Hydroperoxide-Metabolizing Systems in Rat Liver

Eur. .I. Biochem. 57. 503-512 (1975)

Hydroperoxide-Metabolizing Systems in Rat Liver

Helmut SlES and Karl-Heinz SUMMER

Institut fur Physiologische Cheinie und Physikalische Biochernie der Universitiit Miinchen

(Received March 24, May 21, 1975)

1. Metabolism of added hydroperoxides was studied in hemoglobin-free perfused rat liver and in isolated rat hepatocytes as well as microsomal and mitochondria1 fractions.

2. Perfused liver is capable of removing organic hydroperoxides [cumene and fevt-butyl hydro- peroxide] at rates up to 3-4 pmol x rnin-' x gram liver-'. Concomitantly, there is a release of glutathione disulfide (GSSG) into the extracellular space in a relationship approx. linear with hydro- peroxide infusion rates. About 30 nmol GSSG are released per pmol hydroperoxide added per min per gram liver. GSSG release is interpreted to indicate GSH peroxidase activity.

3. GSSG release is observed also with added H202. At rates of H202 infusion of about 1.5 pmol x min-' x gram liver-' a maximum of GSSG release is attained which, however, can be increased by inhibition of catalase with 3-amino-l,2,4-aminotriazole.

4. A contribution of the endoplasmic reticulum in addition to glutathione peroxidase in organic hydroperoxide removal is demonstrated (a) by comparison of perfused livers from untreated and phenobarbital-pretreated rats and (b) in isolated microsomal fractions, and a possible involvement of reactive iron species (e. g . cytochrome P-450-linked peroxidasc activity) is discussed.

5 . Hydroperoxide addition to microsomes leads to rapid and substantial lipid peroxidation as evidenced by formation of thiobarbituric-acid-reactive material (presumably malondialdehyde) and by O2 uptake. Like in other types of induction of lipid peroxidation, malondialdehyde/O, ratios of 1/20 are observed. Cumene hydroperoxide (0.6 mM) gives rise to 4-fold higher rates of malondial- dehyde formation than tert-butyl hydroperoxide (1 mM). Ethylenediamine tetraacetate does not inhibit this type of lipid peroxidation.

6. Lipid peroxidation in isolated hepatocytes upon hydroperoxide addition is much lower than in isolated microsomes or mitochondria, consistent with the presence of effective hydroperoxide- reducing systems. However, when NADPH is oxidized to the maximal extent as evidenced by dual- wavelength spectrophotometry, lipid peroxidation occurs at large amounts.

7. A dependence of hydroperoxide removal rates upon flux through the pentose phosphate path- way is suggested by a stimulatory effect of glucose in hepatocytes from fasted rats and by an increased rate of I4CO2 release from [l -14C]glucose during hydroperoxide metabolism in perfused liver.

In biological systems, hydroperoxides not only occur as an exception such as an intermediate in particular enzymatic reactions but also as a rather ubiquitous attribute of oxygen metabolism. This includes hydrogen peroxide as well as organic hydro- peroxides, and in recent years a number of metabolic effects of such compounds and the enzymatic systems dealing with the reduction of these to water or alcohols, respectively, have been elucidated. Thus, in an intact tissue, hemoglobin-free perfused rat liver, the existence

Enzymes. Glutathione peroxidase (EC I . I I . I .9) ; glutathione reductase (EC 1.6.4.2) ; glucose-6-phosphate dehydrogenase (EC 1.1.1.49); catalase (EC 1.11.1.6).

~~ ~-

of a steady-state rate of H202 generation was demon- strated by spectrophotometry of catalase compound I [I], and rates of H202 generation were calculated [2]. (For the present knowledge on the role of catalase, a peroxisomal enzyme, see a recent review [3,3a]). With respect to organic hydroperoxides, a field that is more complicated due to the diversity of possible reactants, interest was focused largely on lipid per- oxidation and membrane damagc in a number of studies [4- 81, and a central role of the selenoenzyme, glutathione peroxidase [9 - 111, was established for the reduction of organic hydroperoxides [ll , 12,12a]. This also has been subject of a review [13]. In addition,

Eur. J. Biochem. 57 (1975)

504 Hydroperoxide-Mctabolizing Systems in Rat Liver

organic hydroperoxides have increasingly been utiliz- ed in studies with microsomal fractions 114- 161.

Due to the fact that catalase does not react with organic hydroperoxides such as tert-butyl or cumene hydroperoxide, the use of such compounds allows one to study hydroperoxide-metabolizing systems other than catalase in the intact cell. In some respects, externally added organic hydroperoxides may qualify as model lipid hydroperoxides in metabolic studies. However, i t is evident that differences exist between the metabolic role of endogenously generated and externally added hydroperoxides. Nevertheless, a number of metabolic perturbations have been shown to occur upon addition of hydroperoxides to rat liver, and the intervention of glutathione peroxidase was demonstrated by (a) a release of glutathione disulfide (GSSG) froin the cells, and (b) by a substantial oxida- tion of NADPH, most probably due to the activity of GSSG reductase [17,3 81.

In the present investigation, the hydroperoxide removal capacity of rat liver was measured. In addi- tion, it will be shown that besides glutathione per- oxidase another activity associated with the endo- plasmic reticulum operates in hydroperoxide metabo- lism in the intact cell. With respect to the latter, it has been shown recently that cytochrome h, undergoes a substantial oxidation upon addition of hydro- peroxides to hepatocytes [19], and it is possible that this involves the so-called NADH peroxidase of isolated microsomes 1141. However, the concomitant formation of substantial amounts of lipid peroxide products in microsomes indicates that additional reactions take place simultaneously.

MATERIALS AND METHODS

Hemoglobin- Free Liver Perfusion (Open System)

Livers from male Wistar rats of 150 - 180 g body weight, fed on stock diet (Altromin), were perfused as described previously 119,201. Perfusate flow was 3.5 - 4.0 ml/min per gram of liver, the temperature was 36.5-37 C. The perfusion fluid was the one used previously: 13 5 mM NaCI, 5.9 mM KCI, 1.2 mM MgCl,, 1.2 mM NaH,PO,, 1.2 mM Na,SO,, 2.5 mM CaCI,, 25 mM NaHCO,, equilibrated with a CO,/O, mixture (5:95, v/v), and L-lactate and pyruvate as sodium salts, 0.3 mM each. Stepwise additions of substances were performed by infusion of stock solutions into the perfusate using precision micro- pumps. Effluent perfusate was assayed directly for lactate, pyruvate, lactate dehydrogenase, glutathione disulfide, and hydroperoxides.

Isolation of Hepatocytes

Hepatocytes were isolated from rat livers perfused with clostridial collagenase as described previously 1191, a method deriving from the original suggestion of Berry and Friend 1211. The initial perfusion medium was as above except that CaCI, was replaced by 10 mM glucose. After first washing the cells in the same me- dium and filtering them through a gauze filter, hepato- cytes were washed two times further in a medium as above but containing 10 mM glucose and 2.5 mM CaCI,. Hepatocytes were stored on ice. Throughout the experimental period a biosynthetic process, urea formation from 5 mM NH,Cl, operated linearly for 15 min at a rate of 1.2- 1.4 pmol/min per gram wet weight of cells.

lkperiments witlz Hepatocytes

An aliquot of the cell suspension was added to a I-cm glass cuvette containing an incubation buffer at 37 "C. The cuvette was then placed in a temperature- regulated cuvette holder, and the suspension was stirred magnetically. The incubation buffer consisted of the perfusion fluid named above, and glucose was present when indicated. Total volume was 1.5 ml, approx. 6- 12 mg of hepatocyte protein added, open systems with respect to gas exchange. Dual-wavelength recordings were performed with a special photometer designed and constructed by Dr H. Schwab and co- workers (Electronics Department, Sonderforschungs- bereich Medizinische Molekularbiologie und Bio- chemie, Munich, see [22]). Samples were assayed for glutathione disulfide, hydroperoxides and thiobarbi- turic acid reactive material.

Preparation of Mitochondria1 and Microsomal Fractions

Mitochondria were prepared according to 1231, and microsomes were obtained by centrifugation of the 8700 x g supernatant at 105 000 x g for 60 min, then resuspended in 50 mM tris(hydroxymethy1)- aminomethane, 150 mM KCI, pH 7.4, and washed once.

Pretreatment o f Rats

Phenobarbital pretreatment was performed as previously [19] by addition of phenobarbital to the drinking water (1 mg/ml) for at least 7 days. 3-Amino- 1,2,4-triazole pretreatment was performed as pre- viously [19], injecting 1 g per kg 1 h prior to perfusion.

Eur. J. Biochem. 57 (1975)

H. Sies and K.-H. Summer 505

Assuys

L-Lactate, pyruvate, NHZ, urea, and lactate de- hydrogenase activity were determined in enzymatic optical tests using NADH absorbance measurement, based on the procedures laid out in [24]. Similarly, the concentrations of GSSG and of the hydroperoxides were determined in optical tests. GSSG was assayed by addition of 0.6 U of yeast glutathione reductase to a mixture containing 0.5 ml sample, 0.2 ml tri- ethanolamine (0.1 M) and EDTA (I mM), pH 7.6, and NADPH in a final concentration of 30-50 pM. By following the absorbance decrease at 340 nm either in a Hitachi (model 181) or Perkin-Elmer (model 156) spectrophotometer, perfusate GSSG con- centrations of 1 pM could be accurately determined. Recovery of added GSSG was 95- 105 x.

Hydroperoxides were assayed in a coupled optical test as used previously 1171. Usually, the reaction was initiated by addition of glutathione peroxidase in an activity sufficient to bring the assay to completion within 5 - 8 min, the reaction mixture containing 100 mMpotassiumphosphate buffer,pH 7.0,0.15 mM NADPH, 0.3 mM reduced glutathione, and 0.6 U of yeast glutathione reductase. Absorbance decrease at 340 or 366 nm was observed.

The estimation of lipid peroxidation relied on the determination of thiobarbituric-acid-reactive material [25]. It was carried out by addition of 0.5 ml sample to 0.9 ml of 35 mM thiobarbituric acid in 1.2 M tri- chloroacetic acid. The reaction mixture was incubated for 25 min at 95 "C, and after cooling on ice for approx. 30 min and centrifugation, absorbance of the super- natant at 535 nm was measured in a Hitachi spectro- photometer (model 181). Data are expressed as malondialdehyde formed, using E = 156 mM-' tin- ' . In the range studied, malondialdehyde formation was found proportional to sample protein. Protein was determined with a biuret method using bovine serum albumin as standard.

Oxygen concentration was monitored with a Clark-type 0, electrode. In experiments in a closed vessel, rates of 0, uptake were recorded simultaneously with the concentration using a differentiator circuit ( c f Fig. 9).

'"CO, release from [l-'4C]glucose (cf Fig. 11) was measured as previously [20]. The perfusate contained 10 mM glucose, radioactivity was 8.5 nCi/ inl. Rates of I4CO2 release were calculated on the basis of the specific activity of glucose in the perfusate.

Chem icals

tert-Butyl and cumene hydroperoxides as well as di-tert-butyl peroxide were gifts from Peroxid Chemie

Munchen GmbH (Hollriegelskreuth). 3-Amino-l,2,4- triazole was from Fluka (Buchs). Reduced glutathione was a gift from Pharma-Waldhof GmbH (Mann- heim). [1-'"C]Glucose was from NEN Chemicals GmbH (Dreieichenhain).

Bovine erythrocyte glutathione peroxidase was generously provided by L. Flohe, A. Giinzler and A. Wendel (Physiologisch-chemisches Institut, Uni- versitat Tiibingen).

Other chemicals were from Merck (Darmstadt) and biochemicals were from Boehringer (Mannheirn).

RESULTS

Hydropero.de Uptake by Perfused Liver

As shown in Fig. 1, perfused rat liver is capable of removing added organic hydroperoxides, such as tert-butyl hydroperoxide, from the perfusate, as has been previously demonstrated [17]. Furthermore, it was repeatedly observed that effluent hydroperoxide concentration was initially high before coming to a minimum at about 3 min when influent hydroperoxide was as high as 1 mM (Fig. 1). From influent-effluent concentration differences and perfusate flow rate, maximal rates of hydroperoxide uptake were obtained. Data collected from 21 different perfusions with either tert-butyl or cumene hydroperoxide are shown in Fig. 2, demonstrating rates of hydroperoxide removal up to approx. 3 pmol x min- x gram liver-'. No systematic differences were observed for the two hydroperoxides. When high concentrations of hydro-

Eur. .I. Biocheni. 57 (1975)

Hydropcroxide-Metabolizing Systems in Rat Liver

8 0 0

0

0

0 *o 0

0 0

I 1 I

Hydroperoxide infused ((imol . min-' . g liver-')

peroxides were sustained for extended periods, there occurred secondary eKects like spotting of the liver surface, and it was not found advisable, therefore, to further consider the metabolic responses at rates higher than 3 pmol x iiiin-' x grain liver-'. Effluent lactate dehydrogenase activity was 1 - 2 mU/ml during hydroperoxide infusions up to 8 inin.

Hydroperoxide removal is strongly temperature- dependent. In the experiment shown in Fig. 3, the perfusate temperature was shifted from 37 C to 32 C for a short interval, and it can be seen that effluent hydroperoxide concentration rises during this interval. While temperature dependence was not studied systematically, it explains the lower rate of hydroperoxide removal observed previously at 33 ' C

71.

GSSG Re l~asc

There is a release of GSSG from the liver cells during metabolism of added hydroperoxides [ 171, a phenomenon interpreted to indicate the activity of glutathione peroxidase. However, relatively little is known at present on the mechanism and properties of GSSG transport froin the cells. As shown in Fig. 1. CSSG release is reversible. When the inaxiinal rates of rclcase of GSSG, attained 3 to 5 min after the onset of hydroperoxide infusion, are plotted against the hydroperoxide infusion rates, an approximately linear relationship is observed for rcrt-butyl and cumene hydroperoxides (Fig. 4). About 30 ninol of GSSG are maxiinally released from the liver per min per grain when 1 pmol of hydroperoxide is removed per inin pcr gram. The linear relationship observed in

I terf- Butyl hydroperoxide 1 . I mM

Fig. 4 suggests that the rate of GSSG release into the extracellular space may be indicative of, although not necessarily equal to, the intracellular concentration of GSSG. On an average, the rate of 30 ninol GSSG x inin-' x gram liver-' corresponds to 10 pM GSSG in the effluent.

It is of interest to note that H,O, when added to the perfusate also leads to a release of GSSG from rat liver [17]. With the lower rates of H202 infusion, no differences were observed with respect to GSSG release when compared to the organic hydroperoxides (Fig. 4). However, at rates > 1.5-2 pmolxmin- ' xgram liver-', no further increases above a rate of about 40 iiinol GSSG x min-' x gram liver-' were found. This is taken to indicate the intervention of catalase which does not allow a further increase of effective H,O, concentration at the site of glutathione peroxi- dase. Support for this conclusion is given by the observation of an augmented rate of GSSG release in perfused livers pretreated with 3-amino-1,2,4- triazole 1 h prior to perfusion. Under these conditions, >95':,, of catalase is inactive [27,28]. At a rate of H,O, infusion of 1.4 pmol x m i f ' x gram liver-'. GSSG release was 30 and 45 ninol x inin-' x gram liver^.' in livers from untreated and 3-ainiiio-l,2,4- triazole-pretreated rats, respectively.

When H20, is generated within the cells by infu- sion of sodium glycolate, there is 110 extra release of

H. Sies and K.-H. Surnincr 507

0

0

0

A

A A O A

I .

0 ? * I I I

0 1 2 3 Hydroperoxide infused ( p o i ’ rnin-’ ‘ g liver-’)

Fig. 4. Depaiclr.ticc~ uf die rufc of GSSG ri~1i~a.w upon the rate of Ii~.~lroprrosick. infusion in prr;/u.setf rui l i w . Maximal GSSG concen- trations obtained about 3 - 5 rnin aftcr onset of hydroperoxide infusion were used. rert-Butyl (0). cumene (0) hydroperoxide, and H,O, (A) data are shown

GSSG into the perfusate observable that would be clearly distinguishable from the controls (Fig. 5) , although an extra rate of H202 formation of about 450 nmol x min-’ x gram liver-’ has been calculated from the steady state occupancy of catalase compound I [2]. However, in contrast to this a marked CSSG release was observed with glycolate by N. Oshino (personal communication) with Sprague-Dawley rats.

GSSG Releuse und Microsomal Sj’stems

Per gram of liver, maximal rates of GSSG release from livers of phenobarbital-pretreated rats are only about goy,, of those observed with livers from un- treated rats, using tut-butyl hydroperoxide (Fig. 6). While such a difference may simply result from the selective proliferation of components of the endo- plasmic reticulum, leaving the total liver content of glutathione peroxidase unaltered, some additional observations were made which might implicate the endoplasmic reticulum as a reactant site in hydro- peroxide-linked reactions. For example, when cumene hydroperoxide was used instead of tert-butyl hydro- peroxide (Fig. 7), the profile of GSSG release during a continuous infusion of this hydroperoxide exhibits no plateau but rather a decline subsequent to a maximal rate of release about 3 min after onset of

Glvcolate tert-Butvl hydroperoxide

I I I 20 30 40

Perfusion time (min)

Fig. 5 . GSSG coni~cntrarioti iti efflncni jwi:fuseitc~ duritix infusion of sodium gl jcdn /c3 und tert-bzirj./ hjdrupcro ~idc’

tert - Butyl hydroperoxide

t-*-*,*.*.*.* i

20 I I

I I 0 20 30

-

tert- Butyl hydroperoxide (0.56 mM)

1

*.*- 2 I

20 30

A

50

0

B

. . 20 30

perfusion time (min) x) 30

Eur. J . Biochem. S7 (1975)

508

0.4

0.3

- 5 1

a, * x .- 0 d 0 0.2 -D, i

a,

E I3

0.1

0 l

0 5 10 Time (minj

Fig, 8. Drcriusr of ciniii'iic l i ~ ~ t l r o j ~ c ~ r o s i r l c ~ concentrution (inirially 0.41 riiM) iir p r i w w v of nzicrosornes (2.1) nig prot i~i i~l tnl) f1.oi?l uw- frcuferl i n ) und p h e r i o h u r h i t u l - ~ ~ r ~ t ~ c ~ ~ ~ ~ c ~ ~ l (0) rut. The incubation mixture contained 10 mM morpholinopropane sulfonic acid, 154 mM NaCI. 6.2 mM KCI, 4 tnM nicotinamidc, and an N A D H - regenerating systcin (L-lactate, 6 mM. NAD'. 0.2 mM, lactate dehydrogenase 0.15 U). pH 7.3. 37 C

the infusion. Furthermore, when a similar experiment is performed with a liver from a phenobarbital- pretreated rat (Fig. 7B), maximal rates of GSSG release are only 50-60% of those observed with a liver from an untreated rat. This striking difference in the response to added cumene hydroperoxide obtains at similar rates of hydroperoxide removal (not shown). There are several possible explanations for such a difference, ranging from differences in the GSSG transport process, in the NADPH and GSH regeneration capacities as well as in the contribution of microsoinal systems to hydroperoxide removal. Evidence for the latter possibility was provided by studying hydroperoxide removal in subcellular frac- tions. Because of the more pronounced differences with cumene hydroperoxide (Fig. 7), and also because this hydroperoxide appears to be the one more widely applied by investigators working with microsomal fractions, cumene hydroperoxide reinoval was follow- ed in KC1-washed microsoinal fractions. As shown in Fig. 8, there is a rapid decrease of cumene hydro- peroxide concentration in the presence of an NADH- regenerating system, and activity is considerably higher with inicrosomes obtained from phenobarbital- pretreated rats than with those obtained froin un- treated rats. Initial rates of hydroperoxide consump-

Hydroperoxidc-Metaboliziiig Systems in Rat Liver

Cumene hyd roperoxide

(0.6 mM)

I I I I 0 1 2 3

Time (min)

Cumene hydroperoxide

(0.6 mM)

I I I I 0 1 2 3

Time (minj

Fig. 9. O.\-!:ycii upfake Lipoil uddifioii of c .u i i i enc~ IrJ,u'ro()er.ositff, h.i, l i w r niicro.sornc.s ,/i.oiw uizlrerztrd i A ) mid pl~c~iioburhitul-prefr.errtcu' i B ) rurs. Incubation condition as in Fig. 8 except that no NADH- regcnerating system and no nicotinamide was present

tion are approx. 3 times higher in the former, amount- ing to > 300 nmol x inin-' x mg microsomal pro- tein- * . Glutathione concentration under these condi- tions was less than 0.7 pM, making a participation of glutathione peroxidase in the washed microsomal fractions unlikely, also in view of the observations by Flohe and Schlegel [29] that this enzyme belongs to the soluble fraction (and the mitochondria1 matrix space) of the cell. Since in the experiment of Fig. 8 the inicrosomal protein was 2 mgjml in both cases, the observation of a decreased rate of GSSG release in perfused liver from phenobarbital-pretreated rats inay be explained by an enhanced contribution of the

Eur. J . Biochciii. 57 (1975)

H. Sies and K.-H. Summer

0 -

-0.02 - B 8 -x a

-0.04 -

-0.06 -

509

endoplasmic reticulum in hydroperoxide removal. This is further substantiated by the finding of a more pronounced oxygen uptake upon cumene hydro- peroxide addition to microsomes from phenobarbital- pretreated rats compared to those from untreated rats (Fig. 9).

The differences between the two hydroperoxides with respect to the contribution of microsomes was also demonstrated by a higher rate of lipid peroxida- tion. Thiobarbituric-acid-reactive material was found to accumulate more rapidly with cumene hydro- peroxide than with tert-butyl hydroperoxide. Although problems may arise in comparisons between different fractions (see [30]), calculation of malondialdehyde formation was performed (Table 1) in order to facili- tate a comparison with the rates of oxygen uptake. Surprisingly, the cumene hydroperoxide-induced lipid peroxidation and oxygen uptake rates (Table 1 and

Table 1. Lipid peroxidation upon hyiroperoxidtJ addition in rat liver fi.ac.r ions

Peroxide Rat liver fraction Malondialdehqdi. formed at 2 iiiiii

nmollmg protein

terr-Hutyl hydro- microsomal 3.9 peroxide 1 mM mitochondria1 0.4

hcpatocyte < 0.2

Cumene hydro- microsomal 11.5 peroxide 0.6 mM plus EDTA, 2 mM 12.1

mitochondria1 3.1 O2 absent 0

hepatocyte 1 .o 0, absent 0

Di-terr-butyl per- microsomal 0 oxide 1 mM

Fig. 9B) are quite similar to those obtained by Hochstein and Ernster [4,31] in their study on ADP- iron activated lipid peroxidation coupled to the NADPH oxidase system of microsomes. However, EDTA (2mM) was not inhibitory with the hydro- peroxides. Hydroperoxide-induced lipid peroxidation was not observed when O2 was excluded, either by equilibration with N2 or by addition of sodium dithionite. Also, no thiobarbituric-acid-reactive ma- terial was found when di-tert-butyl peroxide was used with microsomes.

Hydroperoxide Metabolism in Isolated Hepatocytes

As shown in Table 1, lipid peroxidation upon hydroperoxide addition to hepatocytes is lower than that of particulate fractions, consistent with the presence of defense mechanisms. A suitable parameter to follow the time course of hydroperoxide metabolism in the intact cell appears to be the monitoring of nicotinamide nucleotide absorbance. We have previ- ously reported that the transitions in the nicotinamide nucleotides are predominantly ascribable to the NADP, not the NAD, systems, both by the indicator metabolite method and by measurement of tissue levels of the nucleotides [17,18]. When absorbance difference (350- 380 nm) is recorded against time, the addition of tert-butyl hydroperoxide (Fig. 10) results in a rapid decrease of absorbance difference followed by a return to the steady state after the hydroperoxide is metabolized. The time interval required is dependent on the concentration of the hydroperoxide added. From traces such as those shown in Fig. 10, the time needed for the absorbance difference to return to the half-excursion (recovery time) is shown in Table 2. Interestingly, lipid peroxida-

terf -Butyl hydropemxide 0.25mM 0.5 mM 1 .O mM

1 1 1 f-

t 1 10 min

Fig. 10. Time course uf N A D P H o.ui~lation-rrduction chunges upon addition of tert-hutyl hydroperoxide to isolated hepatocytes, as monitored by ahsorbunce dfierence .spectroplzotomet~y, For evaluation of results, see Table 2 and text

Eur. J . Biochem. 57 (1975)

510 Hydroperoxide-Metabolizing Systems in Rat Liver

Titble 2. ~ ~ i ( ' O t i i l i ~ i i titkc, iIl/c,/i,otic/c, i) \ - i t / t / t io i i ri i l t l l ipid / l ~ " . o . ~ i t l t i l i r ~ / i iil ?,to/t i l i , t / / i ( ' / J C i / ( l ( ' J / ~ ' ~ S L I ~ I I ? iiddiIiO/t of t-hliti.1 /~?, i lr .r~/~c,ros/ t /c Hepatochtes (6.8 ing per nil) wci-c suspended a s described in Material\ and Methods. The basal iuediuin containcd 10 inM glucose ~nilcss stated otherwise. Nicotinamide nucleotidc oxidation was followed by absorbance changes a t 350- 380 nin. and ds of 4.7 mM- cin ' ivas used for this wavelength pair. Some of the tra ire sliown in Fig. 10. At 9 min. incubations were terminated and thiobarbituric acid reactive inaterial determined: ii\ing i _ = 156 i i iM - ' ciii - I ;it 535 nm. d a t a aIc rcprcscntcd a \ inalondinldehq.dc fornictl 11 d . . not dcterniined

~ t i t r i t i o n ~ i l \ ~ ~ i t c ror/-Kut) I hydropcroxidc Nicotinaiiiidc nucleotidc Recovery time Llpld pe lw lda t ion concentration added o x i da t ion malondialdeliqdc

formed ;it 0 min

111 M

Fed 0.25 0.5 1 .o 2.0

Fas led (24 11) 0.4 I 1 0 glLlcosc 0.4

nniol, nip protein min

1 .9 2.1 1.8 5.6 1 .9 9.6

1.5 4.2 1 5 7.7

1.5 (at peak) 0.9

nmol mg protein

0.02 0.07 0.25 0.74

n.d. I1.d.

,' Tnkcn at 1 1 n i in

fert ~ Bu t yl hydroperox ide

I 0 \

tion is not proportional to the duration of nicotin- amide nucleotide oxidation but increases relatively more with the concentration of added hydroperoxide.

The recovery time for nicotinainide nucleotide oxidation is almost doubled in the absence of added glucose with hepatocytes isolated from rats fasted for 24 h prior to the experiment (Table 2). The dependence on glucose suggests an involvement of oxidative metabolism of glucose in the provision of reducing equivalents. Further support for a participation of the pentose phosphate pathway is given by the observa- tion of a reversible increase of labeled CO, release from [l-14C]glucose upon addition of fwt-butyl hydro- peroxide to perfused liver (Fig. 11).

DISCUSSION

Usc of' Lylernully A d d d Hjdroperoxiu'es it1 Inttict System

The study of hydroperoxide metabolism was brought to a new level by the discovery by Mills [9] of an enzyme capable of reducing hydrogen peroxide at the expense of reduced glutathione, and by the observations of Little and O'Brien [11] and Christo- pherscn [12] that organic hydroperoxides wive as substrates as well as H,02 for this enzyme, glutathione peroxidase. The opportunity of investigating hydro- peroxide metabolism ilia glutathione peroxidase in the intact cell in the presence of catalase by use of organic hydroperoxides was exploited by Sies ef ul. [ I 71 with hemoglobin-free perfused rat liver. Sub- sequently, this was applied also to other types of cell, e g . erythrocytes [32].

It is evident that by coupling with GSSG reductase a number of metabolic reactions occur, notably in the NADPHINADP and 2GSHIGSSG systems. Some of these have been discussed previously [17,18], and in the present paper this subject will not be discussed in detail. Rather, it is attempted to demonstrate that, similar to the probleins occurring with endogenously formed lipid hydroperoxides, a full comprehension of the metabolic responses to added hydroperoxides may be hampered by the complexity of possible reactions.

However, it should be first pointed out that a short-term perturbation of liver cell metabolism, e.g. by an infusion of 0.3 - 0.6 mM felt-butyl hydro- peroxide for 5 min, leads to fully reversible changes. and a s such qualifies to probe for effects that sub- stantial changes in the NADPH and GSH redox state may have on cellular functions. While a potassium release is matched by an uptake occurring subsequent

Fur. J . Biocliein. 57 (1975)

51 1

to the presence of the hydroperoxide, n o release of lactate dehydrogenase as a cytosol inarker enzymc was obscrved [18].

A problem arises by the comparison of the efiects of tort-butyl and cuniene hydroperoxides in untreated and phenobarbital-pretreated rat livers. While both are removed linearly with increasing concentration (Fig. 2) in both types of liver, the rates of GSSG release from the liver are lower in phenobarbital-pretreated ones. and a steady state is no1 obtained with cuinene hydroperoxide (Fig. 6, 7). This and the further results obt,dined with the isolated inicrosomes suggest that a considerable part may be played by the endo- plasinic reticulum in hydropcroxide removal in the intact cell.

While Hrycay and O'Brien [15] reported rates at 25 C of about 80 nmol x iniii-' x ing rnicrosonial protein-' for NADH peroxidase activity, they already mentioned that twice this rate was seen for cuinene hydroperoxide removal. Thus, the fate of half of the hydroperoxide remained obscure. It is now suggested that peroxidation of microsomal lipids accounts for a large part of cumene hydroperoxide removal by isolatcd inicrosonies. I n the experiments of Fig. 9, an NADH-regenerating system was absent, indicating that high rates of oxygen uptake can be initiated by cuinene hydroperoxide in the absence of added reduc- ing equivalents. This is not altogether surprising, since organic hydroperoxides are used in the chemistry of polymerization as initiators where free radicals are required [33]. The higher reactivity of cuniene i~crsus rerr-butyl hydroperoxide inay make the latter more suitable for metabolic studies where uncontrolled oxygen-dependent radical chain reactions are un- wanted side events.

The differences between the untreated and pheno- barbital-treated liver could be due to a difference in the pattern of reactive iron species, r .g . cytochrome P-450 participating in an ROOH-dependent per- oxidasc activity, or to a difference in the pattern of polyunsaturated lipids. Davison and Wills [34] re- ported a significant increase of linoleic acid in micro- soinal phosphatidylcholine and phosphatidyl ethanol- aininc from phenobarbital-pretreated rats. It will therefore be interesting to coinpare other types of initiation of lipid peroxidation as used by Hogberg et a/. [35] with hepatocytes from phenobarbital- treated rats.

GSSG Releuse

The linear dependence of GSSG release froin the liver upon the rate of hydroperoxide infusion makes this parameter suitable for assessing organic hydro-

peroxide-linked reactions. It has been documented by Menzel 1361 that GSH is not released froin the liver upon hydroperoxide addition.

As stated above, the mechanism and properties of glutathione transport are unknown. GSSG release does not depend on the presence of hydroperoxides. Glutathione can be oxidized within cells also by other ineans 137,381. In fact, working with erythrocytes and eye lens, Srivastava and Beutler [39,40] originally described GSSG release using CSH-oxidizing agents and postulated an active ATP-dependent GSSG transport system. A better understanding of the factors governing GSSG release will provide a tool for further evaluation of the role of GSSG i n the control of the pentose phosphate cycle by counter- acting the NADPH inhibition of glucose-6-phos- phate dehydrogeiiase, as proposed by Eggleston and Krebs [41]. Thus, a sudden rise of GSSG concentra- tion effected by glutathione peroxidase may initiate increased rates of NADPH formation by two routes, one isiu a direct effect on glucose-6-phosphate dehydro- genase (mediated by a 'cofactor' [41]), and the other via GSSG reductase.

A relationship between hydroperoxide inetabolisin and pentose phosphate pathway activity has been considered by many investigators, r.g. Jacob and Jandl [42], and the effects of glucose added to hepato- cytes from fasted rats (Table 2) as well as the increase of 14C0, release from [l -'4C]glucose upon tert-butyl hydroperoxide addition are in support of this. How- ever, such observations make an explanation of the release of GSSG at the low rates of hydroperoxide infusion (Fig. 4) somewhat difficult. If indeed an active reducing system were operative, one would expect that a rise of GSSG concentration would be most effectively counteracted with the lower rates. Since this is not the case, the possibility exists that GSSG reductase does not operate effectively i~? siru (the Michaelis constants for the enzyme were reported to be 3 pM for NADPH and 50 pM for GSSG [43]), and, in view of the large oxidation in the NADPH system (Table 2, [17,18]), that fine control of NADPH regeneration is insufficiently operative.

Finally, the data on GSSG release upon H,O, infusion (Fig. 4) further substantiate the conclusion that. in addition to catalase, glutathione peroxidase can play a role in hepatocyte H,O, metabolism [44,17]. For an understanding of the mutual relationships between these two pathways, more knowledge on the cytotopical aspects of H,O, formation and decomposi- tion is required, as results also froin the recent work by Oshino c't 01. [45] on microsomal and peroxisomal H,O, production. One obvious function of catalase would be. in the present context, to alleviate loss of glutathione froin the cell.

Eur. .I. Biocheiii. 57 (197.5)

512 H. Sies and K.-H. Summer: Hydroperoxide-Metabolizing Systems in Rat Liver

Expert technical assistance was provided by Miss I . Linke and Mrs. A. Marklstorfer. Fruitful discussions with Drs L. Flohe, H. Menzel and A. Wendel, Tubingen, and with N. Oshino and B. Chance, Philadelphia, during various stages of this work are gratcfully acknowledged. This study was supported by Deutsche Forscliun~.sgemeinsckuft, Sonder-f~r..scliunR~~bereicli 51 ‘Medizinische Molekulurhiologie und Biochemie’, Grant D18.

REFERENCES

I . Sies, H. & Chance, B. (1970) FEBS Lett. I / , 172-176. 2. Oshino, N., Chance, B., Sies, H. & Bucher, Th. (1973) Arch.

3. Sies, H. (1974) Angen,. Chem. 86, 789-801. 3a. Sies, H. (1974) Angen. Chem. Int. Ed. Engl. 13, 706-718. 4. Hochstein, P. & Ernster, L. (1963) Biochem. Biophys. Rex

5. May, H. E., Poyer, J. I. & McCay, P. B. (1965) Biochem. Bio-

6. Tappel, A. L. (1973) Fed. Proc. 32, 1870-1874. 7. Hunter, F. E., Scott, A., Weinstein, J. & Schneider, A. (1964)

J . Biol. Chem. 239, 622-630. 8. Flohe, L. & Zimmermann, R. (1974) in Glutathione (Flohe. L.,

Benohr, H., Sies, H.. Waller, H. & Wendel, A,, eds) pp. 245 - 259, G. Thieme, Stuttgart.

Biochem. Biophys. 154, 117-131.

Conzmun. 12, 388 - 394.

p1zy.r. Rrs. Comnzun. 19, 166- 170.

9. Mills, G. C. (1957) J. Bid. Chem. 229, 189-197. 9a. Mills, G . C. (1960) Arch. Biocliem. Biophys. 86, 1-5.

10. Schneider, F. & Flohe, L. (1967) Hoppe Seyler’s Z. Phy.siol.

11. Little, C. & O’Brien, P. J. (1968) Biochem. Biophys. Res.

12. Christopherscn, B. 0. (1968) Biochim. Biophj’s. Actu, 176,

12a. Giinzler, W. A., Vergin, H., Muller, I. & Flohe, L. (1972)

13. Flohk, L. (1971) Klin. Wochenschr. 4Y, 669-683. 14. Hrycay, E. G. & O’Brien, P. J. (1971) Arch. Biochem. Biophys.

15. Hrycay, E. G. & O’Brien, P. J. (1974) Arch. Biochem. Biophys.

16. Kadlubar, F. F., Morton, K . C. & Ziegler, D. M. (1973) Bio-

17. Sies, H., Gerstenecker, C., Menzel, H. & Flohe, L. (1972)

18. Sies, H., Gerstenecker, C., Summer, K. H., Menzel, H. & Flohe. L. (1974) in Glutathione (Flohe, L., Benohr, H.. Sies,

Chem. 348, 540- 552.

Commun. 3f, 145- 150.

463 - 470.

2. Physiol. Chem. 353, 1001 - 1004.

147, 14-27.

160, 230-245.

chem. Bioph-vs. Res. Commun. 54, 1255- 1262.

FEBS Lett. 27, 171 - 375.

H., Waller, H. & Wendel. A,) pp. 261-276, G. Thieme, Stuttgart.

19. Sies, H. & Grosskopf, M. (1975) Eur. J . Biochem. 57, 513-520. 20. Haussingcr, D., Weiss, L. & Sies, H. (1975) Eur. J . Biochem.

21, Berry, M. N. & Friend, D. S. (1 969) J . Cell Biol. 43, 506 - 520. 22. Sies, H., Haussinger, D. & Grosskopf, M. (1974) Hoppe Seyler’s

23. Estabrook, R. W. & Holowinsky, A. (1961) J . Biochem. Bio-

24. Bergmeyer, H. U., ed. (1974) Methoden der En:ymatischen

25. Bernheim, F., Bernheim, M. L. C. & Wilbur, K . M. (1948)

26. Reference deleted. 27. Heim, W. G., Appleman, 1). & Pyfrom, H. T. (3956) Am. J .

Phj*siol. 186, 19 - 23. 28. Price. V. E.: Sterling, W. R., Tarantola, V. A,, Hartley, R. W.

& Rechcigl, M. (1962) J . Bid. Chem. 237, 3468-3475. 29. Flohe, L. & Schlegel, W. (1971) Z. Phyxiol. Chem. 352, 1401 -

1410. 30. Dahle, L. K., Hill, E. G . & Holman, R. T. (1962) Arch. Bio-

chem. Biophys. 98, 253 - 261. 31. Hochstein, P., Nordenbrand, K. & Ernster, L. (1964) Biocli~ni.

Biophys. Res. Commun. 14, 323-328. 32. Srivastava, S. K., Awasthi, Y. S. & Beutler, E. (1974) Bioc~hem.

J . 139,289-295. 33. Walling, C. & Heaton, L. (1965) J . Am. Chem. Soc. 87, 38-47. 34. Davison, S. C. & Wills, E. D. (1974) Biochem. J . 140,461 -468. 35. Hogberg, J., Orrenius, S. & Larson, R. E. (1975) Eur. J . Bio-

36. Menzel, H. (1973) Dissertation Fachbcreich Chemie, Uni-

37. Beutler, E. (1957) J . Lab. Clin. Med. 4Y, 84-94. 38. Kosower, N. S., Kosower, E. M., Wertheim, B. & Correa,

W. S. (1969) Biochem. Biopliys. Res. Commun. 37, 593 - 596. 39. Srivastava, S. K. & Beutler, E. (1969) .I. Biol. Cl7em. 244, 9- 16. 40. Srivastava, S. K. & Beutler. E. (1969) Biochem. J . fI2. 421-

41. Egglcston, L. V. & Krebs, H. A. (1974) Biochem. J . 138,

42. Jacob, H. S. & Jandl, J. H. (1966) J . B id . Chem. 241. 4243-

43. Mize, C. E. & Langdon, R. G . (1962) J . Biol. Chcm. 237,

44. Hochstein, P. & Utley, H. (1968) Mol. Pharmacol. 4, 574-579. 45. Oshino, N., Jamieson, D. & Chance, B. (1975) Biochem. J .

52,421 -431.

Z. Pliysiol. Chem. 355, 305 - 320.

p h . ~ ~ . Cytol. 9, 19 - 28.

Anulysr, 3rd edn, Verlag Chemie, Weinheim.

J . Bid. C h e m 174, 257 - 264.

chrni. 50, 595 - 602.

versity of Tubingen.

425.

425 -435.

4250.

1589-1595.

146. 53 - 65.

H. Sies and K.-H. Summer, Institut fur Physiologische Chemie und Physikalische Biochemie der Ludwig-Maximilians-Universitat Miinchen, D-8000 Miinchen 2, GoethestraBe 33, Federal Republic of Germany

Eur. J. Biochem. 57 (1975)