Summary Unsaturated fatty acids, including n-3 polyunsaturated fatty acids (PUFAs) such as docosahexaenoic acid (C225, DHA) and eicosapentaenoic acid (C20.5, EPA), and a series of n-6 PUFAs were investigated for their anti-tumour and antimetastatic effects in a subcutaneous (s.c.) implanted highly metastatic colon carcinoma 26 (Co 26Lu) model. EPA and DHA exerted significant inhibitory effects on tumour growth at the implantation site and significantly decreased the numbers of lung metastatic nodules. Oleic acid also significantly inhibited lung metastatic nodules. Treatment with arachidonic acid showed a tendency for reduction in colonization. However, treatment with high doses of fatty acids, especially linoleic acid, increased the numbers of lung metastatic nodules. DHA and EPA only inhibited lung colonizations when administered together with the tumour cells, suggesting that their incorporation is necessary for an influence to be exerted. Chromatography confirmed that contents of fatty acids in both tumour tissues and plasma were indeed affected by the treatments. Tumour cells pretreated with fatty acids in vivo, in particular DHA, also showed a low potential for lung colony formation when transferred to new hosts. Thus, DHA treatment exerted marked antimetastatic activity associated with pronounced change in the fatty acid component of tumour cells. The results indicate that uptake of DHA into tumour cells results in altered tumour cell membrane characteristics and a decreased ability to metastasize.Keywords: docosahexaenoic acid; unsaturated fatty acid; metastasis; colon carcinoma 26As tumour metastasis exerts an adverse influence on the prognosis of patients and is a major cause of cancer death, a considerable number of investigations into its biological, molecular and genetic features have been conducted (Raz and Ben-Ze'ev, 1987; Bertomeu et al, 1993). These investigations have indicated that tumour metastases are established by an inter-related sequence of processes, depending on various factors derived from both the tumour and the host. There is substantial evidence that the membrane properties of tumour cells play a major role in the interactions between themselves and the surrounding environment (Awad and Spector, 1976;Schroeder, 1984;Taraboletti et al, 1989;Dahiya et al, 1992). Studies on mice have revealed that plasma membranes of tumour sublines with high metastatic ability exhibit a more fluid state than those with low metastatic ability, based on differences in lipid composition and lipid-protein ratios (Kier and Franklin, 1991). Furthermore, studies have demonstrated that the chemical and physical properties of cell membranes are modified by both the amount and type of fat in the diet, and this influences growth and/or alters the metastatic ability of tumour cells (Chen et al, 1992;Rose and Hatala, 1994). Diets rich in linoleic acid have been found to enhance the growth and metastasis of transplantable mammary carcinomas in rodents (Rao and Abraham, 1976;. In a study of dietary n-3 PUFAs in a ...

Inhibitory effects of naturally occurring antioxidants on the initiation stage of hepatocarcinogenesis were studied. Group 1 rats were given a diet containing β‐carotene (β8‐CT, 0.02%), α‐tocopherol (α‐TP, 1.5%), glutathione (GLT, 5%), vanillin (VNL, 1%), quercetin (QCT, 1%) or ellagic acid (ELA, 1%), or 3 doses of diallyl sulfide (DAS, 200 mg/kg, i.g) over an 8‐day period. On day 7, the animals received a single dose of 2‐amino‐3‐methylimidazo[4,5‐f]jumoline (IQ, 100 mg/kg, i.g.), 12 h after two‐thirds partial hepatectomy for initiation and 2 weeks thereafter, were placed on promotion regimen comprising phenobarbital (0.05% in diet) and a single dose of D‐galactosamine (100 mg/kg, i.p.). Groups 2 and 3 were treated as described for Group 1, but without test material or IQ, respectively. Survivors were killed at week 11 and antioxidant influence was assessed by comparing values for preneoplastic glutathione S‐transferase placental form‐positive (GST‐P+) foci between Groups 1 and 2. All lesions larger than 70 βm in diameter consisting of approximately 5 cells in cross section were counted. Numbers of GST‐P+ foci/cm2 in Group 1 were: β ‐CT, 7.99; α‐TP, 8.21; GLT, 9.71; DAS, 10.37; VNL, 10.57; QCT, 11.1; ELA, 12.5 (n = 11‐15). All, except ELA, showed a significant decrease as compared with the Group 2 value of 14.54 (n=15). Only β ‐CT showed a significant decrease for the area value. This is the first report to show that β ‐CT, α‐TP, GLT, DAS, VNL, QCT exert inhibitory effects on initiation of hepatocarcinogenesis by the food carcinogen IQ, suggesting that these antioxidants might find application as chemopreventive agents. Furthermore, the current protocol proved practical for the assessment of chemopreventive agents within 11 weeks, a relatively Short period.

The presence of ras family and p53 gene mutations in rat forestomach, intestine and liver tumors and soft tissue sarcomas induced by Nmethyl‐N‐nitrosourea (MNU) was examined using polymerase chain reaction‐single strand conformation polymorphism (PCR‐SSCP) followed by direct sequencing analysis. In the forestomach squamous cell carcinomas (SCC), Ha‐ros and p53 mutations were detected in 2 (40%) and 4 (80%) of 5 cases, respectively. The figures for Ki‐ras and p53 gene mutations in adenocarcinomas of the large and small intestines were 3 (18.8%) and 5 (31.3%) of 16 cases. Soft tissue sarcomas in different sites were found to have mutations of Ki‐ras in 7 (23.3%)and of p53 in 9 (30%) of 30 cases. One forestomach SCC and 2 soft tissue sarcomas had double p53 mutations in different exons. Single cases of forestomach SCC and intestinal adenocarcinoma had mutations in both Ki‐ras and p53 genes. No mutations were found in counterpart benign tumors or hepatocellular adenomas. The p53 mutation spectrum revealed preferential clustering within exon 8 for the forestomach SCCs, and exons 5 and 8 for the intestinal adenocarcinomas, whereas the distribution was evenly spread through exons 5 to 8 in soft tissue sarcomas. All the detected ras or p53 mutations were G:C to A:T transitions. These results indicate firstly that specific Ki‐ras, Ha‐ras and p53 gene mutations in MNU‐induced lesions are related to particular alkylation sites (G:C to A:T transitions) and secondly, although not essential, Ki‐ras, Ha‐ras or p53 gene mutations may be involved in the progression stage of forestomach, intestine and soft tissue neoplasms induced by MNU.

The expression of connexin 32 (Cx32), a major liver gap junction protein, after partial hepatectomy (PH) and during development and progression of hepatocarcinogenesis was studied in the rat. Cx32 was quantitatively analyzed by counting immunohistochemically demonstrated protein spots on the membranes of hepatocytes. Livers were sequentially examined after PH to assess the correlation with cell proliferation. For the analysis of different stages in carcinogenesis, Cx32 was assayed in N-ethyl-N-hydroxyethylnitrosamine-induced enzyme altered foci (EAF), hyperplastic nodules (HN), hepatocellular carcinomas (HCC), pulmonary metastatic HCC and transplanted HCC in relation to their degree of altered enzyme expression. Cx32 showed: (i) a rapid decrease after PH to its lowest levels during and 12 h after the S phase of cell proliferation when 5-bromo-2'-deoxyuridine (BrdU) labeling indices were examined; (ii) a progressive decrease from early preneoplasia EAF to HN and HCC, values for pulmonary metastatic and transplanted HCC being 0; (iii) clearly inverse correlations with increased BrdU index and degree of altered enzyme expression in HN, indicating that these, with the lowest Cx32 count, are closest to HCC. Therefore, the observed decrease appears linked to cell proliferation and progression of hepatocarcinogenesis, providing a reflection of cellular independence and growth advantage.

Organotropic chemopreventive effects of n-3 unsaturated fatty acids were studied using a multiorgan carcinogenesis model in male rats. Rats were treated with diethylnitrosamine (DEN), Nmethyl-N-nitrosourea (MNU), N-butyl-N-4-hydroxybutylnitrosamine (BBN), 1,2-dimethylhydrazine (DMH) and dihydroxy-di-n-propylnitrosamine (DHPN) during the first 7 weeks, and then given unsaturated fatty acid (UFAs), docosahexaenoic acid (n-3, C 22:6 ) (DHA), eicosapentaenoic acid (n-3, C 20:5 ) (EPA), linoleic acid (n-6, C 18:2 ) (LA) or oleic acid (n-9, C 18:1 ) (OA) at a dose of 1.0 ml/rat, 3 times a week by gavage for the consecutive 30 weeks. All rats were fed a low LA basal diet throughout the experiment and a calorie-restricted basal diet during the period of UFAs feeding administration. DHA significantly reduced tumor size and numbers in the large intestine as compared to OA treatment. Furthermore, DHA showed a tendency to inhibit carcinogenesis in the small intestine and lung. EPA also showed a tendency to inhibit intestinal carcinogenesis. On the other hand, LA showed a tendency to inhibit lung carcinogenesis, but to promote large intestinal carcinogenesis. However these UFAs did not influence preneoplastic and neoplastic lesion development in the liver, kidney, and urinary bladder. Levels of the administered fatty acids were clearly increased in the serum and organs. In contrast, arachidonic acid (AA) levels in the large and small intestines and liver were markedly decreased by treatment with DHA and EPA. Decreased levels of AA in the large intestine correlated well with tumor incidence, although the number of glutathione S-transferase-positive (GST-P + ) foci showed an inverse correlation with AA levels. The data thus provide evidence that an organotropism exists with regard to the influence of UFAs on carcinogenesis, which correlates with reduction of tissue AA levels in the target organs.

To assess mechanisms of chemoprevention of hepatocarcinogenesis by trans‐β‐carotene (β‐C), DL‐α‐tocopherol (α‐T), and freeze‐dried whole leaves of Kidachi aloe (Aloe), formation of 2‐amino‐3‐methylimidazo[4,5‐f]quinoline (IQ)‐DNA adducts was measured by 32P‐post‐labeling analysis, and CYP1A1 and CYP1A2 protein levels were analyzed by ELISA. Group 1 rats were fed diet containing 0.02%β‐C, 1.5%α‐T or 30% Aloe over an 8‐day period, while group 2 was given basal diet alone. On day 7, all animals were subjected to two‐thirds partial hepatectomy (PH). Twelve hours after PH, they received a single dose of the carcinogenic food pyrolysate IQ (100 mg/kg) intragastrically, to initiate hepatocarcinogenesis. Rats were killed 6, 12, 24 and 48 h after IQ administration. The levels of adducts, expressed as relative adduct labeling values in rats treated with β‐C, α‐T and Aloe, were decreased as compared with the control group at hour 24 (36 h after PH), with a significant difference in the case of the β‐C group (46.4% of the control value). Similarly, all showed a tendency for decrease at hour 48. Furthermore, the levels of CYP1A2, known to be responsible for activation of IQ, showed a significant reduction at hour 24. It is concluded that β‐C, and possibly also α‐T and Aloe, have the potential to reduce IQ‐DNA adduct formation, presumably as a result of decreased formation of active metabolites. The results may explain, at least in part, the previously observed inhibitory effects of these compounds on induction of preneoplastic hepatocellular lesions.

Two different types of focal preneoplastic lesions, tentatively named Type I and II lesions, were recognized in the liver of rats chronically treated with clofibrate for 104 weeks. Type I lesions were characterized by mostly negative glucose-6-phosphate dehydrogenase (G6PD) activity (6 out of 10, 60%) and positive expression of succinate dehydrogenase (10 out of 10, 100%), in addition to the previously documented complete lack of expression of glutathione S-transferase, placental form (GST-P) and gamma-glutamyl transpeptidase (GGT). Furthermore, most importantly, Type I lesions exhibited a clear decrease in immunohistochemically demonstrated connexin32 (Cx32) spot counts on their hepatocyte membranes, similarly to nitrosamine-induced lesions. In contrast, Type II lesions, mostly small in size and positively expressing GST-P and/or GGT and G6PD, similarly to their previously reported nitrosamine-induced counterparts, did not exhibit a significant decrease in Cx32 count. In addition, spontaneously occurring lesions, again sharing the same enzyme phenotype, did not show a decrease in Cx32. The results indicate that: (i) a clear distinction between the two lesions, with Type I being involved in clofibrate-induced tumors and Type II being more likely to be spontaneous in nature; (ii) a decrease in Cx32 is closely linked to lesion development and possibly stage of progression, irrespective of the enzyme phenotype and the applied carcinogen; (iii) the unaltered condition of Cx32 may suggest a slow growing or non-progressive nature.

The histogenesis of 3 types of rat renal cell tumors (basophilic cell, clear cell, and oncocytic) was stereologically analyzed, with particular attention paid to transitions from normal tubules. Early nitrosamine-induced preneoplastic lesions, including dysplastic tubules (altered tubules), epithelial hyperplasias, and small adenomas, were reconstructed using serially sectioned specimens processed for carbonic anhydrase type II (CA) and periodic acid-Schiff (PAS) (CA-PAS) double staining to allow easier distinction of the nephron segments: Proximal tubules had a PAS-positive brush border and were weakly positive for CA in the cytoplasm; distal tubules were PAS negative and weakly positive for CA; collecting ducts were PAS negative and strongly positive for CA. Similarly, cytochrome c oxidase (CytOx) and CytOx-PAS double staining was also applied to confirm the character of oncocytic lesions. All basophilic lesions (7 of 7) showed transition to proximal tubules. Clear cell lesions positive for CA, on the other hand, showed transition to distal tubules in 4 of 9 (44.4%) lesions and to collecting ducts in 4 of 9 (44.4%) lesions, but in only 1 of 9 (11%) to a proximal tubule. All oncocytic lesions (1,8,9,23,30,32) similar to those observed for lesions in the liver (6,11,12,16,25,32) and other organs (28,31). The histogenesis of experimentally induced renal tumors has been studied on the basis of characteristic features for enzyme-altered tubules considered to be early precursor lesions for renal cell tumors. For example, basophilic tubules with high glucose-6-phosphate dehydrogenase (G6PD) activity and decreased succinate dehydrogenase (SD) appear to arise from proximal tubules, whereas the oncocytic type exhibiting the opposite phenotype of low G6PD and high SD has been suggested to have a histogenesis independent from other segments, possibly collecting ducts (2,9,22,36 (27,29,34,35 cells (26,37). Similarly, cytochrome c oxidase (CytOx) immunostaining was used for the differential recognition of oncocytic lesions (17,19), occasionally in combination with PAS staining (CytOx-PAS). MATERIALS AND METHODSAnimals and Treatments. Groups of male 6-wk-old Fischer 344 (F344) and Sprague-Dawley rats (Charles River Japan, Inc., Atsugi, Japan) were administered 0.1 % Nethyl-N-hydroxyethylnitrosamine (EHEN) (Sakai Chemical Laboratories, Fukui, Japan) for 3 wk (5 rats) (18, 30, 32) and 0.07% N-nitrosomorpholine (NNM) (Nakalai Tesque, Inc., Kyoto, Japan), respectively, mixed in their drinking water for 10 wk (5 rats) (9) and were then returned to tap water. They were then placed on basal diet (Oriental MF, Oriental Yeast Co., Ltd., Tokyo, Japan) throughout the experimental period. Animals were sacrificed 30 wk after the commencement of carcinogen application. Resected kidneys were cut into 3.5-to 4-mmthick slices along the long axis with a razor blade, fixed in ice-cold acetone for 4 wk, and then routinely processed for paraffin embedding.Histological Preparations and Stereological Reconstruction. Paraffin-embedded kidneys, ...