Reprints Available Directly from the Publisher Photocopying Permitted by License Only Clinical and Enzymatic Investigation of Induction of Oxygen Free Radicals by Rep Ischemla and Erfuslon an Human Hepatocellular Carcinoma and Adjacent Liver

Serum concentration ofthiobarbituric acid (TBA) reactants in the hepatic vein were measured before and after transient dearterialization of the liver in five human subjects bearing unresectable hepatocellular carcinoma (HCC). During hour ofthe occlusion ofthe hepatic artery, change in TBA reactants level was slight. However, the mean value of TBA reactants in hour after the reflow increased to 1.50 + 0.11 nmol/ml (mean + S.E.) and was significantly higher (p < 0.05) than those before hepatic dearterialization (1.28 + 0.11 nmol/ml) andjust before the release ofocclusion (1.32 + 0.09 nmol/ml). Further, two endogeneous scavenger enzymes, superoxide dismutase (SOD) and catalase (CAT), and one of the major sources of oxygen free radicals, xanthine oxidase (XOD) were measured in human untreated HCC and the corresponding adjacent liver tissue.


dence."Dr. Ak
ra Yamanoi, Second Department of Surgery, Shimane Medical University, Izumo 693, Japan.

This study was supported in part by a Grant in Aid for Fundamental Scientific Research from the Japanese Ministry of Education (No. 01010046).

an arterial supplyI.Hepatic dearterialization has been widely sed in an attempt to control hepatic tumors and thereby to increase survival 2-8.Transient blockade of the hepatic artery is one of the alternatives of this therapy, with which we are now attempting clinical trials in the treatment of unresectable hepatocellular carcinoma (HCC).Although favorable results have been demon- strated in some cases by us and another group6,9-, it is unclear that how transient blockade ofthe hepatic artery may cause the tumocidal effect.We speculate that oxygen derived free radicals may have some implications in this therapy because transient blockade of the hepatic artery Among the several defense systems against the ischemia-reperfusion injury, we are interested in two endogenous antioxidative enzymes, superoxide dismutase (SOD), that scavenges superoxide, and catalase (CAT), that scavenges hydrogen peroxide.Much ofthe toxicity of the superoxide radicals is believed to be by way of the hydroxyl radical formed in a transition metal iron- catalyzed reaction with hydrogen peroxide 1,3.The combined action ofendogenous OD and CAT therefore The mean value ofTBA reactants in hour after the reflow was 1.50 +_ 0.11 nmol/ml (mean _+ S.E.) and was significantly higher (p < 0.05) than those before the initiation of hour of hepatic dearterialization (1.28 +_ 0.11 nmol/ml) and just before the release of occlusion (1.32 _+ 0.09 nmol/ml).Each number corresponds to the case number.

appears to be ofimportance for a reduction ofthe oxygen free radical toxicity.On the other hand, the inhibitors of xanthine oxidase(XOD) that is one ofthe major sources of oxygen free radicals, can ameliorate this type of tissue injury 14-16.

This study was designed to investigate the hypothesis that oxygen free radicals induced by ischemia and reperfusion might have a tumocidal effect.Lipid peroxidation in tumor-bearing livers was assessed by the measurement ofthiobarbituric acid (TBA) reactants in the hepatic vein before and after 2 hours of transient dearterialization of the liver.The present paper also describes the difference in a pro-oxidant condition (XOD) and antioxidant systems (SOD and CAT) between HCC and adjacent liver tissue, and further discusses whether such differences have some implications in cellular injury caused by ischemia and reperfusion of tumor-bearing livers.

MATERIALS AND METHODS 1. Assessment of lipid peroxidation by transient chemotherapy dearterialization ofHCC-bearing liver.


Patien

From December 1988 t
June 1990, 9 consecutive patients with unresectable HCC(s) were subjected to transient hepatic dearterialization.Among them, 5 patients were randomly selected, well informed and gave consent to this study.Characteristics ofthe patients including the tumor response and the outcome are summarized in Table 1.


Transient blockade of the hepatic artery

The operative technique for implantation ofthe vascular occl

er was in accordance with the method reported
elsewhere ,.Briefly, the hepatic artery is freed for a distance of2 cm, and the vascular occluder, which consists of a silicone rubber cuff connected to a catheter with flexible strap to be sewn around the hepatic artery, was placed around it just distal or proximal to the gastroduodenal artery that was cannulated for arterial infusion.It was confirmed intraoperatively that injection of to 2 ml of saline into the balloon would occlude the hepatic artery and this defined the volume necessary for comp ete obstruction.Repeated hepatic dearterialization was started 7 days after operation and done for hour twice daily.

2. Measurement of SOD, CAT, and XOD in human HCC and adjacent liver tissue Patients Eleven patients with histologically proven diagnosis of HCC were included for measuring the activities of SOD and CAT, and 9 patients for XOD.Patient characteristics are presented in Table 2 and 3.All patients underwent surgical excision of HCC, and none had received either antiblastic or embolization therapy before surgery.

Bloodsamplingandmeasurement of TBA reactants A cannulation tube w

inserted thro
gh the right subclavian vein or the fight internaljugular vein into one of the hepatic veins which was judged as the predominant drainage vein of the tumor.Blood samplings were serially taken before and at the end of hour occlusion of the hepatic artery, and in 15 and 30 minute, and and 2 hour after the reflow of their first transient hepatic dearteria- lization.Sera were obtained by immediate centrifugation nd were stored at-30 C until the assay was available.

TBA reactants were measured according to the fluorimetric method ofYagi 19.Specimens HCC tissue and macroscopically cancer free liver tissue adjacent to HCC were obtained at surgery, and were immediately frozen by liquid nitrogen.Each sample was homogenized in 9 volumes of 100 mM sodium phosphate buffer (pH 7.4) per gram wet weight using a Teflon homogenizer.The homogenate was then sonicated over ice in 30 consecutive 0.5s bursts at 0.5s intervals, and centri

ged at 15,000 G for
0 minutes.The supernatant for XOD assay was desalted with Sephadex G-25 and served as an enzyme sample.


Enzyme Assays

Total SOD concentration (Copper-and Zinc-containing SOD + Manganese-containing SOD) was measured by a method of Nakano et al. (20) involving an inhibition of a crypridina luciferin analog-dependent luminescence induced by the hypoxanthine-xanthine oxidase system, and expressed as units/mg ofprotein.CAT activity was measured by a modified method of von Euler-Josephson and

xpressed
as units/mg of protein 21.

XOD activity was measured according to the method of Sasaoka et al. 22 using high-performance liquid chro- matography with Xuorescence detection.Xanthine dehydrogenase that can be rapidly converted to XOD in ischemic tissue was converted to XOD by adding 2,6- dichlophenolindopherol sodium to the reaction mixture as an electron acceptor.Data were expressed as pmoles!minute/mg ofprotein.

The protein content of tissue preparation was measured by the method of Lowry et al. 23 with bovine serum albumin as a standard.


StatisticalAnalysis

Statistical analysis was done by paired test, and p value less than 0.05 was considered significant.2 Correlations in activity of superoxide dismutase (SOD), catalase (CAT), and xanthine oxidase (XOD) between HCC and corresponding adjacent liver tissue in each patient are shown.A higher level of SOD in HCC than that in liver tissue was found in 9 cases out of (81.8%).A lower level of CAT in HCC than that in liver tissue was found in 8 cases out of 11 (72.7(1/4,).A lower level cfxanthine oxidase in HCC than that in liver tissue was found in all of 9 cases.


RESULTS

l. TBA reactants in the hepatic vein before and after transient hepatic dearterialization

Changes in TBA reactants ofeach patient and the mean value are concomitantly shown in Figure 1.The characte- ristic findings of change in TBA reactants were present to some degree in 4 (case 1-3, and 5) of 5 patients, and were as follows; During hour of occlusion of the hepatic artery, change in TBA reactants level was slight, but it rapidly elevated after the release of occlusion, and reached its peak value in 30 minutes or hour.In case 4, the value of TBA reactants hour after the release of occlusion however was not higher than that obtained just before the release ofocclusion.The mean value ofTBA reactants in hour after the reflow was 1.50 + 0.11 nmol/ml (mean + S.E.) and was significantly higher (p < 0.05) than those before the initiation of hour of hepatic dearterializati n (1.28 + 0.11 nmol/ml) and just before the release of occlusion (1.32 + 0.09 nmol/ml).

2. SOD, CAT, and XOD in human HCC and adjacent liver tissue

The results given in Table 4 show the mean value + S.E.

of SOD, CAT, and XOD amounts in both HCC and adjacent liver tissue.The mean amount of SOD in HCC was sign

icantly inc
eased compared with liver tissue, but adversely, the mean value ofCAT in HCC was significantly decreased compared with liver tissue.The mean activity of XOD in HCC was significantly lower than that in liver tissue.Figure 2 Shows each correlation ofSOD, CAT, and XOD level between HCC and corresponding liver tissue.

A higher level ofSOD in HCC than that in liver tissue was found in 9 cases out of 11 (81.8%).One ofthe two cases in which level of SOD in HCC was lower than in the liver was associated with macronodular liver cirrhosis and another was with chronic active hepatitis, and histopathology in both two cases showed trabecullar pattern and Edmondson 2 grade.A lower level of CAT in HCC than that in liver tissue was found in 8 cases out of 11 (72.7%).

There was however no significant difference in clinical background between these 8 and other 3 cases.A lower level ofxanthine oxidase in HCC than that in liver tissue was found in all of 9 cases.

The correlation between SOD and CAT level in HCC is shown in Figure 3.No strict correlation between SOD and CAT in both HCC and liver tissue (not shown) was found.The correlation between superoxide dismutase and catalase activities in HCC.No obvious correlation was found.


DISCUSSION

During the last decade, investigations ofthe role of oxygen derived free radicals in biological systems have seen a rapid expansion14-18.It has been postulated that the generation ofsuperoxide radicals initiates the biochemical cascade that produces cellular injury.An increase in oxygen free radicals and/or an impairment of scavenging systems oftoxic oxygen species may lead to severe damage of cell structures.Recent evidence indicates that such action may be involved not only in ischemia-reperfusion injury, but in chromosomal aberration, mutation, and carcinogenesis 27-32.Interruption of both the hepatic artery and the portal vein can induce oxygen free radicals, and these toxic intermediates may cause ischemic liver injury upon reperfusion.Puntis et al. 24 have demonstrated that the liver tissue was injured by oxygen free radicals in partial liver ischemia induced by only hepatic artery occlusion in rats.In the current study, serial chan e in TBA reactants was evaluated in the hepatic vein before and after hour of transient hepatic artery interruption in patients with unresectable HCC.Our results showed that the level of TBA reactants was rapidly elevated after release of the occlusion.TBA reactants are the metabolites of lipid peroxidation and are generally considered as indirect evidence of free radical production.It is therefore suggested that oxygen free radicals were generated at the time of reperfusion, and that reperfusion would be the important component in this therapy.This observation has provided the first clinical evidence of oxygen free radical activity related to ischemia-reperfusion injury induced by partial liver ischemia.It would however be impossible to distinguish by this study whether such injury had been induced in normal hepatic tissue, in tumors, or in both.To know this would be essential to clearly show the role of oxygen free radicals in the current therapy.Investigations of the precise relation between the tumor response and the degree ofTBA reactants elevation would be helpful to clarify this matter, although various patient characteristics and lack ofcompletion ofthe therapy made it difficult to evaluate by this study.

SOD is a protective enzyme that efficiently and specifically scavenges the superoxide radical by catalyzing its dismutation to hydrogen peroxide and oxygen.It is thus believed to play a key role in the enzymatic defense of he cell against oxygen free radicals.We demonstrated an increase in SOD content in 81.8% of human HCC compared with adjacent liver tissue.An increased activity of SOD might lead to a depressed level of reactive oxygen metabolites, and thus might lead ultimately to decreased lipid peroxidation.It has been suggested that one of the characteristic biological changes occurring in malignant tissues is an impairment of antioxidant systems and increased lipid peroxidation 30-32.The results however may indicate that the system of defense against oxygen free radicals is not impaired in human HCC compared with corresponding adjacent liver.This observation disagrees with the result by Corrocher et al. 30, in which a remark- able reduction of CAT and glutathione in human HCC was shown.Our result also showed lowered CAT activities in 72.2% ofHCC.However, it would be premature to draw a general conclusion about the antioxidant system in human HCC, because the content of SOD is different in Corrocher's and our own studies.Although the reason for the increase in SOD content of HCC is unclear in the current study, it is not surprising that HCC gains several, particular lines of defense systems against the oxidative attack.

Interest in XOD as a source of oxidizing agents has increased since it has been implicated in the pathogenesis ofischemia-reperfusion injury.There is growing evidence that oxygen free ra icals generated by XOD are primarily responsible for the cellular injury associated with reoxy- genation ofhypoxic tissues 14-17.It is therefore important to know the content of XOD in HCC in order to know if HCC can produce oxygen free radicals upon ischemia and reperfusion.The current study shows that the activity of XOD in HCC is significantly reduced compared with the adjacent liver tissue.Existence ofXOD in HCC, even ifthe content is low, would however be enough to suggest that oxygen free radicals may be generated not only in the liver tissue but in HCC upon ischemia and reperfusion.It should be mentioned that the difference in the severity of tissue injury caused by oxygen free radicals between HCC and the liver tissue remains unclear, because the value of lipid peroxidation should be determined by the combi- nation with antioxidant systems that counteract the damaging action.Cheeseman et al. 32 have shown that malondialdehyde production stimulated by NADPH + ADP + iron in suspensions of isolated Yoshida hepatoma cell was reduced compared with normal rat hepatocytes.Such direct measurement of lipid peroxidation may be mandatory to reveal whether antioxidant system in human HCC is impaired or not.

We demonstrated the difference in the content of two scavenger enzymes, SOD and CAT, and the key enzyme to produce oxygen free radical upon ischemia and reperfusion, XOD, between human HCC and adjacent liver tissue.Our data support the hypothesis that oxygen free radicals can be generated in HCC when transient hepatic artery blockade is applied.However, it remains unclear whether such oxidative stress is effectively tumorcidal or not, because the content ofXOD is low and SOD is high in HCC.It can be however assumed that under arterial occ