.12

0.18

0.24

IIS

60

60

IO

5 Time

15

I 20

(min)

FIG. 2. Concentrationand time-dependence of effect of NAPQI on the fluorescence of fluo-3-loaded PBMC. Changes in fluorescence were monitored in unidirectional, vacuum filtration plates, using an automated protocol as described under Experimental Procedures. Data are means of eight replicate determinations (SE < 5%). Inhibition by the lower concentrations of NAPQI (3.9 and 7.8 &M) was minimal and hence these data have been omitted from the main figure for clarity. Control fluorescence (for 2.5 X 10’ cells incubated with Me,SO) was 6153 FU and was independent of time. The inset shows a Lineweaver-Burke plot of the data obtained at 5 min, where V is the rate of decrease in fluorescence and 5’ is the NAPQI concentration

(0-500pM).

solution. Hence, at a final concentration of 50 PM, the rate of NAPQI decomposition was some sixfold less than that at 500 PM (peaks I, II, III, IV, and V representing 95.66, 3.15, 0.84, 0.11, and 0.22% of the total peak height at 60 min). Measurement of the fluorescence of the samples generated from the NAPQI degradation studies at the excitation/emission wavelength pairs for the most widely used fluorescent calcium probes yielded valuable information (Fig. 7). Essentially, these data show that, with the exception of fluo-3, the decomposition of NAPQI resulted in the formation of products which fluoresced intensely at the wavelengths used to monitor [Ca2+], with calcium indicators. The relative contributions of acetaminophen, benzoquinone, and hydroquinone to this effect were neglible, confirming that the “polymers” were the major fluorescent species (data not shown).

257

INTERACTIONS

cium homeostasis is a critical early target for chemicals which cause hepatotoxicity. Depletion of intracellular glutathione via extensive oxidation and/or arylation by these reactive compounds is presumed to result in their interaction with critical protein thiols, which may then lead to a perturbation in [Ca2+li, altered cell morphology, and, ultimately, cell death (28). In addition, toxic elevations of cytosolic calcium may activate a number of catabolic processes, mediated by proteases, phospholipases, and endonucleases (29). Nicotera and colleagues (20) have recently suggested that N-acetyl-p-benzoquinone imine mediates the hepatotoxicity of the widely used analgesic acetaminophen through thiol depletion and an early, sustained elevation in [Ca2+],. Furthermore, data obtained with NAPQI and two dimethylated analogues supported earlier findings that, although both arylation and oxidation of protein thiols may result in an elevation in cytosolic calcium and cytotoxicity, arylation appears to be the more lethal insult (30). The initial aim of the present study was to determine whether similar events occur in PBMC challenged with this model reactive metabolite and to compare and contrast the results to those obtained with a number of other (pro-) reactive metabolites (such as SMX-HA and its nitroso derivative) in an attempt to delineate early cellular effects of drug metabolite toxicity. Preliminary experiments employed fluo3, one of a family of more recent fluorescent calcium indicators excitable at visible wavelengths (19), to monitor chemically induced changes in cytosolic calcium levels. However, although the results obtained confirm that this fluorescein analogue may be used as a sensitive marker for calcium fluxes produced in response to mitogen stimulation and ionophore addition (Fig. l), the applications of fluo-3 to the study of chemically induced perturbations in [Ca2+], may be limited because of its chemistry, Exposure of PBMC loaded with fluo-3 to NAPQI (250 PM) resulted in a marked, rapid decrease in fluorescence similar to that observed following the addition of a heavy metal ion to a calcium-fluorescent probe complex. Further studies indicated that this interaction was not due simply to competitive displacement of calcium from the probe but rather to a direct effect of NAPQI on fluo-3 (Fig. 3). This interaction decreased both the ability of fluo-3 to chelate calcium and its inherent fluorescent properties. The precise chemistry of this interaction remains to be elucidated but may involve inner-filtering by the colored NAPQI degradation products, intermolecular quenching, or covalent modification of the probe through a Michael addition.4 The finding that NAPQI quenched the fluorescence of fluo-3 in

DISCUSSION

In vitro studies with isolated hepatocytes (4-6) and subcellular fractions (26,27) have suggested that cal-

4 Dr. D. N. Butler, nication.

Dalton

Chemical

Laboratories,

personal

commu-

258

RILEY

* =

A

10

ET AL.

B

g, E ::

6

f4 0 a2 s

300

ld!LL

Go

200

250

300

350

Wavelrngth

o!

I. 5

0

.

Time

.

I 10

400

WSWength

(nm)

.

.

‘.

I

”

15

350

4OD

(nm)

I

0

5

Tlme

(mln)

15

10

(min)

FIG. 3. Effect of NAPQI on the fluorescence of fluo-3 free acid (10 FM) in PBS (A) or Ca’+/Mg”-free PBS containing EGTA (B). Flue-3 fluorescence was continuously monitored in a fluorescence spectrophotometer with an excitation/emission setting of 4851535 nm. The insets show typical fluorescence excitation spectra (emission at 436 run) obtained immediately after the effects of NAPQI (N, 250 PM) had been monitored. Each trace is representative of at least three separate experiments. Control fluorescence was 239 FU (A) and 10 FU (B) with a photomultiplier setting of 400 V.

the absence of calcium (Fig. 3) makes the latter suggestion the most likely of these alternatives. Indeed, this effect appeared to be a function of the chemical structure of the reactive metabolite investigated, since the addition of SMX-HA (100 or 500 PM) to fluo-3-loaded PBMC did not result in a significant change in the fluorescent signal (Fig. 1). Moreover, these results suggest that, although SMX-HA may in theory undergo redox

cycling within the target cells and/or arylate essential macromolecules (via the formation of the nitroso-derivative (31)), calcium homeostasis is not an early critical target in the manifestation of cytotoxicity produced by this proreactive metabolite. Subsequent studies showed that the addition of NAPQI to PBMC loaded with another Auoresent calcium probe, indo- 1, produced a rapid increase in Auores-

-s

0

2

4

6 Time

8 (min)

10

12

‘

0

2

I

,

6

4

Time

s-lb

1;

(mln)

FIG. 4. Effects of NAPQI on the fluorescence of indo-l-loaded PBMC. (A) NAPQI (N, 250 PM) was added to the cells once a stable baseline (B) Fluorescence-time profile had been achieved, EGTA (E, 4.0 mM) or MnCl, (Mn, 2.0 mM) was then added once the signal approached F-. for NAPQI added to buffer alone (phosphate-free Hepes, pH 7.4). Each trace is representative of duplicate determinations and the baselines shown represent 679 FU (A) and 76 FU (B).

FLUORESCENT

PROBE-METABOLITE

0.008 -

0.006 z % E 0.004 s 5 0

0.000 ULI~. , IL , !I jl , 0

5

10

15

Elution

time

25

20 (min)

FIG. 5. Typical chromatogram of reversed-phase HPLC tion of NAPQI decomposition products. Trace shown represents ucts isolated from incubations of 500 PM NAPQI (added Me&SO) in 0.1 M phosphate-buffered saline, pH 7.4, conducted min. Reactions were terminated by the addition of methanol:water (9O:lO) containing 2.0 mM ascorbic acid and the mixtures on ice. HPLC methodology is described under mental Procedures.

separaprodin 1% for 60 ice-cold placing Experi-

cence (to around F,,,,, for this dye) similar in terms of both its chronology and its magnitude to that observed in isolated hepatocytes, using quin-2 as the calcium indicator (20). However, an almost identical fluorescence-time profile could also be obtained in the absence of both target cells and the dye (Fig. 4). Preliminary scans suggested that NAPQI and/or products of its decomposition may fluoresce intensely at the wavelength settings used for a number of commonly used calcium probes.

The decomposition of NAPQI in aqueous buffer has been previously reported and involves a complex sequence of reactions, including reduction, hydrolysis, and radical coupling reactions (25). In these earlier studies, the half-life for NAPQI decomposition was estimated at 11 min. HPLC analysis of the samples generated from the degradation of NAPQI in phosphate-buffered saline confirmed this rapid decay (Fig. 6) and revealed the presence of several products. These were tentatively identified as acetaminophen polymers previously characterized as products of acetaminophen oxidation catalyzed by horseradish peroxidase (32), cytochrome P450 (33), and prostaglandin H-synthetase (34). The data shown in Fig. 6 suggest that the rate of NAPQI degradation observed may have been significantly less than that reported by Dahlin and Nelson (25); however, it should be borne in mind that peak I probably represents acetaminophen produced via disproportionation of two semiquinone imine radicals and/or reduction of NAPQI by ascorbic acid and, hence, the rate of decrease in peak I is not an accurate reflection of the rate of NAPQI decomposition. These “acetaminophen polymers” fluoresced intensely at the excitation/emission wavelengths used for quin-2, fura-2, andindo-1 (Fig. 7). Furthermore, the fluorescentproperties of these polymerization products were such that they selectively (but not exclusively) fluoresced at the wavelength pairs used to monitor the calcium-bound form of the dyes, which would therefore preclude the use of a ratio method of calibration. In summary, the data presented herein demonstrate that the most widely used, commercially available fluorescent probes cannot be used to monitor perturbations in cytosolic calcium levels produced by NAPQI. For fluo-3, this highly reactive intermediate interacts chemi-

12

&30!

,

0

30

.

,

,

,

,

60

90

120

150

Time FIG. 6.

Time (B) Concomitant

course for the decomposition time-dependent increase

(min)

.

259

INTERACTIONS

1

( 180

0

30

60

90

Time

120

150

180

(min)

of NAPQI (500 pM) in aqueous buffer. (A) Decrease in peak I (predominantly in peaks II, III, IV, and V (tentatively identified as acetaminophen polymers).

acetaminophen).

260

RILEY

ET

AL.

B g

10000

c a r

6000

& 2 P 5

6000

c

4000

iI ts

2000

E a 0

30

60 flme

90

120

150

+

+

-

Quill-2

160

-

+

Fun-2

(min)

Flue-3

IlldO-

Fuorescent

probe

FIG. 7. Fluoresence of the samples generated from the decomposition of NAPQI at the excitation/emission wavelengths used for a number of commonly used calcium probes. (A) Detailed fluorescence-time profile conducted at the settings used for the calcium-bound (0) and the calcium-free (M) forms of quin-2. Identical profiles were observed for furaand indo-1. (B) The fluorescence of products formed from NAPQI degradation (after 20 min) at the wavelengths for the calcium-bound (+) and calcium-free (-) forms of several fluorescent calcium indicators (see Ref. (24)). *denotes that the actual fluorescence reading was off-scale (>9999).

tally with the fluorescent probe. The chemical interaction appears complex and not characteristic of proreactive or reactive metabolites per se as neither SMX-HA nor its nitroso-derivative exhibited this effect. In addition, NAPQI readily decomposes in aqueous buffer (and, to a lesser degree, Me,SO) to yield products (presumably polymers) which interfere with the measurement of intracellular calcium by quin-2, fura-2, and indo-1. Antioxidants such as ascorbic acid, glutathione, and NADPH (33) have been shown to prevent this polymerisation in vitro and therefore it is conceivable that the protective role of dithiothreitol against NAPQI-induced elevations in [Ca2’], (20) may have been due to its ability to inhibit the formation of these polymers. Hence, these results suggest that the early, sustained rise in [Ca2+], previously observed in response to NAPQI (20) may have been, at least in part, an artifact of NAPQI polymerization rather than a consequence of inhibition of the plasma membrane calcium pump, although they do not conclusively exclude the possibility that intracellular calcium homeostasis is a critical target for NAPQI-induced cytotoxicity (7,16). The fluorescent properties of the acetaminophen polymers may simply preclude an accurate assessment of the effects of NAPQI on cytosolic calcium through the use of fluorescent probes, such as quin-2. Evidently, further work is required to elucidate the precise mechanism by which NAPQI produces a disruption in [Ca2+],, and the results presented in this communication suggest that such studies may necessitate the use of indirect biochemical markers of calcium regulation.

REFERENCES 1. Horning, Environ.

E. C., Thenot, J. P., and Helton, Health 4, 341-361.

2. Guengerich,

F. P., and Liebler,

E. D. (1978)

D. C. (1985)

CRC

J. Toxicol.

Crit.

Reo. TOG-

col. 14, 259-307. 3. Parke,

D. V. (1987)

4. Moore,

L. (1980)

Arch.

Toxicol.

Biochem.

60,515.

Phnrmacol.

29,2505-2511.

5. Jones, D. P., Thor, H., Smith, M. T., Jewell, S. (1983) J. Biol. Chem. 288,6390-6393. 6. Nicotera, Orrenius,

S. A., and Orrenius,

P., McConkey, D., Svensson, S.-A., S. (1988) Toxicology 52, 55-63.

7. Moore, Orrenius,

M.,

Bellomo,

G., and

Thor, H., Moore, G., Nelson, S., Moldeus, S. (1985) J. Biol. Chem. 260, 13,035-13,040.

P., and

8. McConkey, D. J., Nicotera, P., Hartzell, P., Bellomo, A. H., and Orrenius, S. (1989) Arch. B&hem. Biophys.

G., Wyllie, 269,365-

370. 9. Kerr, J. F. R., Wyllie, Cancer 26,239-257. 10. Wyllie,

A. H. (1987)

11. Pritchard,

A. H., and Arch.

A. R. (1972)

Sll,

Toricol.

D. J., and Butler,

Currie,

Brit.

J.

3-10.

W. H. (1989)

J. Pathd.

P., Nicotera,

P., and

168, 253-

260. 12. McConkey, D. J., Hartzell, (1989) FASEB J. 3,1843-l&19. 13. Spielberg,

14. Riley, Park,

S.

P. (1980)

J. Phnrmacol.

Exp.

16. Boobis, A. R., Tee, L. B. G., Hampden, (1986) Food Chem. Toxicol. 24,731-736. Parker,

C. W. (1982)

Phurmacol.

N. R., and

A., and Spielberg, C. E., and Davies,

Reu. 34,85-104.

S.

213,395-398.

R. J., Maggs, J. L., Lambert, C., Kitteringham, B. K. (1988) Brit. J. Clin. Pharmncol. 26,577-566.

15. Leeder, J. S., Cannon, M. M., Nakhooda, (1988) Phurmncol. Erp. Ther. 248,956-962.

17.

Ther.

Orrenius,

S.

P.

D. S.

FLUORESCENT 18. Park, them.

PROBE-METABOLITE

B. K., Coleman, J. W., and Kitteringham, Pharmacol. 36,581-590.

19. Kao, J. P. Y., Harootunian, Chem. 264,8179-8184.

N. R. (1987)

A. T., and Tsien,

R. Y. (1989)

20. Nicotera, P., Rundgren, M., Poruhek, D. J., Cotgreave, deus, P., Orrenius, S., and Nelson, S. D. (1989) Chem.

Bio-

J. Biol.

I., MolRes. Toxi-

col. 2, 46-50. 21. Boyum, Lamgone, Academic

A. (1984) in Methods in Enzymology J. J., and Van Vunakis, H., Eds.), Vol. Press, San Diego, CA.

(Di Sahato, G., 108, pp. 88102,

22. Leeder, J. S., Dosch, H.-M., Harper, P. A., Lam, berg, S. P. (1989) Anal. Biochem. 177, 364-372.

23. Potter, Chem. 24. Cobbold, 25. Dahlin,

886.

D. W., Miller, 260,12,174-12,181.

D. W.,

P. H., and Rink, D. C., and Nelson,

and Hinson,

T. J. (1987) S. D. (1982)

P., and Spiel-

J. A. (1985)

Biochem. J. Med.

J. Biol.

J. 248,313-328. Chem.

26,885-

261

INTERACTIONS

26. Bellomo, G., Mirabelli, F., Richelmi, FEBS Lett. 163,136-139.

P., and Orrenius,

S. (1983)

27. Thor, H., Hartzell, P., Svensson, S.-A., Orrenius, S., Mirabelli, F., Marinoni, V., and Bellomo, G. (1985) B&hem. Phurmncol. 34, 3717-3723. 28. Orrenius, S., and Nicotera, P. (1987) Arch. Z’oxicol. Sll, 11-19. 29. Orrenius, S., McConkey, D. J., Bellomo, G., and (1989) Trends Pharmacol. Sci. 10,281-285. 30. Gant, T. W., Rao, D. N. R., Mason, R. P., and Cohen, Chem. Biol. Interact. 65, 157-173. 31. Kappus, H. (1987) Arch. Toxicol. 60, 144-149. 32. Potter, macol. 33. Potter, 973. 34. Potter, 980.

D. W., Miller, D. W., and Hinson, J. A. (1986) 29,155-162. D. W., and Hinson, J. A. (1987) J. Biol. Chem. D. W., and Hinson,

J. A. (1987)

J. Biol.

Chem.

Nicotera,

P.

G. M. (1988)

Mol.

Phur-

262,966262,974-