Cytotoxicity of Natural Compounds in Hepatocyte Cell Culture Models. The Case of Quaternary Benzo[c]phenanthridine Alkaloids
|
|  Yüklə 66.5 Kb.
|
| Cytotoxicity of Natural Compounds in Hepatocyte Cell Culture Models. The Case of Quaternary Benzo[c]phenanthridine Alkaloids
Jitka Ulrichová1*, Zdeněk Dvořák1, Jaroslav Vičar1, Jan Lata2, Jana Smržová2, Aleksi Šedo3 and Vilím Šimánek1 1Institute of Medical Chemistry and Biochemistry, Palacký University, 775 15 Olomouc, Czech Republic, 23rd Med.-Gastroenterology Department of University Hospital, 639 00 Brno, Czech Republic and 31st Institute of Medical Chemistry and Biochemistry, Charles University, 121 08 Prague, Czech Republic Key words: sanguinarine; chelerythrine; fagaronine; Macleaya cordata; human hepatocyte; porcine hepatocyte --------------------------- Corresponding author: Jitka Ulrichová, Ph.D., Institute of Medical Chemistry and Biochemistry, Palacký University, 775 15 Olomouc, Czech Republic, fax: ++420-68-5632966, phone: ++420-68-5632312, Email: ulricho@tunw.upol.cz ABSTRACT The quaternary benzo[c]phenanthridine alkaloids (QBA) produce a plethora of species- and tissue-specific effects but the molecular basis of their biological activities remain mysterious. The objective of the present study was to investigate the cytotoxicity of QBA alkaloids, sanguinarine (SA), chelerythrine (CHE), fagaronine (FA), and the extract from Macleaya cordata in primary cultures of human and porcine hepatocytes. The cellular damage was assessed by the MTT assay, lactate dehydrogenase (LDH) leakage and the determination of intracellular glutathione (GSH) levels. The results are summarized as follows: (i) The alkaloids tested in doses 0.1 and 10 M did not display statistically significant cytotoxicity for 0 – 3 h incubation. (ii) SA and CHE showed the dose- and time-dependent toxicity within the range 25 to 100 M whereas FA was not toxic. (iii) The LDH leakage into the medium was higher for SA than for CHE, thus revealing a potent potential of SA to disturb cell-membrane integrity. (iv) After 3h incubation with 100 M SA/CHE, mitochondrial dehydrogenase activity (MTT assay) and the cellular GSH levels decreased to residual values of about 40% suggesting that mitochondria are unlikely to be a primary target for SA/CHE in the cell. (v) No differences were found in the response to QBA application in human vs. porcine hepatocyte. INTRODUCTIONSanguinarine (SA), chelerythrine (CHE) and fagaronine (FA), are quaternary benzo[c]phenanthridine alkaloids (QBA) and display a wide spectrum of non-specific biological activities (Walterová, 1995). In contrast to SA and CHE, FA has a different substitution pattern in the A-ring (Fig.1). QBA are found in plants of families Caprifoliaceae, Fumariaceae, Meliaceae, Papaveraceae, and Rutaceae (Šimánek, 1985). QBA are the elicitor-inducible secondary metabolites and are considered phytoalexines because of their anti-microbial activity. Two standardized QBA extracts are available on the market. Sanguiritrine, a QBA extract from Macleaya species (mixture of SA, CHE and three minor QBA), serves as a remedy for internal uses in myopathy therapy and as antimicrobial drug in Russia, and sanguinaria, a QBA extract from Sanguinaria canadensis (mixture of SA, CHE and four minor QBA), is used as the antiplaque component in dentifrices and oral rinses manufactured in Europe and USA (Tenenbaum, 1999). Sanguinaria was used in the USA as expectorans but FDA classified this remedy as unsafe for uses in drugs (MARTINDALE, 1996). Interestingly, the veterinary preparation Sangrovit, which contains approx. 4 % of the QBA extract – sanguinaria, is advertised as a weight gain stimulant in farm animals (Anonymous, 2000). FA possesses antileukemic activity (Suffness, 1985), inhibits HIV-1 and HIV-2-reverse transcriptase (Tan, 1991; Kerry, 1998), and DNA topoisomerases I and II (Wang, 1993). FA is being considered as a potential antitumor drug (Kerry, 1998). As with many natural compounds, QBA exhibit toxic side effects in addition to their desired pharmacological activity. Contamination of edible oils with the oil from seeds of Argemone mexicana and A. ochroleuca was suggested to be the cause for the disease epidemic dropsy (Das, 1997). The principal constituent responsible for argemone oil toxicity was identified as sanguinarine. Damm et al. (Damm, 1999) reported that the long term use of oral products containing sanguinaria appears to be associated with an increased prevalence of leukoplakia of the maxillary vestibule. However, this conclusion has been criticized (Munro, 1999). From in vivo experiments with rats and guinea pigs, it was concluded that the lipoid structures in the gastrointestinal tract, liver and kidneys are the target sites for SA (Tandon, 1992). SA and CHE adverse effects may be due to their non-selectivity towards specific target structure or to toxicity that results from metabolic activation to reactive intermediates, which are capable of covalent binding to cellular macromolecules. Their covalent binding to proteins can cause cell death by disturbing essential biochemical processes or immunological reactions, whereas binding to DNA may initiate the process leading to cancer. The positive charge on nitrogen atom facilitates the interaction binding with DNA. In this respect, it is FA that is most active (Walterová, 1995). Recently, the toxicity of SA and CHE in rat hepatocytes (Ulrichová, 1996) and cytotoxic effects of SA in cultured human fibroblasts (Karjalainen, 1988) were described. More studies on the molecular level are needed to clarify the adverse effects of both alkaloids on cells in vitro, and to expand our knowledge about potential toxicity in vivo. Primary human hepatocytes are currently considered to be an appropriate model for the assessment of safety of pharmaceutical and nutritional products. Furthermore, the studied compounds are a part of a veterinary preparation for pigs. Therefore, in this study, both human and porcine hepatocytes were used. MATERIALS AND METHODSAlkaloids and chemicalsSanguiritrine, a QBA extract from M. cordata, containing 71% QBA in ratio SA:CHE:minor QBA (6.3:2.4:1.3, determined by HPLC) was purchased from CAMAS Technologies, Inc. (Broomfield, USA). Sanguinarine and chelerythrine were isolated from sanguiritrine using column chromatography on alumina (Dostál, 1992). Sanguinarine in 98.1% purity, MP 279–282oC ((Southon, 1989) 277–280oC) and chelerythrine in 95% purity, MP 200-204oC ((Southon, 1989) 202-203oC) were obtained. Fagaronine was synthesized by Šmidrkal (Šmidrkal, 1988), MP 203–206oC ((Southon, 1989) 202oC). IR, UV, MS and NMR spectra were consistent with the structures of the above alkaloids. Collagenase cruda was purchased from Sevac (Czech Republic), Trypan blue from Merck (Germany). William’s medium E and other chemicals were from Sigma. Hepatocyte isolation and cultivationHuman liver tissue was obtained from multiorgan donors, with the approval of Czech Ethical Committee. Porcine liver were obtained from experimental mini-pigs (MeLiM strain, 25 kg, Institute of Animal Physiology and Genetics, Liběchov, Czech Republic). The hepatocytes were isolated from the HTK pre-washed liver using two-step collagenase perfusion (Pichard, 1990). Yield and viability of cells were assessed by Trypan blue exclusion test. Hepatocytes were resuspended in ISOM medium, consisting of 1:1 mixture of Ham F12 and William´s E, supplemented with additives as follows: glucose (7 mM), glutamine (2.4 mM), penicillin (100 U/mL), streptomycin (10 µM), sodium pyruvate (0.4 mM), dexamethasone (1.8 µM), holo-transferrin (5 mg/L), ethanolamine (1 µM), insulin (350 nM), glucagon (0.2 mg/L), linolic acid (11 µg/L), ascorbic acid (15 mg/L), amphotericin B (1.4 mg/L), pH 7.2. (Modrianský, 2000). Cells were seeded on collagen-coated 6-well plates (area 9.4 cm2) in density of 1.25 x 105 cells/cm2. For the first 4 h, culture medium was enriched with 5% of foetal calf serum to improve cell attachment. Following the stabilisation period of 24–48 h, hepatocytes were treated for 1–4 h with test compounds (final concentrations 0.1, 10, 25, 50, 75, 100 M) and/or with vehicle (DMSO) for control. The cultivation and exposures were carried out under sterile conditions, using a humidified incubator at 37°C under an atmosphere containing 5% CO2. Integrity of the cell membrane was determined by the leakage of lactate dehydrogenase (LDH) into the medium. Depletion of reduced glutathione (GSH) was estimated by 5,5’-dithio-bis(2-nitrobenzoic acid) (DTNB). Functionality of hepatocytes was assessed by the MTT test (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). Cultures were routinely checked under an Olympus CK-2 microscope. LDH assayEnzyme activity was measured in 50 L samples by determining the decrease in absorbance at 340 nm for 3 min in reaction mixture of 1.22 mM pyruvate (1 mL) in 50 mM phosphate buffer (pH 7.5) + 20 L NADH (6.2 mg/mL) and expressed in nkat/106 cells. GSH levelHepatocytes were washed and harvested in 0.5 mL of PBS with 0.1% of Triton-X100. An aliquot of obtained lysate (0.36 mL) was denatured with 0.04 mL cold trichloroacetic acid (25%). After 10 min of incubation on ice, the mixture was centrifuged (3000 x g, 10 min, 4°C) and 0.3 mL of the supernatant was mixed with 1.0 mL of Tris-base (0.8 M) – EDTA (0.02 M) buffer, pH 8.9. Following the addition of 0.1 mL of DTNB (0.01 M) in methanol, absorbance of the sample was measured at 412 nm against DTNB blank. The level of GSH (nmol/106 cells) was calculated using absorptivity of 13 600 cm-1.M-1. The standard solution of GSH (1.0 mM dissolved in deionised water) was used for the control of the recovery and reproducibility of the measurement. MTT assayAfter the treatment, culture medium was replaced with fresh and 100 L of MTT reagent (5 mg/mL PBS) was added. Three h later, the medium was removed, cells washed with PBS and lysed for 5 min with 1 mL of DMSO containing 1% ammonia. The lysate was diluted 20 times with DMSO (+ 1% ammonia) and absorbance at 540 nm was measured against blank (DMSO + 1% ammonia). Results were normalised to the control value (i.e. 100 x Asample/Acontrol) and expressed as percentage of control. Statistical analyses Data were analysed by ANOVA, followed by Dunnett´s test. Results are presented as mean ± SD. A p value of < 0.05 was considered to be statistically significant. RESULTS The addition of alkaloids SA and CHE to primary cultures of human and porcine hepatocytes resulted in time and dose–dependent cell death as monitored by LDH leakage, GSH depletion and decrease of mitochondrial dehydrogenase activity (MTT assay), while FA did not affect cell viability and functionality by any of the parameters mentioned above (Tab. 1). Similar results were obtained with rat hepatocytes (Ulrichová, 1996). SA was more potent than CHE in hepatocytes from all three species. Sanguiritrine was identical to sanguinarine by all evaluated parameters of hepatocyte damage. The time course of the evaluated parameters displayed a rapid increase in LDH leakage into the culture medium (i.e. increased plasma membrane permeability) reaching a value corresponding to nearly 100% damage after 1 h incubation with SA (Fig. 2) and a slower rate of decrease of GSH levels and of MTT reduction capacity. Both parameters required 3 h incubation with SA 100 μM to reach residual values of about 40% of the controls (Tab. 1; Fig. 3, 4, 5). The data on sanguiritrine were, again, almost identical to those of SA on all evaluated parameters of hepatocyte damage (data not shown). In FA–treated hepatocytes the LDH, GSH as well as MTT time courses were almost identical to the controls. Quality of primary cultures was assessed microscopically. Primary cultures of control cells maintained a typical polygonal shape with intercellular contact throughout the whole incubation period, forming a homogenous monolayer attached to the collagen surface (Fig. 6A). Addition of SA, CHE and sanguiritrine to the medium induced destruction of the monolayer, with destruction of cellular contacts, a change in shape from polygonal to circular, and detachment of the cells (Fig. 6B). DISCUSSION One of the prerequisites for QBA biological activity is the presence of an iminium bond. In SA and CHE, this bond is susceptible to a nucleophilic attack and consequently plays a key role in the inhibition of SH-proteins. Both alkaloids interconvert between the cationic vs. neutral form (i.e. hydroxide adduct or pseudobase) displaying sort of “Dr Jekyll and Mr Hyde” duality. They penetrate across the cell membrane in the hydrophobic pseudobase form acting as the pro-drugs and convert into the active cationic form once inside the cell. It is well known that pH plays a pivotal role in many cellular events (Johnson, 1998). Unlike the two alkaloids, FA, due to its different ring substitution pattern, exists at physiological pH as a cation only, is a weak electrophile and its iminium bond is not readily attacked by nucleophiles. Our results can be correlated with these differences in the iminium bond character, as SA and CHE always reacted differently when compared with FA. In line with our previous results on rat hepatocytes (Ulrichová, 1996), we observed that rapid damage to plasmatic membranes resulting in 100% LDH leakage preceded a slower decrease in GSH and MTT down to about 40% of their initial values. The slow decrease and the existence of residual mitochondrial activities show that alkaloid penetration into the mitochondria is slow and that mitochondrial functions are partially maintained even after toxic (100 μM) alkaloid doses. Consequently, mitochondria are likely not a target of the toxic action of these QBA alkaloids. Measurements of malondialdehyde production (MDA), a marker of lipid peroxidation, did not reveal any significant increase in lipid peroxidation (results not shown) which excludes oxidative injury as the cause of QBA toxicity. In conclusion, we attribute cytotoxic action of SA and CHE to their binding to the vitally important SH-peptides (glutathione) and cellular proteins (most probably within the cytochrome P-450 enzyme family). Although our results demonstrate cytotoxic properties of sanguinarine and chelerythrine on porcine and human hepatocyte cultures, we still can not conclude unambiguously whether similar oral doses of either alkaloid are toxic or not. In vivo experiments reporting SA hepatotoxicity employ a very high dose (10 mg/kg bw ip to mice) relative to potential exposure under normal use (Dalvi, 1985; Williams, 2000). Furthermore, the method for SA determination in organ tissues is problematic and reported values questionable (Tandon, 1992). On the other hand, chronic administration of sanguinarine to rats (0.2 mg/kg bw ip, 56 days) did not induce any detectable histopathological changes of the liver tissue (Ulrichová, 1996). Since we have found that QBA alkaloids possessing an iminium bond susceptible to nucleophile attack strongly bind to SH-proteins (Šedo, 2001), we might expect that the extent of their transport from the gastrointestinal tract to the liver is low. To investigate this, experiments with labeled QBA alkaloids are under way. Considering available data together, it seems that it is still very difficult to interpret both in vivo and in vitro sanguinarine and chelerythrine toxicity in terms of mechanism of their action. The resulting toxicity reflects the sum of individual biological effects whose character and proportion we are unable to assess. Nevertheless, before we have more conclusive arguments, prudence is recommended in using QBA-containing plant extracts in human as well as in veterinary applications.
Das, M., Khanna, S.K., 1997. Clinicoepidemiological, toxicological, and safety evaluation studies on argemone oil. Crit. Rev. Toxicol 27(3), 273-297. Dostál, J., Táborská, E., Slavík, J., 1992. Preparative column chromatography of quaternary benzophenanthridine alkaloids of Dicranostigma lactucoides. Fitoterapia 63, 67-70. Johnson, I., 1998. Fluorescent probes for living cells. Histochem. J. 30(3), 123-140. Karjalainen, K., Kaivosola, S., Seppa, S., Knuuttila, M.,1988. Effects of sanguinaria extract on leucocytes and fibroblasts. Proc. Finn. Dent. Soc. 84(3), 161-165. Kerry, M.A., Duval, O., Waigh, R.D., Mackay, S.P., 1998. The role of the iminium bond in the inhibition of reverse transcriptase by quaternary benzophenanthridines. J. Pharm. Pharmacol. 50, 1307-1315. Martindale The Extra Pharmacopeia, 1996. Edition 31, The Pharmaceutical Press, London, pp. 1410. Modrianský, M., Ulrichová, J., Bachleda, P., Anzenbacher, P., Anzenbacherová, E., Walterová, D., Šimánek, V., 2000. Human hepatocyte--a model for toxicological studies. Functional and biochemical characterization. Gen.Physiol.Biophys. 19(2), 223-235.
Suffness, M., Cordell, G.A., 1985. Antitumor alkaloids. In: Brossi, A. (Ed.), The Alkaloids, Vol. 25. Academic Press, New York, pp. 3-345. Tan, G.T., Pezzuto, J.M., Kinghorn, A.D., Hughes, S.H., 1991. Evaluation of natural products as inhibitors of human immunodeficiency virus type 1 (HIV-1) reverse transcriptase. J. Nat. Prod. 54, 143-154. Tandon, S., Das, M., Khanna, S.K., 1992. Biometabolic elimination and organ retention profile of argemone alkaloid, sanguinarine, in rats and guinea pigs. Drug Metab. Dispos. 21(1), 194-197. Tenenbaum, H., Dahan, N., Soell, M., 1999. Effectiveness of a sanguinarine regimen after scaling and root planing. J. Periodontol. 70, 307-311. Ulrichová, J., Walterová, D., Vavrečková, C., Kamarád, V., Šimánek, V.,1996. Cytotoxicity of Benzo[c]phenanthridinium Alkaloids in Isolated Rat Hepatocytes. Phytother. Res. 10, 220-223. Walterová, D., Ulrichová, J., Válka, I., Vičar, J., Vavrečková, C., Táborská, E., Harkrader, R.J., Meyer, D.L., Černá, H., Šimánek, V., 1995. Benzo[c]phenanthridine alkaloids sanguinarine and chelerythrine: biological activities and dental care applications. Acta Univ. Palacki. Olomuc. (Olomouc), Fac. Med. 139, 7–16. Wang, L.K., Johnsoh, R.K., Hecht, S.M., 1993. Inhibition of topoisomerase I function by nitidine and fagaronine. Chem. Res. Toxicol. 6(6), 813-818. Williams, M.K., Dalvi, S., Dalvi, R.R., 2000. Influence of 3-Methylcholanthrene Pretreatment on Sanguinarine Toxicity in Mice. Vet. Hum. Toxicol. 42, 196-198. Table 1. Effect of alkaloids on human and porcine hepatocytes after 3 h incubation. The data are mean + SD, n = 5. Values for fagaronine were not significantly different from control. Values for sanguinarine, chelerythrine, fagaronine and sanguiritrine in concentrations 0.1, 1 and 10 µM were not significantly different from control. All values for sanguinarine, chelerythrine and sanguiritrine were different at p0.05 vs control values. aConcentration of sanguiritrine was expresed relative to combined sanguinarine and chelerythrine content.
FIGURE LEGENDS Fig. 1. Structures of quaternary benzo[c]phenanthridine alkaloids. Fig. 2. Effect of sanguinarine (SA, 50 µM), chelerythrine (CHE, 50 µM) and fagaronine (FA, 50 µM) on LDH leakage in human hepatocytes (A) and porcine hepatocytes (B). % = the portion of the enzyme activity released into the medium; data are mean + SD, n = 5. Comparisons: SA vs CHE (p0.01); SA and CHE vs FA and control (p0.001). Fig. 3. Effect of sanguinarine (SA, 50 µM), chelerythrine (CHE, 50 µM) and fagaronine (FA, 50 µM) on intracellular GSH content in human hepatocytes (A) and porcine hepatocytes (B). Data are mean + SD, n = 5. Comparisons: SA vs CHE (Non-significant); SA and CHE vs FA (p0.01); FA vs control (p0.05). Fig. 4. Effect of sanguinarine (SA, 50 µM), chelerythrine (CHE, 50 µM) and fagaronine (FA, 50 µM) on mitochondrial dehydrogenases activity in human hepatocytes (A) and porcine hepatocytes (B). % - data were normalised on control value. Values are mean + SD, n = 5. Comparisons: SA vs CHE (Non-significant); SA and CHE vs FA (p0.001); FA vs control (p0.01). Fig. 5. Effect of sanguinarine (SA), chelerythrine (CHE), fagaronine (FA) and sanguiritrine (SI) on mitochondrial dehydrogenases activity in human hepatocytes (A) and porcine hepatocytes (B) after 3 h of treatment. % - data were normalised on control value. Values are mean + SD, n = 5. The values of IC50 were calculated from plots as follows: (A) Human hepatocytes: SA = 71 5 µM, CHE > 100 µM, SI = 42 3 µM, FA >> 100 µM; (B) Porcine hepatocytes: SA = 47 4 µM, CHE = 83 6 µM, SI = 42 3 µM, FA >> 100 µM. Fig. 6. Effect of sanguinarine on human hepatocytes; (A) control cells 48 h after plating, (B) cells treated with 25 µM sanguinarine for 2 h. Similar changes were observed at longer exposure periods (3 h) and for higher concentrations of CHE (data not shown). | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||