- Email: [email protected]
Interaction between cultured endothelial cells and macrophages: In Vitro model for studying flavonoids in redox-dependent gene expression
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
387
Synopsis This article focused on two methods to measure the activity of NF-KB. Both methods evaluate "post-IKB pbosphorylation" stages in the NF-KB activation cascade. In fact, EMSA performed with nuclear extracts provides an information only on NF-KB nuclear translocation and its ability to bind KB-DNA sequences. Likewise, the reporter gene assay is limited to assessing NF-KB-dependent gene expression no matter the mechanism that originally activated NF-KB. Nevertheless, the latter assay represents a more physiological and more reproducible way of measuring NF-KB activity in mammalian cells than the EMSA does. In order to obtain further insights into NF-KB signal transduction pathways, investigating IKB degradation and phosphorylation are recommended. The cloning and characterization of IKB kinases provided new testing possibilities based on the measure of their activity.
[35] Interaction between Cultured Endothelial Cells and Macrophages: In Vitro Model for Studying Flavonoids in Redox-Dependent Gene Expression By GERALD
RIMBACH, CLAUDE SALIOU, RAFFAELLA CANALI,
and
FABIO VIRGILI
Introduction A coculture model composed of human endothelial cells and murine macrophages mimicking the proinflammatory environment characterizing atherosclerotic plaque formation is described. Interferon-',/(IFN--,/)-activated macrophages generate reactive oxygen and nitrogen species (ROS and RNS). In turn, ROS and RNS induce the activation of the redox-sensitive transcription factor NF-KB, its transactivation activity, and the expression of monocyte chemotactic protein-1 (MCP-1), a gene containing NF-KB in its promoter side. We utilized this coculture model to study the role of nutritional antioxidants, such as polyphenols, in molecular events associated with the early stages of atheroma development. This article describes the effect of the flavonoid-rich extract from Ginkgo biloba leaves, EGb761, on redox-sensitive gene expression. Results described herein suggest that the model is suitable to assess: (i) the protective effect of flavonoids on oxidative stress induced by reactive oxygen and nitrogen species and (ii) the effect of flavonoids on the activation of signaling pathways leading to the expression of specific response genes, such as MCP- 1.
METHODS IN ENZYMOLOGY,VOL.335
Copyright© 2001 by AcademicPress All rightsof reproductionin any formreserved. 0076-6879/00$35.00
388
BIOLOGICALACTIVITY
[351
RAW 264.7 monocyte/macrophages activated by IFN-y
Human unbilical vein endothelial cells HUVEC FIG. 1. Schemeof the coculturemodel: RAW 264.7 cells are grown on a 0.45-1xmfilter. Following activationwith IFN-3,,RAW264.7cells are placed on top ofa ECV/HUVECdish in directcontactwith the culture medium (1.5 ml). After an appropriate coincubationtime, during which underlying cells are challenged by the reactive oxygen and nitrogen species generated by activated macrophages,the filter is removed and specific parameters (transcription factors activation, mRNAexpression, cellular antioxidant levels) are assessed. A Coculture System to Study Role of ROS and RNS in Macrophage-Endothelial Cell I n t e r a c t i o n The coculture model, illustrated in Fig. 1, is composed of primary human endothelial ceils (HUVEC) and the monocyte-macrophage murine cell line RAW 264.7. It is characterized by a long-term, sustained generation of high levels of ROS and RNS, thus mimicking the occurrence of a complex though mild "physiological" oxidative stress. This is confirmed by the significant accumulation of nitrite and nitrate (NO 2 and NO~-) in the medium after cell activation, which is an index of both nitric oxide (NO.) generation and NO* reaction with superoxide (see Fig. 2). This mode of generation of RNS appears somehow different from that induced by the addition of NO- donors such as 3-morpholinosydnonimine-Nethylcarbamide (SIN-1), sodium nitroprusside (SNP), and compounds either contalning or generating reactive oxygen (H202, organic peroxides, transition metals), l In fact, macrophages activated by the bacterial wall component lipopolysaccharide (LPS) or by cytokines such as IFN-~/ express the enzymatic machinery that generates both ROS and RNS. In such a condition, a direct interaction of NO"
1G. Rimbach, E Virgili, Y. C. Park, and L. Packer, Redox Rep. 4, 171 (1999).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
i + activated macrophages
389
+
+t+l++tWt++t+l[t+l+t+Wml +tt+t
D
+ dormant macrophages
NO3. + N O 2NO2_
controls
0
50
100
150
200
FIG. :2. NO~- and NO 3 accumulation in ECV/HUVEC medium (p,M) after a 24-hr exposure to a filter containing RAW 267.4 monocyte-macrophages activated by IFN-~,. NO 2 and NO 3 levels in the medium are quantified using the automated NOx analyzer based on Griess reagent. Nitrate is determined after reduction to nitrite using an A7200 copperized cadmium reduction column and is then quantified by Griess reagent.
with 0 2 ° - is likely to occur, leading to ONOO- formation. 2-6 This reaction can be observed in different cell lines and tissues such as macrophages, neutrophils, and cultured human endothelial cells. 3'70NOO- is a powerful oxidant molecule, which, due to its weak O-O bond strength, decomposes spontaneously to form the hydroxyl radical and nitrogen dioxide. 8 0 N O O - may react directly with many biological targets, such as lipids and sulfhydryl groups. It may also nitrate tyrosine in proteins, thereby forming 3-nitrotyrosine residues. Nitration of tyrosine can result in an inhibition of protein tyrosine phosphorylation and thus inhibit signal transduction. 2,9 We have already utilized this coculture model to demonstrate that the exposure to activated macrophages is associated with a significant decrease of a-tocopherol and glutathione levels in endothelial cells. In the same studies, we demonstrated
2 G. E. Arteel, K. Briviba, and H. Sies, FEBS Lett. 445, 226 (1999). 3 M. C. Carreras, G. A. Pargament, S. D. Catz, J. J. Poderoso, and A. Boveris, FEBS Lett. 341, 65 (1994). 4 G. L. Squadrito and W. A. Pryor, Free Radic. Biol. Med. 25, 392 (1998). 5 S. Moncada, R. M. Palmer, and E. A. Higgs, Pharmacol. Rev. 43, 109 (1991). 6 j. S. Beckman and W. H. Koppenol, Am. J. Physiol. 271, C1424 (1996). 7 N+ W. Kooy and J. A. Royall, Arch. Bioche. Biophys. 310, 352 (1994). 8 W. H. Koppenol, J. J. Moreno, W. A. Pryor, H. Ischiropoulos, and J. S. Beckman, Chem. Res. Toxicol. 5, 834 (1992). 9 H. Ischiropoulos, Arch. Biochem. Biophys. 356, 1 (1998).
390
BIOLOGICALACTIVITY
[35]
that procyanidins extracted from the bark of Pinus maritima protect the cellular antioxidant network from macrophage-induced oxidative challenge.l' 10 S o m e M e t h o d o l o g i c a l R e m a r k s R e g a r d i n g Cell L i n e s In the coculture model reported herein, a combination of two nonhomo-specific cell lines in the experimental design (RAW 264.7, a murine-derived cell line, with either ECV304 or primary endothelial ceils of human origin) was chosen in order to avoid the activation of endothelial cells by the cytokine utilized for macrophage activation. In fact, no changes were observed either on antioxidant levels or other functional cell parameters when ECV304 was exposed to the medium utilized for RAW 264.7 culture after activation with IFN-3,, l,l° confirming that recombinant routine IFN-~/utilized in our studies displays a species-specific activity. However, reports alerting the utilization of ECV304 as an appropriate model for the endothelium have appeared. In fact, several distinct differences between ECV304 and human umbilical vein-derived endothelial cells (HUVEC), as well as a strong genetic similarity between ECV304 and T24/83, a human bladder cancer cell line, have been reported. 11 The authors suggested that even though ECV304 displays some endothelial characteristics and may he a useful tool for the study of receptor pharmacology, it should not be considered equivalent to HUVEC and is therefore an inappropriate cell line to study endothelial cell biology. These observations led us to reconsider our model; we now use primary endothelial cells obtained from collagenase treatment of the human umbilical vein known as HUVEC instead. Therefore, data presented here only refer to the coculture of RAW 264.7 cells together with HUVEC. However it is worth mentioning that some of the features and responses displayed by "authentic" HUVEC are quite similar to those we have already reported for ECV304. This observation allows us to suggest that even taking into account the limitations mentioned earlier, some of the studies performed utilizing ECV304 as a model for endothelial cells should still be considered valid. Cell C u l t u r e Human Umbilical Vein Endothelial Cells
Primary human endothelial cells are obtained from the umbilical cord vein. According to the method of Jaffe and co-workers, 12 the main vein of umbilical cords (kindly provided by the nursery of the University of Roma, "La Sapienza") is 10 E Vipgili, D. Kim, and L. Packer, FEBS Lett. 431,315 (1998). 11 j. Brown, S. J. Reading, S. Jones, C. J. Fitehett, J. Howl, A. Martin, C. L. Longland, E Miehelangeli, Y. E. Dubrova, and C. A. Brown, Lab. Invest. ~ , 37 (2000). 12 E. Jaffe, R. L. Nachmann, C. G. Beeker, and C. R. Minick, J. Clin. Invest. 153, 5008 (1973).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
391
washed thoroughly with phosphate-buffered saline (PBS) and then treated with 10 ml collagenase (0.2%, dissolved in Medium 199) from Clostridium histiolyticum (type XI, Sigma St. Louis, MO) for 20 min at 37 °. The endothelial cell layer is removed from the vein using 50 ml of Hanks' buffer (Sigma). Cells are then pelleted by centrifugation and cultured in six-well culture dishes (3.5 cm diameter, Falcon, Franklin Lakes, NJ) pretreated with 1.5% gelatine (type B from bovine skin, Sigma). Cells are grown utilizing 199 Medium (GIBCO-BRL, Rockville, MD) supplemented with 20% fetal calf serum (FCS, University of California cell culture facilities, San Francisco, CA), 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 100 ~g/ml heparin, streptomycin-penicillin, and 50 ~g/ml endothelial cell growth supplement (ECGS, Sigma). HUVEC cells are utilized for experiments at 90-100% apparent confluence and within passage 3 through 6. Passages are performed according to standardized protocols and by diluting the cell population 1 to 3. RAW 264.7 Murine Monocyte-Macrophage RAW 264.7, a murine cell line of monocyte-macrophages, are obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells are seeded on 0.45-1~m pore-size, polyethylene terephthalate cell culture inserts (Falcon) and cultured in Dulbecco's modified Eagle's medium (DMEM, GIBCO-BRL) containing 10% FCS, 2 mM glutamine, 1 mM sodium pyruvate, and streptomycinpenicillin. RAW 264.7 are stimulated with 10 U/ml IFN-y (Genzyme, Cambridge, MA). Three hours after stimulation, the filters containing macrophages are placed on top of HUVEC cells in direct contact with the culture medium. Cells are then cocultured at 37 ° and, after an appropriate coincubation time, the filter is removed and specific parameters (transcription factors activation, mRNA expression) are assessed on HUVEC cells. Ginkgo biloba E x t r a c t EGb 761 EGb 761 is a standardized extract of G. biloba leaves. It is used commonly in Asia, Europe, and the United States to treat a variety of pathological conditions, such as peripheral arterial diseases and organic brain syndromes. EGb 761 contains 24% flavonoids (ginkgo-flavon glycosides) and 6% terpenois (ginkgolides, bilobalides) as active components. The flavonoid fraction is composed mainly of the flavonols quercetin, kaempherol, isorhamnetpferol, and isorhamnetin, which are linked to a sugar moiety. 13 This plant extract has been shown to have strong scavenging activity toward hydroxyl radicals, superoxide anion radicals, and NO'. EGb 761 may also contribute significantly to LDL protection from oxidation in 13 O. Sticher, Planta Med. 59, 2 (1993).
392
BIOLOGICAL ACTIVITY
[35]
vitro and to a lesser extent in v i v o . 14 Furthermore, it has been demonstrated in our laboratory and by others that EGb 761 inhibits the activation of NF-KB in endothelial cells and macrophages challenged by different stimuli. 15 All these features render EGb 761 potentially able to contribute to the protection from the development of artheriosclerotic lesion both by acting as a bona fide antioxidant and by modulating the signaling pathway that leads to an endothelial cell response to prooxidation and dysfunction. This article reports on some observation obtained utilizing the macrophageendothelial cell coculture model to study the effect of EGb 761 in the modulation of the endothelial cell response to the exposure to activated macrophages. In the experiments described herein, EGb 761 is dissolved in dimethyl sulfoxide (DMSO) and added to HUVEC culture medium at the indicated concentrations. Cells are supplemented with the extract for 16 hr prior to any further treatment. The final DMSO concentration in the medium is less than 0.1%. Control cells, i.e., those not supplemented with EGb 761, are treated with identical amounts of DMSO. C o c u l t u r i n g M a c r o p h a g e s a n d E n d o t h e l i a l Cells a s S u i t a b l e M o d e l for S t u d y of F l a v o n o i d s in A c t i v a t i o n of NF-KB P a t h w a y I n d u c e d by Oxidative-Nitrosative Stress Coincubation with Activated RAW 264. 7 Monocytes/Macrophages to Induce Transfer of NF-KB into Nucleus and DNA Binding in HUVEC Figure 3 shows the electromobility shift assay (EMSA) performed according to Suzuki and co-workers 16 on the nuclear extract of HUVEC cells after exposure to RAW 264.7 monocytes/macrophages either quiescent or activated by treatment with 10 U/ml murine IFN-~/. In the basal condition (lane 1, Fig. 3), HUVEC cells do not show significant DNA-binding activity, which is induced strongly by TNFot (lane 2, Fig. 3). Similarly, following the exposure to the medium utilized to culture RAW 264.7 macrophages, transferred to the top of HUVEC cells 3 hr after the addition of murine IFN-~/, no significant changes are observed in the level of NF-KB binding to the specific DNA consensus sequence (lane 3, Fig. 3). After the exposure to nontreated RAW 264.7 cells, the level of NF-KB-binding activity to DNA increases slightly (lane 5, Fig. 3). The increase is more dramatic after the coincubation with macrophages activated by IFN-~/ (lane 4, Fig. 3). This result indicates that activated macrophages significantly induce the activation of the signaling pathway leading to NF-KB transfer to the nucleus and possibly the expression of downstream genes in HUVEC cells. It is noteworthy that this observation
14 L. J. Yan, M. T. Droy-Lefaix, and L. Packer, Biochem. Biophys. Res. Commun. 212, 360 (1995). 15 Z. Wei, Q. Peng, B. H. Lau, and V. Shah, Gen. Pharmacol. 33, 369 (•999). 16 y. j. Suzuki, M. Mizuno, and L. Packer, J. Immunol. 153, 5008 (1994).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
393
FIG. 3. DNA-binding activity of NF-KB in HUVEC cocultured for 24 hr with nonstimulated or IFN-'y-stimulated RAW 264.7 macrophages. One representative experiment, out of three, is presented. Lane 1, nonstimulated HUVEC cells; lane 2, HUVEC cells stimulated by 10 ng/ml TNF-c~;lane 3, HUVEC. exposed to the medium used to culture RAW 264.7 cells and containing 10 units/ml murine IFN-',/; lane 4, HUVEC cells after exposure to RAW 264.7 macrophages stimulated by 10 units/ml IFN-~/; and lane 5, HUVEC cells after exposure to nonactivated RAW 264.7 macrophages. overlaps with u n p u b l i s h e d experiments conducted on the endothelial-like cell line E C V 304, suggesting that the immortalized cell line E C V 304 displays a similar pattern of activation of NF-KB than authentic primary endothelial cells, H U V E C .
Coincubation with Macrophages to Induce NF-KB Transactivation Activity in HUVEC: Effect of Preincubation with Flavonoid-Rich Plant Extract (EGb 761) E M S A provides information about activation of the NF-KB pathway. However, this assay is not suitable to understand whether transactivation activity occurs subsequently to the b i n d i n g of the transcription factor to the consensus sequence of target genes. A n expedient tool to study NF-KB transactivation is a reporter gene assay based on the simultaneous transfection of H U V E C with two plasmids containing either firely or Renilla luciferase hooked up to a multiple NF-KB consensus sequence and to a simple t h y m i d i n e kinase promoter, respectively. The details of the m e t h o d are given elsewhere. 17 17C. Saliou, M. Kitazawa, L. McLaughin, J. E Yang, J. K. Lodge, T. Tetsuka, K. lwasaki, J. CiUard, T. Okamoto, and L. Packer, FreeRadic. Biol. Med. 26, 174 (1999).
394
BIOLOGICAL ACTIVITY
~"
['---] no macrophages ~ nonaetivated macrophages ~ activated macrophages
.~ a
~
[351
activated macrophages + EGb 761
.4
=1
o "9
i' ~2
0 FIG. 4. NF-KB transactivation activity in HUVEC coincubated with RAW 264.7 macrophages treated with 10 U/ml IFN-~. Twelve hours before the exposure to macrophages, HUVEC cells are transfected transiently with two plasmids containing either firefly or Renilla luciferases hooked up to a multiple NF-KB consensus sequence and to a simple thymidine kinase promoter, respectively. Twelve hours after coincubation, HUVEC cells are collected and both Renilla and firefly luciferase activities are measured in cell lysate. The effect of the preincubation with Ginkgo biloba extract EGb 761 is also shown. Results are expressed as the ratio of Renilla to firefly luciferase activity in the cell lysate.
The coincubation of HUVEC with nonstimulated macrophages is associated with a slight but significant increase in the ratio between firefly to Renilla luciferase activity in comparison with control endothelial cells (Fig. 4). In agreement with results obtained by EMSA, the observed increase suggests that some transactivation and NF-KB-driven gene expression occur in the presence of nonstimulated macrophages. As suggested earlier, this effect can be due to the presence of a small but significant amount of active macrophages in the population of RAW 264.7, which generate both ROS and RNS and possibly other molecules triggering the cellular response mediated by NF-KB signaling. After stimulation with IFN-~/, RAW 264.7 are able to induce a much stronger expression of NF-KB transactivation, as measured by luciferase activity, with respect to the basal value. Thus, the exposure of HUVEC to stimulated macrophages
[35]
INTERACTIONBETWEENENDOTHELIALCELLSAND MACROPHAGES
395
is associated with an increase of NF-KB transactivating activity, which eventually leads to the expression of specific response genes. The preincubation with EGb 761 significantly affects the expression of NF-KB-dependent luciferase gene in HUVEC exposed to activated macrophages, whereas no effect is observed both in control cells and in cells exposed to quiescent macrophages. Thus, our observations indicate that the interaction of activated macrophages with endothelial cells is associated with both the transfer of NF-KB to the nucleus and to the activation of its transactivating activity. Coincubation with Macrophages to Induce Expression of MCP-1 in HUVEC Cells: Effect of Preincubation with Flavonoid-Rich Plant Extract (EGb 761)
Different genes have been reported to be expressed in response to proinflammatory stimuli partially depending on the activation of the NF-KB transcription factor. Among these genes, MCP- 1 plays a central role in encoding for a secretory protein, which further recruits monocytes in the inflammatory environment.18 This recruitment is followed by a complex process of recognition, adhesion, diapedesis, and activation, which eventually leads to the infiltration of macrophages in the subluminal area.19 If this process is not perfectly controlled and regulated, a vicious circle may be triggered and a chronic inflammation may be established. MCP-1 has different NF-KB sites in the promoter, along with other transcription factors; NF-KB is one of the major regulatory units for the expression of MCP-l.2° We have therefore assayed the coincubation of macrophages with endothelial cells to study its impact on the expression of the mRNA encoding for the MCP-1 protein. In order to investigate this issue, HUVEC cells are coineubated with either quiescent macrophages or macrophages activated by IFN-~/. After 6 hr of coincubation, total RNA is isolated to study the expression of MCP-I-mRNA by a reverse transcription polymerase chain reaction (RT-PCR). Some of the methodological details of this approach are described elsewhere. 21 Figure 5 shows that HUVEC express a moderate although detectable level of the MCP-1 gene under basal conditions (lane 1). Preincubation with 100 and 200 Ixg EGb 761 (lanes 2 and 3, Fig. 5) slightly affected the expression of the MCP-1 gene. Following 6 hr of exposure to RAW 264.7 macrophages activated by IFN-~/, MCP-1 expression increase significantly (lane 4, Fig. 5). Interestingly, the preincubation with EGb 761 reduces the expression of the MCP-1 gene at both concentrations tested, but not in a dose-dependent fashion (lanes 5 and 6, Fig. 5;). 18j. M. Wang,W. P. Shen, and S. B. Su, TrendsCardiovas. Med. 8, 169 (1998). 19j. p. Cai, S. Hudson, M. W. Ye, and Y. H. Chin, Celllmmunol. 167, 269 (1996). 20U. Widmer, K. R. Manogue, A. Cerami, and B. Sherry,J. lrnmunol. 150, 4996 (1993). 2mF. Virgili, H. Kobuchi,and L. Packer,FreeRadic. Biol. Med. 24, 1120 (1998).
396
BIOLOGICALACTIVITY
[35]
FIG. 5. MCP-1 gene expression in HUVEC cells coincubated with either dormant RAW 264.7 macrophages or those treated with 10 U/ml IFN-y. After 6 hr of coincubation, HUVEC are collected and total RNA is isolated to study the expression ofMCP- 1-mRNA using a RT-PCR approach. The effect of preincubation with EGb 761 is also shown. Lane 1, control HUVEC; lanes 2 and 3, preincubation with 100 and 200 izg EGb 761, respectively; lane 4, following 6 hr of exposure to RAW 264.7 macrophages activated by IFN-',/; and lanes 5 and 6, same as lane 4 but in the presence of 100 and 200 I~g/ml EGb 761, respectively. Results are shown as the gel photograph of PCR-amplified cDNAs from MCP- 1 and G3PDH mRNAs.
Conclusions Artheriosclerosis is a complex multifactorial disease in the course of which a series of critical events occurs. This includes endothelial dysfunction, infiltration of inflammatory cells into the vessel wall, alteration of the vascular cell phenotype, and vascular remodeling. Before the manifestation of typical dramatic outcomes, such as stenosis or occlusive thrombosis of the arterial lumen, a number of subtle dysfunctions are displayed at cellular and molecular levels at the beginning of disease progression. These events lead to the loss of homeostatic functions of the endothelium and include the modification of the pattem of gene expression, cell proliferation, and apoptosis. 22 Macrophages are the principal inflammatory cell type in the artheriosclerotic microenvironment. In the early stages of atheroma formation, macrophages 22 G. H. Gibbons and V. J. Dzau, Science 272, 689 (1996).
[35]
INTERACTIONBETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
397
interplay with endothelial cells. A shift from normal homeostasis possibly triggers a vicious circle, which finally leads to endothelial dysfunction. 23'24 Macrophage activity displays the features o f a proinflammatory condition, including the generation of reactive oxygen and nitrogen species. These events are likely to induce a condition of oxidative-nitrosative stress. Indeed, oxidative stress induced injury has been observed in the environment of artheriosclerotic plaque, suggesting that free radicals may be important in the etiology of disease. 22 Also, the oxidation of lowdensity lipoprotein has been described to play a central role in the development of the atheriosclerotic l e s i o n Y Epidemiological surveys strongly suggest a negative association between the dietary consumption of antioxidants such as polyphenols and the incidence of cardiovascular disease, including atheriosclerosis. 26 Experimental studies have given ground to this evidence. 27'28 In fact, polyphenols are attracting growing interest due to their strong antioxident capacity. Phenolic acids and flavonoids have been demonstrated to participate in the cellular antioxidant network and to protect cells from injury induced by both oxygen and nitrogen reactive species. 29-31 The coculture system described herein provides an expedient experimental tool to study the complex interaction between macrophages and endothelial cells and to better understand the molecular basis of the beneficial role of antioxidant flavonoids in the protection from the risk of artheriosclerotic disease.
23j. L. Witztum and D. Steinberg, J. Clin. Invest. 88, 1785 (1991). 24j. m. Berliner, M. Navab, A. M. Fogelman, J. S. Frank, L. L. Demer, P. A. Edwards, A. D. Watson, and A. J. Lusis, Circulation 91, 2488 (1995). 25I. Jialal and S. Devaraj, J. Nutr. 126, S1053 (1996). 26 A. S. St Leger, A. L. Cochrane, and F. Moore, Lancet 8124, 1017 (1979). 27 C. H. Hennekens, Circulation 97, 1095 (1998). 28 S. Hercberg, P. Preziosi, P. Galan, H. Faure, J. Arnaud, N. Duport, D. Malvy, A. M. Roussel, S. Briancon, and A. Favier, Food Chem. ToxicoL 37, 925 (1999). 29C. A. Rice-Evans, N. J. Miller, P. G. Bolwell, P. M. Bramley, and J. B. Pridham, Free Radic. Rec. 22, 375 (1995). 3oj. A. Vinson and Y. A. Dabbagh, FEBS Lett. 433, 44 (1998). 31 W. Bors and C. Michel, Free Radic. Biol. Med. 27, 1413 (1999).
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
387
Synopsis This article focused on two methods to measure the activity of NF-KB. Both methods evaluate "post-IKB pbosphorylation" stages in the NF-KB activation cascade. In fact, EMSA performed with nuclear extracts provides an information only on NF-KB nuclear translocation and its ability to bind KB-DNA sequences. Likewise, the reporter gene assay is limited to assessing NF-KB-dependent gene expression no matter the mechanism that originally activated NF-KB. Nevertheless, the latter assay represents a more physiological and more reproducible way of measuring NF-KB activity in mammalian cells than the EMSA does. In order to obtain further insights into NF-KB signal transduction pathways, investigating IKB degradation and phosphorylation are recommended. The cloning and characterization of IKB kinases provided new testing possibilities based on the measure of their activity.
[35] Interaction between Cultured Endothelial Cells and Macrophages: In Vitro Model for Studying Flavonoids in Redox-Dependent Gene Expression By GERALD
RIMBACH, CLAUDE SALIOU, RAFFAELLA CANALI,
and
FABIO VIRGILI
Introduction A coculture model composed of human endothelial cells and murine macrophages mimicking the proinflammatory environment characterizing atherosclerotic plaque formation is described. Interferon-',/(IFN--,/)-activated macrophages generate reactive oxygen and nitrogen species (ROS and RNS). In turn, ROS and RNS induce the activation of the redox-sensitive transcription factor NF-KB, its transactivation activity, and the expression of monocyte chemotactic protein-1 (MCP-1), a gene containing NF-KB in its promoter side. We utilized this coculture model to study the role of nutritional antioxidants, such as polyphenols, in molecular events associated with the early stages of atheroma development. This article describes the effect of the flavonoid-rich extract from Ginkgo biloba leaves, EGb761, on redox-sensitive gene expression. Results described herein suggest that the model is suitable to assess: (i) the protective effect of flavonoids on oxidative stress induced by reactive oxygen and nitrogen species and (ii) the effect of flavonoids on the activation of signaling pathways leading to the expression of specific response genes, such as MCP- 1.
METHODS IN ENZYMOLOGY,VOL.335
Copyright© 2001 by AcademicPress All rightsof reproductionin any formreserved. 0076-6879/00$35.00
388
BIOLOGICALACTIVITY
[351
RAW 264.7 monocyte/macrophages activated by IFN-y
Human unbilical vein endothelial cells HUVEC FIG. 1. Schemeof the coculturemodel: RAW 264.7 cells are grown on a 0.45-1xmfilter. Following activationwith IFN-3,,RAW264.7cells are placed on top ofa ECV/HUVECdish in directcontactwith the culture medium (1.5 ml). After an appropriate coincubationtime, during which underlying cells are challenged by the reactive oxygen and nitrogen species generated by activated macrophages,the filter is removed and specific parameters (transcription factors activation, mRNAexpression, cellular antioxidant levels) are assessed. A Coculture System to Study Role of ROS and RNS in Macrophage-Endothelial Cell I n t e r a c t i o n The coculture model, illustrated in Fig. 1, is composed of primary human endothelial ceils (HUVEC) and the monocyte-macrophage murine cell line RAW 264.7. It is characterized by a long-term, sustained generation of high levels of ROS and RNS, thus mimicking the occurrence of a complex though mild "physiological" oxidative stress. This is confirmed by the significant accumulation of nitrite and nitrate (NO 2 and NO~-) in the medium after cell activation, which is an index of both nitric oxide (NO.) generation and NO* reaction with superoxide (see Fig. 2). This mode of generation of RNS appears somehow different from that induced by the addition of NO- donors such as 3-morpholinosydnonimine-Nethylcarbamide (SIN-1), sodium nitroprusside (SNP), and compounds either contalning or generating reactive oxygen (H202, organic peroxides, transition metals), l In fact, macrophages activated by the bacterial wall component lipopolysaccharide (LPS) or by cytokines such as IFN-~/ express the enzymatic machinery that generates both ROS and RNS. In such a condition, a direct interaction of NO"
1G. Rimbach, E Virgili, Y. C. Park, and L. Packer, Redox Rep. 4, 171 (1999).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
i + activated macrophages
389
+
+t+l++tWt++t+l[t+l+t+Wml +tt+t
D
+ dormant macrophages
NO3. + N O 2NO2_
controls
0
50
100
150
200
FIG. :2. NO~- and NO 3 accumulation in ECV/HUVEC medium (p,M) after a 24-hr exposure to a filter containing RAW 267.4 monocyte-macrophages activated by IFN-~,. NO 2 and NO 3 levels in the medium are quantified using the automated NOx analyzer based on Griess reagent. Nitrate is determined after reduction to nitrite using an A7200 copperized cadmium reduction column and is then quantified by Griess reagent.
with 0 2 ° - is likely to occur, leading to ONOO- formation. 2-6 This reaction can be observed in different cell lines and tissues such as macrophages, neutrophils, and cultured human endothelial cells. 3'70NOO- is a powerful oxidant molecule, which, due to its weak O-O bond strength, decomposes spontaneously to form the hydroxyl radical and nitrogen dioxide. 8 0 N O O - may react directly with many biological targets, such as lipids and sulfhydryl groups. It may also nitrate tyrosine in proteins, thereby forming 3-nitrotyrosine residues. Nitration of tyrosine can result in an inhibition of protein tyrosine phosphorylation and thus inhibit signal transduction. 2,9 We have already utilized this coculture model to demonstrate that the exposure to activated macrophages is associated with a significant decrease of a-tocopherol and glutathione levels in endothelial cells. In the same studies, we demonstrated
2 G. E. Arteel, K. Briviba, and H. Sies, FEBS Lett. 445, 226 (1999). 3 M. C. Carreras, G. A. Pargament, S. D. Catz, J. J. Poderoso, and A. Boveris, FEBS Lett. 341, 65 (1994). 4 G. L. Squadrito and W. A. Pryor, Free Radic. Biol. Med. 25, 392 (1998). 5 S. Moncada, R. M. Palmer, and E. A. Higgs, Pharmacol. Rev. 43, 109 (1991). 6 j. S. Beckman and W. H. Koppenol, Am. J. Physiol. 271, C1424 (1996). 7 N+ W. Kooy and J. A. Royall, Arch. Bioche. Biophys. 310, 352 (1994). 8 W. H. Koppenol, J. J. Moreno, W. A. Pryor, H. Ischiropoulos, and J. S. Beckman, Chem. Res. Toxicol. 5, 834 (1992). 9 H. Ischiropoulos, Arch. Biochem. Biophys. 356, 1 (1998).
390
BIOLOGICALACTIVITY
[35]
that procyanidins extracted from the bark of Pinus maritima protect the cellular antioxidant network from macrophage-induced oxidative challenge.l' 10 S o m e M e t h o d o l o g i c a l R e m a r k s R e g a r d i n g Cell L i n e s In the coculture model reported herein, a combination of two nonhomo-specific cell lines in the experimental design (RAW 264.7, a murine-derived cell line, with either ECV304 or primary endothelial ceils of human origin) was chosen in order to avoid the activation of endothelial cells by the cytokine utilized for macrophage activation. In fact, no changes were observed either on antioxidant levels or other functional cell parameters when ECV304 was exposed to the medium utilized for RAW 264.7 culture after activation with IFN-3,, l,l° confirming that recombinant routine IFN-~/utilized in our studies displays a species-specific activity. However, reports alerting the utilization of ECV304 as an appropriate model for the endothelium have appeared. In fact, several distinct differences between ECV304 and human umbilical vein-derived endothelial cells (HUVEC), as well as a strong genetic similarity between ECV304 and T24/83, a human bladder cancer cell line, have been reported. 11 The authors suggested that even though ECV304 displays some endothelial characteristics and may he a useful tool for the study of receptor pharmacology, it should not be considered equivalent to HUVEC and is therefore an inappropriate cell line to study endothelial cell biology. These observations led us to reconsider our model; we now use primary endothelial cells obtained from collagenase treatment of the human umbilical vein known as HUVEC instead. Therefore, data presented here only refer to the coculture of RAW 264.7 cells together with HUVEC. However it is worth mentioning that some of the features and responses displayed by "authentic" HUVEC are quite similar to those we have already reported for ECV304. This observation allows us to suggest that even taking into account the limitations mentioned earlier, some of the studies performed utilizing ECV304 as a model for endothelial cells should still be considered valid. Cell C u l t u r e Human Umbilical Vein Endothelial Cells
Primary human endothelial cells are obtained from the umbilical cord vein. According to the method of Jaffe and co-workers, 12 the main vein of umbilical cords (kindly provided by the nursery of the University of Roma, "La Sapienza") is 10 E Vipgili, D. Kim, and L. Packer, FEBS Lett. 431,315 (1998). 11 j. Brown, S. J. Reading, S. Jones, C. J. Fitehett, J. Howl, A. Martin, C. L. Longland, E Miehelangeli, Y. E. Dubrova, and C. A. Brown, Lab. Invest. ~ , 37 (2000). 12 E. Jaffe, R. L. Nachmann, C. G. Beeker, and C. R. Minick, J. Clin. Invest. 153, 5008 (1973).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
391
washed thoroughly with phosphate-buffered saline (PBS) and then treated with 10 ml collagenase (0.2%, dissolved in Medium 199) from Clostridium histiolyticum (type XI, Sigma St. Louis, MO) for 20 min at 37 °. The endothelial cell layer is removed from the vein using 50 ml of Hanks' buffer (Sigma). Cells are then pelleted by centrifugation and cultured in six-well culture dishes (3.5 cm diameter, Falcon, Franklin Lakes, NJ) pretreated with 1.5% gelatine (type B from bovine skin, Sigma). Cells are grown utilizing 199 Medium (GIBCO-BRL, Rockville, MD) supplemented with 20% fetal calf serum (FCS, University of California cell culture facilities, San Francisco, CA), 2 mM glutamine, 1 mM sodium pyruvate, 10 mM HEPES, 100 ~g/ml heparin, streptomycin-penicillin, and 50 ~g/ml endothelial cell growth supplement (ECGS, Sigma). HUVEC cells are utilized for experiments at 90-100% apparent confluence and within passage 3 through 6. Passages are performed according to standardized protocols and by diluting the cell population 1 to 3. RAW 264.7 Murine Monocyte-Macrophage RAW 264.7, a murine cell line of monocyte-macrophages, are obtained from the American Type Culture Collection (ATCC, Rockville, MD). Cells are seeded on 0.45-1~m pore-size, polyethylene terephthalate cell culture inserts (Falcon) and cultured in Dulbecco's modified Eagle's medium (DMEM, GIBCO-BRL) containing 10% FCS, 2 mM glutamine, 1 mM sodium pyruvate, and streptomycinpenicillin. RAW 264.7 are stimulated with 10 U/ml IFN-y (Genzyme, Cambridge, MA). Three hours after stimulation, the filters containing macrophages are placed on top of HUVEC cells in direct contact with the culture medium. Cells are then cocultured at 37 ° and, after an appropriate coincubation time, the filter is removed and specific parameters (transcription factors activation, mRNA expression) are assessed on HUVEC cells. Ginkgo biloba E x t r a c t EGb 761 EGb 761 is a standardized extract of G. biloba leaves. It is used commonly in Asia, Europe, and the United States to treat a variety of pathological conditions, such as peripheral arterial diseases and organic brain syndromes. EGb 761 contains 24% flavonoids (ginkgo-flavon glycosides) and 6% terpenois (ginkgolides, bilobalides) as active components. The flavonoid fraction is composed mainly of the flavonols quercetin, kaempherol, isorhamnetpferol, and isorhamnetin, which are linked to a sugar moiety. 13 This plant extract has been shown to have strong scavenging activity toward hydroxyl radicals, superoxide anion radicals, and NO'. EGb 761 may also contribute significantly to LDL protection from oxidation in 13 O. Sticher, Planta Med. 59, 2 (1993).
392
BIOLOGICAL ACTIVITY
[35]
vitro and to a lesser extent in v i v o . 14 Furthermore, it has been demonstrated in our laboratory and by others that EGb 761 inhibits the activation of NF-KB in endothelial cells and macrophages challenged by different stimuli. 15 All these features render EGb 761 potentially able to contribute to the protection from the development of artheriosclerotic lesion both by acting as a bona fide antioxidant and by modulating the signaling pathway that leads to an endothelial cell response to prooxidation and dysfunction. This article reports on some observation obtained utilizing the macrophageendothelial cell coculture model to study the effect of EGb 761 in the modulation of the endothelial cell response to the exposure to activated macrophages. In the experiments described herein, EGb 761 is dissolved in dimethyl sulfoxide (DMSO) and added to HUVEC culture medium at the indicated concentrations. Cells are supplemented with the extract for 16 hr prior to any further treatment. The final DMSO concentration in the medium is less than 0.1%. Control cells, i.e., those not supplemented with EGb 761, are treated with identical amounts of DMSO. C o c u l t u r i n g M a c r o p h a g e s a n d E n d o t h e l i a l Cells a s S u i t a b l e M o d e l for S t u d y of F l a v o n o i d s in A c t i v a t i o n of NF-KB P a t h w a y I n d u c e d by Oxidative-Nitrosative Stress Coincubation with Activated RAW 264. 7 Monocytes/Macrophages to Induce Transfer of NF-KB into Nucleus and DNA Binding in HUVEC Figure 3 shows the electromobility shift assay (EMSA) performed according to Suzuki and co-workers 16 on the nuclear extract of HUVEC cells after exposure to RAW 264.7 monocytes/macrophages either quiescent or activated by treatment with 10 U/ml murine IFN-~/. In the basal condition (lane 1, Fig. 3), HUVEC cells do not show significant DNA-binding activity, which is induced strongly by TNFot (lane 2, Fig. 3). Similarly, following the exposure to the medium utilized to culture RAW 264.7 macrophages, transferred to the top of HUVEC cells 3 hr after the addition of murine IFN-~/, no significant changes are observed in the level of NF-KB binding to the specific DNA consensus sequence (lane 3, Fig. 3). After the exposure to nontreated RAW 264.7 cells, the level of NF-KB-binding activity to DNA increases slightly (lane 5, Fig. 3). The increase is more dramatic after the coincubation with macrophages activated by IFN-~/ (lane 4, Fig. 3). This result indicates that activated macrophages significantly induce the activation of the signaling pathway leading to NF-KB transfer to the nucleus and possibly the expression of downstream genes in HUVEC cells. It is noteworthy that this observation
14 L. J. Yan, M. T. Droy-Lefaix, and L. Packer, Biochem. Biophys. Res. Commun. 212, 360 (1995). 15 Z. Wei, Q. Peng, B. H. Lau, and V. Shah, Gen. Pharmacol. 33, 369 (•999). 16 y. j. Suzuki, M. Mizuno, and L. Packer, J. Immunol. 153, 5008 (1994).
[35]
INTERACTION BETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
393
FIG. 3. DNA-binding activity of NF-KB in HUVEC cocultured for 24 hr with nonstimulated or IFN-'y-stimulated RAW 264.7 macrophages. One representative experiment, out of three, is presented. Lane 1, nonstimulated HUVEC cells; lane 2, HUVEC cells stimulated by 10 ng/ml TNF-c~;lane 3, HUVEC. exposed to the medium used to culture RAW 264.7 cells and containing 10 units/ml murine IFN-',/; lane 4, HUVEC cells after exposure to RAW 264.7 macrophages stimulated by 10 units/ml IFN-~/; and lane 5, HUVEC cells after exposure to nonactivated RAW 264.7 macrophages. overlaps with u n p u b l i s h e d experiments conducted on the endothelial-like cell line E C V 304, suggesting that the immortalized cell line E C V 304 displays a similar pattern of activation of NF-KB than authentic primary endothelial cells, H U V E C .
Coincubation with Macrophages to Induce NF-KB Transactivation Activity in HUVEC: Effect of Preincubation with Flavonoid-Rich Plant Extract (EGb 761) E M S A provides information about activation of the NF-KB pathway. However, this assay is not suitable to understand whether transactivation activity occurs subsequently to the b i n d i n g of the transcription factor to the consensus sequence of target genes. A n expedient tool to study NF-KB transactivation is a reporter gene assay based on the simultaneous transfection of H U V E C with two plasmids containing either firely or Renilla luciferase hooked up to a multiple NF-KB consensus sequence and to a simple t h y m i d i n e kinase promoter, respectively. The details of the m e t h o d are given elsewhere. 17 17C. Saliou, M. Kitazawa, L. McLaughin, J. E Yang, J. K. Lodge, T. Tetsuka, K. lwasaki, J. CiUard, T. Okamoto, and L. Packer, FreeRadic. Biol. Med. 26, 174 (1999).
394
BIOLOGICAL ACTIVITY
~"
['---] no macrophages ~ nonaetivated macrophages ~ activated macrophages
.~ a
~
[351
activated macrophages + EGb 761
.4
=1
o "9
i' ~2
0 FIG. 4. NF-KB transactivation activity in HUVEC coincubated with RAW 264.7 macrophages treated with 10 U/ml IFN-~. Twelve hours before the exposure to macrophages, HUVEC cells are transfected transiently with two plasmids containing either firefly or Renilla luciferases hooked up to a multiple NF-KB consensus sequence and to a simple thymidine kinase promoter, respectively. Twelve hours after coincubation, HUVEC cells are collected and both Renilla and firefly luciferase activities are measured in cell lysate. The effect of the preincubation with Ginkgo biloba extract EGb 761 is also shown. Results are expressed as the ratio of Renilla to firefly luciferase activity in the cell lysate.
The coincubation of HUVEC with nonstimulated macrophages is associated with a slight but significant increase in the ratio between firefly to Renilla luciferase activity in comparison with control endothelial cells (Fig. 4). In agreement with results obtained by EMSA, the observed increase suggests that some transactivation and NF-KB-driven gene expression occur in the presence of nonstimulated macrophages. As suggested earlier, this effect can be due to the presence of a small but significant amount of active macrophages in the population of RAW 264.7, which generate both ROS and RNS and possibly other molecules triggering the cellular response mediated by NF-KB signaling. After stimulation with IFN-~/, RAW 264.7 are able to induce a much stronger expression of NF-KB transactivation, as measured by luciferase activity, with respect to the basal value. Thus, the exposure of HUVEC to stimulated macrophages
[35]
INTERACTIONBETWEENENDOTHELIALCELLSAND MACROPHAGES
395
is associated with an increase of NF-KB transactivating activity, which eventually leads to the expression of specific response genes. The preincubation with EGb 761 significantly affects the expression of NF-KB-dependent luciferase gene in HUVEC exposed to activated macrophages, whereas no effect is observed both in control cells and in cells exposed to quiescent macrophages. Thus, our observations indicate that the interaction of activated macrophages with endothelial cells is associated with both the transfer of NF-KB to the nucleus and to the activation of its transactivating activity. Coincubation with Macrophages to Induce Expression of MCP-1 in HUVEC Cells: Effect of Preincubation with Flavonoid-Rich Plant Extract (EGb 761)
Different genes have been reported to be expressed in response to proinflammatory stimuli partially depending on the activation of the NF-KB transcription factor. Among these genes, MCP- 1 plays a central role in encoding for a secretory protein, which further recruits monocytes in the inflammatory environment.18 This recruitment is followed by a complex process of recognition, adhesion, diapedesis, and activation, which eventually leads to the infiltration of macrophages in the subluminal area.19 If this process is not perfectly controlled and regulated, a vicious circle may be triggered and a chronic inflammation may be established. MCP-1 has different NF-KB sites in the promoter, along with other transcription factors; NF-KB is one of the major regulatory units for the expression of MCP-l.2° We have therefore assayed the coincubation of macrophages with endothelial cells to study its impact on the expression of the mRNA encoding for the MCP-1 protein. In order to investigate this issue, HUVEC cells are coineubated with either quiescent macrophages or macrophages activated by IFN-~/. After 6 hr of coincubation, total RNA is isolated to study the expression of MCP-I-mRNA by a reverse transcription polymerase chain reaction (RT-PCR). Some of the methodological details of this approach are described elsewhere. 21 Figure 5 shows that HUVEC express a moderate although detectable level of the MCP-1 gene under basal conditions (lane 1). Preincubation with 100 and 200 Ixg EGb 761 (lanes 2 and 3, Fig. 5) slightly affected the expression of the MCP-1 gene. Following 6 hr of exposure to RAW 264.7 macrophages activated by IFN-~/, MCP-1 expression increase significantly (lane 4, Fig. 5). Interestingly, the preincubation with EGb 761 reduces the expression of the MCP-1 gene at both concentrations tested, but not in a dose-dependent fashion (lanes 5 and 6, Fig. 5;). 18j. M. Wang,W. P. Shen, and S. B. Su, TrendsCardiovas. Med. 8, 169 (1998). 19j. p. Cai, S. Hudson, M. W. Ye, and Y. H. Chin, Celllmmunol. 167, 269 (1996). 20U. Widmer, K. R. Manogue, A. Cerami, and B. Sherry,J. lrnmunol. 150, 4996 (1993). 2mF. Virgili, H. Kobuchi,and L. Packer,FreeRadic. Biol. Med. 24, 1120 (1998).
396
BIOLOGICALACTIVITY
[35]
FIG. 5. MCP-1 gene expression in HUVEC cells coincubated with either dormant RAW 264.7 macrophages or those treated with 10 U/ml IFN-y. After 6 hr of coincubation, HUVEC are collected and total RNA is isolated to study the expression ofMCP- 1-mRNA using a RT-PCR approach. The effect of preincubation with EGb 761 is also shown. Lane 1, control HUVEC; lanes 2 and 3, preincubation with 100 and 200 izg EGb 761, respectively; lane 4, following 6 hr of exposure to RAW 264.7 macrophages activated by IFN-',/; and lanes 5 and 6, same as lane 4 but in the presence of 100 and 200 I~g/ml EGb 761, respectively. Results are shown as the gel photograph of PCR-amplified cDNAs from MCP- 1 and G3PDH mRNAs.
Conclusions Artheriosclerosis is a complex multifactorial disease in the course of which a series of critical events occurs. This includes endothelial dysfunction, infiltration of inflammatory cells into the vessel wall, alteration of the vascular cell phenotype, and vascular remodeling. Before the manifestation of typical dramatic outcomes, such as stenosis or occlusive thrombosis of the arterial lumen, a number of subtle dysfunctions are displayed at cellular and molecular levels at the beginning of disease progression. These events lead to the loss of homeostatic functions of the endothelium and include the modification of the pattem of gene expression, cell proliferation, and apoptosis. 22 Macrophages are the principal inflammatory cell type in the artheriosclerotic microenvironment. In the early stages of atheroma formation, macrophages 22 G. H. Gibbons and V. J. Dzau, Science 272, 689 (1996).
[35]
INTERACTIONBETWEEN ENDOTHELIAL CELLS AND MACROPHAGES
397
interplay with endothelial cells. A shift from normal homeostasis possibly triggers a vicious circle, which finally leads to endothelial dysfunction. 23'24 Macrophage activity displays the features o f a proinflammatory condition, including the generation of reactive oxygen and nitrogen species. These events are likely to induce a condition of oxidative-nitrosative stress. Indeed, oxidative stress induced injury has been observed in the environment of artheriosclerotic plaque, suggesting that free radicals may be important in the etiology of disease. 22 Also, the oxidation of lowdensity lipoprotein has been described to play a central role in the development of the atheriosclerotic l e s i o n Y Epidemiological surveys strongly suggest a negative association between the dietary consumption of antioxidants such as polyphenols and the incidence of cardiovascular disease, including atheriosclerosis. 26 Experimental studies have given ground to this evidence. 27'28 In fact, polyphenols are attracting growing interest due to their strong antioxident capacity. Phenolic acids and flavonoids have been demonstrated to participate in the cellular antioxidant network and to protect cells from injury induced by both oxygen and nitrogen reactive species. 29-31 The coculture system described herein provides an expedient experimental tool to study the complex interaction between macrophages and endothelial cells and to better understand the molecular basis of the beneficial role of antioxidant flavonoids in the protection from the risk of artheriosclerotic disease.
23j. L. Witztum and D. Steinberg, J. Clin. Invest. 88, 1785 (1991). 24j. m. Berliner, M. Navab, A. M. Fogelman, J. S. Frank, L. L. Demer, P. A. Edwards, A. D. Watson, and A. J. Lusis, Circulation 91, 2488 (1995). 25I. Jialal and S. Devaraj, J. Nutr. 126, S1053 (1996). 26 A. S. St Leger, A. L. Cochrane, and F. Moore, Lancet 8124, 1017 (1979). 27 C. H. Hennekens, Circulation 97, 1095 (1998). 28 S. Hercberg, P. Preziosi, P. Galan, H. Faure, J. Arnaud, N. Duport, D. Malvy, A. M. Roussel, S. Briancon, and A. Favier, Food Chem. ToxicoL 37, 925 (1999). 29C. A. Rice-Evans, N. J. Miller, P. G. Bolwell, P. M. Bramley, and J. B. Pridham, Free Radic. Rec. 22, 375 (1995). 3oj. A. Vinson and Y. A. Dabbagh, FEBS Lett. 433, 44 (1998). 31 W. Bors and C. Michel, Free Radic. Biol. Med. 27, 1413 (1999).