Effect of beta-carotene supplementation on indices of colonic cell proliferation.

BACKGROUND: Epidemiologic studies have shown that consuming foods containing beta-carotene is associated with a decreased incidence of colon cancer. The validity of this association has recently been questioned. It is not known if the rate of colonic cell proliferation differs among individuals with or without a history of colonic polyps or cancer and if proliferation changes in response to beta-carotene. PURPOSE: This study was intended to (a) determine whether differences exist in colonic cell proliferation in individuals with and without prior colonic polyps or tumors, (b) demonstrate that beta-carotene accumulates in colonic mucosa following dietary supplementation, and (c) determine whether mucosal beta-carotene accumulation influences colonic cell proliferation. METHODS: Subjects were enrolled in the phase I study from June 1991 until February 1994. The participants included 20 individuals (11 males and nine females, aged 62.3 +/- 8.9 years [means +/- SD]) with normal colons (as judged by recent colonoscopy), 40 (24 males and 16 females, aged 59.6 +/- 10.1 years) with a history of colonic polyp(s), and 41 (30 males and 11 females, aged 67.2 +/- 9.7 years) with prior colon cancer. The subjects in the last two groups consumed either 30 mg of beta-carotene or placebo each morning for 3 months. This dose of beta-carotene has no known toxic effects, but it can increase the serum level by approximately 10-fold. beta-carotene concentration in serum and colonic tissue was quantitated by high-pressure liquid chromatography in samples collected before and after supplementation with beta-carotene or placebo. Cellular proliferation was assessed on the basis of tissue ornithine decarboxylase activity, urinary polyamine excretion, and proliferating cell nuclear antigen expression. The differences in colonic cell proliferation parameters due to beta-carotene supplementation, within and among different groups, were evaluated by the Wilcoxon matched-pairs signed ranked test and the Mann-Whitney test, respectively. All statistical tests were two-sided. RESULTS: Colonic cell proliferation did not differ in samples obtained from individuals with and without prior colonic polyp(s) or cancer. beta-carotene concentrations in serum and colonic tissue were significantly increased in groups receiving beta-carotene (P < .001). However, cell proliferation did not differ, as judged by any of the three measures, among samples from all experimental groups collected before and after supplementation with beta-carotene. CONCLUSIONS: Dietary supplementation with beta-carotene for a period of 3 months does not alter colonic cell proliferation in individuals with a history of colonic polyps or cancer. IMPLICATIONS: The mechanism by which beta-carotene might reduce colon cancer incidence does not appear to involve or result in a change in cell proliferation in the normal colonic mucosa as studied in individuals with a history of colonic polyps or cancer.

Aspirin toxicity for human colonic tumor cells results from necrosis and is accompanied by cell cycle arrest.

Chemoprevention of colorectal cancer using aspirin has been demonstrated in rodents and has been suggested by data from epidemiological studies. The mechanism that accounts for this prevention is unknown, but it is thought to relate to an irreversible inhibition of cyclooxygenase and, subsequently, prostaglandin production. The effect of aspirin on the growth of human colonic tumor cells was determined in an effort to gain insight into a possible mechanism of action. In the two cell lines studied, SW 620 and HT-29, we observed a significant dose- and time-dependent increase in aspirin toxicity in a concentration range of 1.25-10 mM. This result was independent of prostaglandin production, because there was no measurable prostaglandin E2 in cell culture medium. As compared with controls, cells in cultures that contained aspirin were not detached, which suggests that the mechanism of cell death was not apoptosis. Flow cytometric analysis revealed an increase in S phase and G2-M populations as well as the number of subdiploid nuclei in cultures treated with high-dose aspirin. Confirmation that cells were undergoing necrosis in response to aspirin was evident from the presence of cells that bound annexin V and accumulated propidium iodide in the absence of a population that bound annexin alone. The results suggest that aspirin induces cell cycle arrest and causes necrosis at high concentrations in vitro, but does not induce apoptosis. Collectively, these two events, necrosis and cell cycle arrest, may contribute to the chemopreventive effect that seems to result from long-term administration of aspirin in vivo.

Effect of aspirin on prostaglandin E2 and leukotriene B4 production in human colonic mucosa from cancer patients.

Results from epidemiological studies indicate that chronic administration of aspirin reduces the incidence of colon cancer. The mechanism that accounts for this reduction is not known, but it may be related to the decreased production of prostanoids that results from aspirin inhibition of cyclooxygenase. However, it is not known whether aspirin has a local effect on prostanoid production in the colonic mucosa and whether this effect is dose dependent. In this study, we determined the effect of oral administration of aspirin on the production of the prostanoid prostaglandin E2 (PGE2) in the intact human colonic mucosa. Inhibition of cyclooxygenase could result in an increased availability of arachidonic acid and a corresponding increase in production of other eicosanoids. To determine whether such an effect occurs, we also quantitated the concentration of leukotriene B4 (LTB4) in colonic mucosal samples. Mucosal samples were obtained during sigmoidoscopy from the colons of 17 subjects with a history of colonic cancer prior to and following 60 days of self-administration of 325 mg aspirin/day and again 60 days after administration of 650 mg aspirin/day. PGE2 and LTB4 concentrations were determined by enzyme immunoassay for tissue samples that were flash frozen after removal from the biopsy forceps and also in medium that was collected from tissue samples that were incubated for 4 h following removal from the subject. PGE2 concentrations were decreased significantly in samples collected after 60 days of consumption of 325 mg aspirin. An additional 60 days of consuming 650 mg aspirin/day did not result in a further significant decrease relative to that attained after consumption of 325 mg/day. Similar results were obtained using colonic explants, and the addition of aspirin to medium further reduced PGE2 production. LTB4 in tissue and medium was not significantly different in pre-versus post-aspirin samples, with the exception of an increased concentration in medium samples collected after consumption of 650 mg/day relative to pre-aspirin samples. The results indicate that aspirin affects eicosanoid production in the colonic mucosa of humans, but the effect is most likely restricted to products of the cyclooxygenase-dependent pathway. It appears that 325 mg of aspirin is sufficient to affect PGE2 production and that increasing the dosage to 650 mg daily provides an additional decrease in PGE2 synthesis. However, the higher dosage was associated with a considerable increase in complaints of gastric discomfort. Additional study is needed to establish whether doses less than 325 mg also provide a significant decrease in PGE2 production.

Flow cytofluorimetric analysis of drug accumulation by multidrug-resistant Trypanosoma brucei brucei and T. b. rhodesiense.

Dual laser flow cytofluorimetry has been used to compare accumulation of compounds representing three major classes of trypanocidal drugs by drug sensitive and drug resistant clones of Trypanosoma brucei brucei and T. b. rhodesiense. Clones selected for resistance to melarsoprol were shown to be cross-resistant in vivo to two diamidines, pentamidine and Berenil, but not to suramin. At 35 degrees C, bloodstream forms of these multidrug-resistant clones accumulated lower intracellular concentrations of the diamidines 4',6-diamidino-2-phenyl-indole (DAPI) and Hoechst 33342, the phenanthridine ethidium bromide, and the acridine acriflavine than drug sensitive parasites. Accumulation of all four drugs was saturable. Drug concentrations giving half-maximal rates of accumulation were increased in the resistant clones relative to the sensitive parent clones. The rate of DAPI accumulation by both resistant and sensitive parasites was strongly temperature dependent and increased sharply above 27 degrees C. Two distinct populations were resolved in mixtures of sensitive and resistant clones exposed to DAPI. Resistant and sensitive cells accumulated identical intracellular concentrations of DAPI following brief treatment with the detergent Triton X-100. The results suggest that alterations in the surface membrane of multidrug-resistant trypanosomes reduce accumulation of several drugs relative to drug sensitive parasites.

The major surface glycoprotein (GP63) is present in both life stages of Leishmania.

Leishmania exist as extracellular promastigotes which multiply in the gut of the sandfly insect vector and as intracellular amastigotes which divide in the phagolysosome of mononuclear phagocytic cells of the mammalian host. Promastigotes express a major surface glycoprotein of 63 kDa, referred to as GP63. The expression of GP63 in both Leishmania life stages was studied using rabbit antibodies against native GP63 as well as rabbit antibodies against recombinant GP63 that was synthesized in an Escherichia coli expression system. Immunofluorescence staining detected GP63 in intracellular amastigotes contained within a macrophage cell line and within freshly isolated lesion amastigotes. Western blot analysis using anti-recombinant GP63 antibodies also demonstrated that amastigotes synthesize GP63 which may undergo differential post-translational processing as compared to promastigote GP63.

P-glycoprotein expression by human colonic tumor cells influences the growth of tumor cells in vitro.

The stability of P-glycoprotein expression by high and low expressing cells obtained by flow sorting from two cell lines established from colonic tumors was assessed. Low expressing cells rapidly recovered expression to a level that was comparable to that determined for uncloned cells. High expressing cells maintained a high level of expression relative to the parental uncloned lines, although there was continuous reversion to a level of expression determined for the parental line with time. The population growth rates for high expressing cells was greater than that of parental and low expressing cells when assessed immediately following plating of sorted cells. The data suggest that high expression of P-glycoprotein confers a growth advantage which may increase repopulation and dissemination of tumor cells following drug therapy.

Cytotoxic effect of beta-carotene in vitro is dependent on serum concentration and source.

In a previous study we reported that beta-carotene solubilized in tetrahydrofuran (THF) is toxic for human colonic tumor cells in vitro using media containing 10% fetal calf serum (FCS). Cytotoxicity was evident using beta-carotene concentrations that are achieved in human serum as a result of supplementation with 30 mg beta-carotene/day. In an attempt to determine the mechanism for this toxicity we investigated the effect of beta-carotene when present in human serum as a result of dietary supplementation. This effect was compared to that observed for cells incubated in THF-solubilized beta-carotene. The results indicate that human serum from subjects with a high concentration of beta-carotene is not cytotoxic. Subsequent analysis revealed that in contrast to the results using FCS, THF-solubilized beta-carotene is not cytotoxic in the presence of human serum. In addition, the effect observed with FCS is blunted with increasing FCS concentration from > or = 90% cytotoxicity using 10% FCS to 36% using 100% FCS. The difference in results obtained using FCS and human serum may be due to a serum component that is relatively lacking in FCS as compared to human serum.

Detection of Porphyromonas gingivalis in oral plaque samples by use of the polymerase chain reaction.

Periodontitis is believed to be caused by bacteria which inhabit periodontal pockets. The identification of these periodontal pathogens by currently available methods requires considerable time and expertise. In this study, we have used a polymerase chain reaction (PCR) assay that is quick, relatively simple, and can detect low numbers of a putative periodontal pathogen. Primers specific for the fimbrial gene of Porphyromonas gingivalis were selected from the published sequence and used for amplification of a 131-basepair sequence of genomic DNA. The PCR detected as few as 100 P. gingivalis cells obtained from pure cultures. Three other bacteria (Actinobacillus actinomycetemcomitans, Prevotella intermedia, and Capnocytophaga gingivalis) that are also putative periodontal pathogens yielded no PCR product at any of the cell concentrations used. This assay was also used for detection of P. gingivalis in subgingival plaque. Five of 13 subgingival bacterial plaque samples obtained from four advanced adult periodontitis patients and two samples from a prepubescent child with advanced periodontitis contained P. gingivalis. The protocol developed is relatively simple and can be completed within four hours of the time of sample acquisition.

Suramin is synergistic with vinblastine in human colonic tumor cell lines: effect of cell density and timing of drug delivery.

Suramin is a multi-targeted antiproliferative drug developed for the treatment of African trypanosomiasis but with potential efficacy for the treatment of human cancer. Cell growth inhibition was determined in vitro for three human colonic tumor cell lines using three different doses of suramin (50, 100 and 200 microM). At the lower suramin concentration cell growth was stimulated relative to control cultures in all three cell lines. At the higher dose which is at the upper end of the tolerable dose in humans suramin reduced cell numbers by greater than 50%. Inhibition of cellular proliferation was reduced relative to increases in cell plating density. Addition of vinblastine six and to a lesser extent 72 hours post suramin (200 microM) resulted in an inhibition of cell growth and/or toxicity that exceeded that which occurred as a result of exposure to either suramin or vinblastine alone. To investigate the possible mechanism by which suramin sensitizes cells to vinblastine we determined the effect of suramin on expression of the multidrug resistance (mdr1) gene. A decrease in mdr1 mRNA was evident in one colon tumor cell line and a slight decrease detected in a second line. The results establish that suramin is effective in controlling growth in colonic tumor cells and confirms that suramin activity is synergistic with other chemotherapeutics. The effect of suramin on MDR is of potential value that needs to be more thoroughly investigated particularly in cancers such as the those of the colon that are often drug refractory.

Chronic inflammation and cancer: potential role of Bcl-2 gene family members as regulators of cellular antioxidant status.

The incidence of cancer is increased in all gastrointestinal (GI) tissues as a result of chronic inflammation. These cancers may develop in cells that are selected for resistance to inflammatory products by virtue of overexpressing the antioxidant protein Bcl-2 or other Bcl-2 gene family members that have antioxidant activity. Unfortunately, Bcl-2 also functions as an anti-apoptotic. Bcl-2 overexpression can increase cellular lifetime and increase the likelihood that other cancer-related mutations will accumulate in individual cells. Thus, Bcl-2 expression is cytoprotective, but its expression also increases the risk of cancer incidence. This is perhaps more evident in the GI tract, which is exposed to more potential mutagens relative to other tissues.

Changes in albumin levels in blood and urine of Microtus montanus chronically infected with Trypanosoma brucei gambiense.

Serum albumin and glucose concentrations and urinary excretion of alpha-keto acids and proteins were determined in samples obtained throughout a chronic Trypanosoma brucei gambiense infection in Microtus montanus. An increase in urinary excretion of alpha-keto acids and proteins during the terminal stage of disease was accompanied by a decrease in serum glucose concentration. This terminal hypoglycemia reflected a depletion of liver glycogen in most animals. In contrast (and the major focus of this study) serum albumin concentration was decreased by the second week of infection and in the final sample obtained was less than 50% of that measured in preinfection samples. Female animals survived approximately 1 wk longer than males and were less susceptible during the acute phase of disease. This relative resistance was most likely due to the fact that female animals were relatively more efficient in limiting parasitemia during the first week of infection. The similarity between humans and voles in terms of protein and alpha-keto acid excretion and changes in serum concentrations of glucose and albumin during trypanosome infection further validate the use of Microtus as an experimental model for trypanosomiasis in humans.

Tissue alterations in Microtus montanus chronically infected with Trypanosoma brucei gambiense.

Changes in liver, spleen, kidneys, heart, and brain are reported for Microtus montanus chronically infected with Trypanosoma brucei gambiense. An increase in body weight of infected animals was attributable to a significant increase in total mass of spleen, liver and kidney. Cellular infiltrate consisting primarily of lymphocytes and plasma cells was observed in all organs and was particularly evident in intralobular connective tissue of the liver, adipose tissue of the hilum, and adjacent medullary region of the kidney, spleen, and the meninges. Disruption of normal metabolism and the pathological changes observed in liver and kidney suggest that the survival of trypanosome-infected voles is dependent largely on the physiological response occurring in these organs.

Trypanosoma brucei gambiense and T. b. rhodesiense: concanavalin A binding to the membrane and flagellar pocket of bloodstream and procyclic forms.

We have measured binding of fluorescein-conjugated succinyl-concanavalin A (Fl-s-Con A) to bloodstream and procyclic forms of Trypanosoma brucei gambiense and to bloodstream forms of T. b. rhodesiense by flow cytofluorimetry. Bloodstream forms bound an order of magnitude less lectin than procyclic forms. Trypsin-treating cells enhanced binding of Fl-s-Con A to bloodstream forms 3-16-fold depending on the strain and the length of trypsinization but had little effect on Fl-s-Con A binding by procyclics. The trypsinization protocol used did not remove major common glycoproteins detected on lectin blots of either life cycle form but removed greater than 95% of the variant specific glycoprotein and fragments derived from this protein of bloodstream forms. Microscopically detectable Fl-s-Con A binding to bloodstream forms was confined to the flagellar pocket. Trypsinized bloodstream forms and procyclics bound Fl-s-Con A in the flagellar pocket, on the flagellum, and on the cell surface. Lectin remained cell associated but appeared to redistribute towards the flagellum and pocket when cells that had bound lectin on ice were subsequently incubated at physiological temperatures. The Fl-s-Con A binding had specificity characteristic of the interaction between the lectin and oligosaccharides. These results are consistent with the hypothesis that the variant specific surface glycoprotein blocks binding of the lectin to surface glycoproteins of bloodstream forms and suggest that concanavalin A-binding glycoproteins are abundant in the flagellar pocket of both life cycle forms.

Porphyromonas gingivalis, Actinobacillus actinomycetemcomitans and Treponema denticola detection in oral plaque samples using the polymerase chain reaction.

Detection of putative pathogens is critical for delineating the etiology and progression of periodontitis. In the present study, we have used a polymerase chain reaction (PCR) assay utilizing primers specific for the lkt A gene of Actinobacillus actinomycetemcomitans, the fimbrial gene of Porphyromonas gingivalis, and tdp A gene of Treponema denticola in order to determine the presence of these pathogens in subgingival plaque samples from periodontitis sites. These gene specific primers were also used to assess the detection of different strains of bacteria in the PCR assays. Primers for P. gingivalis detected P. gingivalis strain 33277, but no product was detected when primers were used with extracts from 4 species of Capnocytophaga, 3 species of Prevotella, 2 species of Porphyromonas other than P. gingivalis, Bacteroides levii, Escherichia coli, 3 strains of A. actinomycetemcomitans and 3 strains of T. denticola. PCR analysis using primers for the lkt A gene of A. actinomycetemcomitans also did not result in a product with any of these bacteria with the exception of a positive result with 3 different strains of A. actinomycetemcomitans. Primers selected from the tdp A gene of T. denticola did not identify any of the bacteria strains tested except T. denticola serovars a, b, and c. Thus, these primers were shown to amplify gene segments that are specific to either P. gingivalis (33277), A. actinomycetemcomitans (33384, 43717 and 43718) or T. denticola (35405, 33521 and 35404). The PCR assay may be used to rapidly detect the presence of periodontal pathogens in the future.

Trypanosoma brucei brucei, T. brucei gambiense, and T. brucei rhodesiense: common glycoproteins and glycoprotein oligosaccharide heterogeneity identified by lectin affinity blotting and endoglycosidase H treatment.

Whole cell extracts of 10 clones of bloodstream forms of African trypanosomes representing two strains of Trypanosoma brucei gambiense, one strain of T. b. rhodesiense and one strain of T. b. brucei were fractionated on sodium dodecyl sulfate-polyacrylamide gels, electrophoretically transferred to nitrocellulose paper, and probed with horseradish peroxidase conjugated lectins to detect glycoproteins. Variant specific glycoproteins of all 10 clones bound peroxidase labeled concanavalin A, but peroxidase labeled wheat germ agglutinin bound to the variant specific glycoproteins of only 3 of the 10 clones examined. In addition, 22 other glycoproteins expressed in common by all clones bound peroxidase labeled concanavalin A; 19 common glycoproteins bound peroxidase labeled wheat germ agglutinin. Lectin binding to transferred glycoproteins was specifically inhibited by appropriate monosaccharides, alpha-methyl mannoside for concanavalin A and N-acetyl glucosamine for wheat germ agglutinin. Prior incubation of blots in endo-beta-N-acetylglucosaminidase H eliminated binding of peroxidase-labeled concanavalin A to most of the 22 common glycoproteins. Two glycoproteins, designated Gp 81 and Gp 110, were the major Endoglycosidase H resistant components. Endoglycosidase H treatment also reduced binding of peroxidase labeled concanavalin A to the variant specific glycoproteins of 7 clones. The variant specific glycoproteins from the 3 clones that bound peroxidase labeled concanavalin A following enzyme treatment were those that bound peroxidase labeled wheat germ agglutinin. These results show that African trypanosomes express a greater number of glycoproteins than has been reported previously and that only a limited number of these glycoproteins bear Endoglycosidase H resistant oligosaccharides.

Trypanosoma brucei brucei and Trypanosoma brucei gambiense: stage specific differences in wheat germ agglutinin binding and in endoglycosidase H sensitivity of glycoprotein oligosaccharides.

Gel electrophoresis, lectin affinity blotting, and endoglycosidase H digestion have been used to analyze the glycoprotein profiles of bloodstream and procyclic forms of Trypanosoma brucei brucei and T. b. gambiense. Proteins resolved by polyacrylamide gel electrophoresis were stained with silver nitrate or electrophoretically transferred to nitrocellulose and probed with a horseradish peroxidase conjugate of either concanavalin A or wheat germ agglutinin. Silver staining showed, as expected, that the expression of the variant specific glycoprotein was restricted to the bloodstream forms. Twenty-three concanavalin A binding proteins were resolved in blots of bloodstream forms. Concanavalin A binding molecules corresponding in electrophoretic mobility to 21 of these 23 bloodstream form glycoproteins were detected in blots of procyclic forms. The two concanavalin A binding glycoproteins present only in bloodstream form extracts were variant specific glycoprotein and an 81-kDa protein designated glycoprotein 81b. One concanavalin A binding molecule of 84 kDa, glycoprotein 84p, was detected only in procyclic forms. The 19 major wheat germ agglutinin binding glycoproteins expressed by bloodstream forms were not detected in procyclic forms; only small proteins or protein fragments in procyclic form extracts bound wheat germ agglutinin. Incubating transferred proteins in endoglycosidase H eliminated subsequent binding of concanavalin A to most of the 22 common glycoproteins of bloodstream forms. Three major concanavalin A binding glycoproteins of bloodstream forms, variant specific glycoprotein, glycoprotein 81b, and a 110-kDa molecule (glycoprotein 110b), and other minor glycoproteins carried sugar chains that resisted endoglycosidase H digestion. In contrast, concanavalin A did not bind to any procyclic form glycoproteins, including a 110-kDa concanavalin A binding molecule (glycoprotein 110p) after endoglycosidase H treatment.(ABSTRACT TRUNCATED AT 250 WORDS)

Trypanosoma brucei sspp.: cleavage of variant specific and common glycoproteins during exposure of live cells to trypsin.

Intact bloodstream forms of Trypanosoma brucei brucei, T.b. gambiense, and T.b. rhodesiense and procyclic forms of T.b. brucei and T.b. gambiense were incubated in trypsin, solubilized for gel electrophoresis, and analyzed for removal of surface molecules. Silver-stained gels and transfer blots probed with horseradish peroxidase-conjugated or radiolabeled lectins revealed that only three glycoproteins, Gp120p, Gp91p, and Gp23p, were removed from the surface of procyclic forms by trypsin. The variant specific glycoproteins, Gp23b, Gp120b, and in some clones Gp91b were surface molecules cleaved from bloodstream forms. Greater than 90% of the variant specific glycoprotein (VSG) was removed from the surface of all clones studied within 1 hr following the addition of trypsin. The removal of VSG was coincident with appearance of 37 to 50 kDa glycopeptide fragments of VSG with different clones yielding different sized fragments. Detailed kinetic analysis of proteins from whole cell extracts and supernatants of the DuTat 1.1 clone of T.b. rhodesiense using concanavalin A (Con A) and polyclonal antibodies revealed that three major VSG fragments were released during trypsinization. The electrophoretic mobility of the three VSG fragments of DuTat 1.1 was not altered when samples were boiled in sodium dodecyl sulfate to inhibit the endogenous phospholipase C. Antiserum to the cross-reactive determinant bound to intact VSG, but did not bind VSG fragments. Thus, the major Con A binding fragments of DuTat 1.1 VSG and perhaps those of the other clones we studied were probably derived from the N-terminal domain of the molecule. The data suggest that VSG is cleaved by trypsin in situ at the hinge region, but remains attached to the cell surface via weak interaction with neighboring molecules.

In vitro beta-carotene toxicity for human colon cancer cells.

Experiments were conducted to determine the effect of beta-carotene on human colon cancer cells in vitro. beta-Carotene solubilized in tetrahydrofuran (THF) was determined to be cytotoxic for three different cell lines: LS 180, SW 620, and HCT-15. The number of LS 180 and SW 620 cells surviving treatment with 2.9 microM beta-carotene was significantly reduced relative to THF-treated cells, and a similar reduction was achieved in HCT-15 cells with use of 5.8 microM beta-carotene. These concentrations are in the range achieved in serum of individuals supplemented with beta-carotene at 30 mg/day. There was no beta-carotene cytotoxicity in the concentration range that characterizes serum of unsupplemented individuals. Vitamin E at > 200 microM was not cytotoxic and at higher concentrations slightly stimulated proliferation of all three cell lines. Exposure of cells to vitamin E did not diminish the cytotoxicity of beta-carotene, suggesting that the toxic effect of beta-carotene is not due to prooxidant activity. Percent cytotoxicity was increased by extending the duration of exposure of cells to beta-carotene. Interestingly, beta-carotene cytotoxicity decreased with increasing cell density. This density-dependent toxicity was attributable to a higher beta-carotene concentration per cell for cells plated at lower densities. Thus toxicity of beta-carotene for colon cancer cells is dose, time, and cell density dependent and occurs in vitro at concentrations that can be achieved safely in humans.

Expression of c-myc in human colonic tissue in response to beta-carotene supplementation.

Dietary supplementation with beta-carotene at 30 mg/day results in an increased serum trans-retinoic acid concentration in patients with a prior colonic polyp. In a number of human cell lines, trans-retinoic acid upregulates c-myc mRNA expression in colonic mucosa by reverse transcription polymerase chain reaction and correlated the results with serum concentrations of all-trans- (ATRA), 13-cis-(13-cRA), and total retinoic acid. Serum and colonic biopsy samples were obtained before and 90 days after administration of a placebo (n = 7) or 30 mg of beta-carotene (n = 5) daily. An increase in c-myc expression after supplementation was observed in 6 of 12 subjects, but 5 of these 6 subjects had decreased total serum retinoic acid concentration and 4 had decreased ATRA concentration. In addition, five of the six subjects with increased c-myc expression had received a placebo. Conversely, c-myc expression was increased in only two of five paired samples from subjects whose total serum retinoic acid concentration increased during the 90-day supplementation period. We conclude that c-myc expression is not correlated with ATRA, 13-cRA, or total retinoic acid concentration in vivo and that increased serum retinoic acid secondary to increased tissue beta-carotene is not sufficient to activate c-myc transcription.

Immunohistochemical evaluation of Bcl-2 gene family expression in liver of hepatitis C and cirrhotic patients: a novel mechanism to explain the high incidence of hepatocarcinoma in cirrhotics.

OBJECTIVE: The purpose of this study was to determine whether there is an increase in expression of bcl-2 and related bcl-2 gene family members bcl-X and bax in liver biopsy samples obtained from patients with either hepatitis C infection or cirrhosis. Bcl-2, bcl-X, and bax, as well as other bcl-2-related proteins, function coordinately through homo- and heterodimerization to regulate apoptosis. Bcl-2, which is characterized as an antiapoptotic, also functions as an antioxidant. We hypothesized that a mechanism that could account for increased hepatocellular carcinoma in patients with hepatitis C and cirrhosis is selection of bcl-2 expressing cells. This selection would be due to the capacity of individual cells to resist the toxic effects of inflammatory byproducts, specifically reactive oxygen species. METHODS: Sections cut from archived liver biopsy samples embedded in paraffin were probed with antibody specific for bcl-2, bcl-X, or bax. Liver samples were from normal (N = 5), hepatitis C patients (N = 19), and cirrhotics (N = 10). Percent positive staining and intensity of staining were judged independently for hepatocytes, bile ducts, mononuclear cells, and Kupffer cells. RESULTS: Bcl-2 expression was evident in bile ducts and mononuclear cells of hepatitis C patients, but was not commonly present in hepatocytes (two of 10). In the cirrhotic liver, bcl-2 expression was also detected in bile ducts and mononuclear cells, but in contrast to hepatitis patients was also expressed in hepatocytes (nine of 10). A similar pattern of expression was evident for bcl-X, but in general the level of expression was limited relative to that of bcl-2. Bax expression was infrequently present in sections from any of the three patient groups. CONCLUSIONS: The data indicate that bcl-2 expression is elevated in the liver of cirrhotics, but is not evident in the liver of hepatitis C patients. This increase in expression of bcl-2 in cirrhotic patients may correlate with development of hepatocellular carcinoma given the anti-apoptotic/oncogenic potential of bcl-2.