Green tea is the most widely consumed beverage. It has attained high reputation as a health-promoting dietary component ascribed to the antioxidant activity of (-)-epigallocatechin-3-gallate (EGCG), its main polyphenolic constituent. Evidence is increasing that tea constituents can be cell damaging and pro-oxidant themselves. These effects were suggested to be due to spontaneous H2O2 generation by polyphenols in solution. In the present study, we investigated the oxidant and antioxidant properties of green tea extracts (GTE) and of EGCG by means of the rodent macrophage-like RAW 264.7 and human promyelocytic leukemic HL60 cell lines. The results obtained show that both under cell-free conditions and in the presence of cells the oxidant activities of GTE and EGCG exceeded those of spontaneously generated H2O2 (FOX assay). Increase of intracellular oxidative stress was indicated by 2',7'-dichlorofluorescin probing, and the enhanced genotoxicity was demonstrated by the alkaline comet assay and by the micronucleus assay (cytokinesis block). Time- and dose-dependent induction of cell death was monitored by trypan blue exclusion, MTT assay, and Hoechst staining. Furthermore, in our systems in vitro, EGCG neither directly scavenges H2O2 nor mediates other antioxidant activities but rather increased H2O2-induced oxidative stress and DNA damage. In conclusion, our data suggest that detailed mechanistic studies on the effects of GTE and EGCG should be performed in vivo before excessive intake and/or topical application of green tea products can be recommended to healthy and/or diseased persons.

A great number of drugs, toxicants, and growth factors induce the generation of intermediary reactive oxygen species (ROS). The human promyelocytic leukemia HL60 cell line differentiated along the macrophage or neutrophil lineage is a model system that is frequently used for the generation of ROS by various agents. As a primary source of ROS the superoxide anion produced by an enzymatic complex, NADPH oxidase, is well established. The present study shows that nondifferentiated HL60 cells contain NADPH oxidase and can be used as a model for the assessment of oxidant as well as antioxidant compounds. The expression of the multicomponent NADPH oxidase was demonstrated in nondifferentiated HL60 cells at the molecular level by detection of the mRNAs of the components gp91phox, p47phox, and p67phox as well as functionally by phorbol 12-myristate-13-acetate (PMA)-stimulated generation of superoxide, which was susceptible to inhibition by diphenyleneiodonium. The functional assay was performed using the cells in a log growth phase by adapting a standard microplate assay based on the classic superoxide dismutase-inhibitable reduction of cytochrome c. Validation of the microplate assay was carried out both with nonadherent differentiated HL60 cells and the adherent mouse monocyte-macrophage-like RAW 264.7 cell line, as well as with various compounds of oxidant (bleomycin sulfate, cis-diammineplatinum(II), camptothecin, TNF-alpha, IL-1 beta), nonoxidant (4 alpha-PMA, piracetam), and antioxidant (alpha-tocopherol, ascorbic acid) activity. In summary, we established a highly specific, reproducible and--with the aid of the nondifferentiated HL60 cell line--time-saving superoxide microplate assay as a valuable tool for the rapid screening of compounds for oxidative and antioxidative activity.

Hepatocellular carcinoma almost always arises in chronically inflamed livers. We developed a culture model to study the role of non-parenchymal cells (NPCs) for inflammation-driven hepatocarcinogenesis. Rats were treated with the carcinogen N-nitrosomorpholine, which induced initiated hepatocytes expressing the marker placental glutathione-S-transferase (GSTp). After 21 days two preparations of hepatocytes were made: (i) conventional ones (Hep-conv) containing NPCs and (ii) hepatocytes purified of NPCs (Hep-pur). Initiated hepatocytes, being positive for GSTp (GSTp-pos) were present in both preparations and were cultured along with normal hepatocytes, being negative for GSTp (GSTp-neg). Under any culture condition DNA synthesis was approximately 4-fold higher in GSTp-pos than in GSTp-neg hepatocytes demonstrating the inherent growth advantage of the first stages of hepatocarcinogenesis. Hepatocytes showed approximately 3-fold lower rates of DNA synthesis in Hep-pur than in Hep-conv, which was elevated above Hep-conv levels by addition of NPC or NPC-supernatant. Pretreatment of NPCs with proinflammatory lipopolysaccharide (LPS) further increased DNA synthesis. Thus, NPCs release soluble growth stimulators. Next we investigated the effect of specific cytokines produced by NPCs. Tumour necrosis factor alpha and interleukin 6 barely altered DNA synthesis, whereas hepatocyte growth factor (HGF), keratinocyte growth factor (KGF) and the heparin-binding epidermal growth factor-like growth factor (HB-EGF) were potent inducers of DNA replication in both, GSTp-neg and GSTp-pos cells. In conclusion, DNA synthesis of hepatocytes is increased by factors released from NPCs, an effect augmented by LPS-stimulation. NPC-derived cytokines, such as KGF, HGF and HB-EGF, stimulate DNA synthesis preferentially in initiated hepatocytes, presumably resulting in tumour promotion. Similar mechanisms may contribute to carcinogenesis in human inflammatory liver diseases.

We hypothesized that superoxide from Kupffer cells (KC) contributes to hepatocarcinogenesis. p47phox(-/-) mice, deficient in phagocyte NADPH oxidase and superoxide generation, received a single dose of the hepatocarcinogen diethylnitrosamine (DEN). The following hepatic effects were observed at time points between 30 min and 35 days. Liver damage after DEN was manifested by loss of body and liver mass and of liver DNA and by an increase in apoptosis, necrosis and signs of inflammation. These effects were massive in wild-type (wt) male mice, but only very mild in p47phox(-/-) mice. Regenerative DNA synthesis subsequent to liver damage was high in wt male mice, but weak in p47phox(-/-) mice. In females the apparent protection by p47phox(-/-) was less pronounced than in males. Therefore, further experiments were performed with males. In KC isolated from wt mice superoxide production was enhanced by DEN pretreatment in vivo. Also, in vitro addition of DEN to KC cultures induced superoxide release, similarly to lipopolysaccharide, a standard KC activator. Thus, DEN directly activates wt KC to produce superoxide. KC from p47phox(-/-) mice did not release superoxide. TNFalpha production by isolated KC was transiently depressed 0.5 h after DEN treatment in vivo, but recovered rapidly. In blood serum TNFalpha levels of wt mice were elevated for the initial 6 h. TNFalpha in KC cultures and in serum of p47phox(-/-) mice was reduced. DEN in vivo induced DNA damage ('comets') in hepatocytes. This damage was extensive in wt mice but much less in p47phox(-/-) mice. These studies suggest two conclusions: (i) superoxide generation by phagocytes during liver damage and inflammation aggravates genotoxic and cytotoxic effects in hepatocytes and may thus contribute to tumor initiation and promotion; (ii) DEN has a direct stimulatory effect on KC to release superoxide and TNFalpha.

There have been numerous reports that chemicals which induce peroxisomes in rodent liver increase DNA synthesis in isolated hepatic parenchymal cells, but not as well in vitro as in vivo. It is also known that tumour necrosis factor α (TNFα) is mitogenic in isolated hepatocytes. Since Kupffer cells are a major source of TNFα in the liver and have recently been shown to be activated by peroxisome proliferators, the possibility exists that the effect of peroxisome proliferators on DNA synthesis in parenchymal cells is via Kupffer cell contamination of isolated hepatocyte preparations. The purpose of this study was to evaluate this hypothesis by studying the effect of model peroxisome proliferators on purified hepatocyte preparations. Hepatocytes were prepared from rat liver by standard calciumfree and collagenase perfusion. Subsequently, cells were centrifuged through Percoll to remove contaminating nonparenchymal cells. Cells were at least 99.9% pure as assessed by cell counting using specific markers for hepatocytes (resorufin O-glucoside) and Kupffer cells (FITClabelled latex beads). Hepatocytes were cultured in Williams medium ϩ 10% fetal bovine serum for 24 h followed by culture for 48 h in Williams medium plus or minus drug or mitogen additions. Under these conditions epidermal growth factor stimulated DNA synthesis assessed by incorporation of [ 3 H]thymidine~5-fold over control levels. The peroxisome proliferators WY,14-643 and nafenopin, however, had no effect on DNA synthesis, although they did increase acyl-CoA oxidase as expected. In contrast, TNFα increased cell proliferation nearly 10-fold in purified hepatocytes, an effect nearly doubled by WY-14,643. Further, when conditioned medium from purified Kupffer cells incubated with WY-14,643 was added to pure hepatocytes, DNA synthesis was increased over 2-fold in a time-dependent manner. Collectively, these data support the hypothesis that peroxisome proliferators do not influence DNA synthesis in isolated hepatocytes per se. Rather, they stimulate cytokine production by Kupffer cells which in turn increases DNA synthesis in parenchymal cells. An increase in mitogenic cytokine production by Kupffer cells is necessary for stimulation of DNA synthesis in purified rat parenchymal cells.