- Email: [email protected]
Arachidonic acid metabolism as a reporter of skin irritancy and target of cancer chemoprevention
Toxicology Letters 96,97 (1998) 111 – 118
Arachidonic acid metabolism as a reporter of skin irritancy and target of cancer chemoprevention F. Marks *, G. Fu¨rstenberger, K. Mu¨ller-Decker Deutsches Krebsforschungszentrum (German Cancer Research Center), Forschungsschwerpunkt Tumorzellregulation, D-69120 -Heidelberg, Germany
Abstract Keratinocytes respond to skin irritation and injury by cytokine release and a rapid but transient activation of arachidonic acid metabolism along both the cyclooxygenase and lipoxygenase pathways. In the first part of this article results are reviewed indicating that the release of pro-inflammatory mediators such as eicosanoids and interleukin-1 from keratinocytes provides a suitable in vitro parameter of irritancy. Based on this response an assay system has been established which may partially replace animal tests such as the Draize test. A permanent overactivation of arachidonic acid metabolism appears to be a driving force of tumor development in both experimental animals and man. Inhibition of the enzymes involved (such as cyclooxygenases by nonsteroidal antiinflammatory drugs) provides, therefore, a powerful and promising measure of cancer chemoprevention. The state of the art in this rapidly developing field is briefly reviewed in the second part of this article. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Skin irritancy; Eicosanoids; Interleukin-1; Nonsteroidal antiinflammatory drugs; Cancer; Cyclooxygenase
1. Introduction A major function of the skin is to protect the body from environmental hazards. This function is accomplished by both terminal differentiation of epidermal keratinocytes into a resistant and impermeable horny layer and a very rapid and characteristic tissue response to damage and irritation. This response, called hyperplastic transfor-
* Corresponding author.
mation (Marks and Fu¨rstenberger, 1993), includes defence reactions becoming apparent as inflammation, as well as protective and repair processes such as development of epidermal hyperplasia and wound healing. Since skin inflammation and epidermal hyperplasia in general constitute a combined response it may be proposed that both originate from a common molecular trigger mechanism despite the fact that the overall reaction results from highly sophisticated interactions between the different compartments and cell types of the skin.
0378-4274/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0378-4274(98)00057-5
112
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
As the most exposed outpost of the skin the keratinocytes of the epidermis have a major control function in that upon irritation and damage they become rapidly activated to produce and release various signaling factors such as cytokines, growth factors, pro-inflammatory mediators etc. (Kupper and Groves, 1995). Some of these factors may act back on keratinocytes thereby inducing cellular proliferation, terminal differentiation, cytokine release etc., others recruit and control other cell types involved in inflammation and tissue repair. Considering its regulatory function, epidermis may be, therefore, looked upon as a highly effective signaling interface between the environment and the body. The cascades of intercellular signaling triggered by activated keratinocytes are probably started off by the release of a few key compounds which are stored by the cells in an inactive form or as precursor molecules. These key factors include interleukin-1a (Il-1a) and arachidonic acid. While Il-1a provides a major regulator of cytokine production (Stadnyk, 1994; Kupper and Groves, 1995), arachidonic acid is rapidly metabolized to a wide variety of highly active compounds, i.e. the eicosanoids such as prostaglandins, thromboxanes, leukotrienes etc. (Fig. 1), which act mainly as short-lived local mediators being involved in the control of cell proliferation and differentiation, progammed cell death, erythema and oedema formation, leukocyte invasion etc. Thus, both Il-1a and arachidonic acid may be consid-
Fig. 1. The major pathways of eicosanoid formation. Upon release from phospholipids (as catalyzed, for instance, by signal-controlled phospholipase A2) arachidonic acid is immediately transformed by various enzymes into a large number of biologically highly active compounds, i.e. the eicosanoids.
ered to represent bona fide pro-inflammatory key mediators. Provided the release of these factors is a general response of keratinocytes to irritation and damage, it would offer an attractive possibility to develop an in vitro test for skin irritants thus aiming at the replacement of animal experiments. Moreover, eicosanoid formation has been shown to be a critical event in tumor formation in skin and perhaps also in other tissues (Fu¨rstenberger et al., 1989; Marnett, 1992). Thus, arachidonic acid metabolism may provide a promising target of cancer chemoprevention.
2. The release from keratinocytes of pro-inflammatory key mediators as an in-vitro parameter of skin irritancy We have investigated the possibility to use the release of both Il-1a and arachidonic acid from cultivated keratinocytes as in vitro parameter of skin irritancy. To circumvent from the beginning the problem of species differences we have chosen the human keratinocyte line HPKII, which can be easily grown in vitro (Mu¨ller-Decker et al., 1992). To measure the release of arachidonic acid the cells were prelabeled for 18 h with [114 C]arachidonic acid. After treatment with the test substances, arachidonic acid and eicosanoids were extracted from the culture medium, separated by thin layer chromatography and quantitated by radiodensitometry. The release of Il-1a was determined by means of an enzyme immuno assay, and cell viability was monitored using the MTT test (Mu¨ller-Decker et al., 1992). Fifteen structurally unrelated chemicals with graded irritant activities were tested and the concentrations required for the induction of the halfmaximal arachidonic acid release (SC50 values) or a 10-fold stimulation of Il-1a release (ED10 values, SC50 values could not be determined because in higher concentrations the test substances interacted with the Il-1a enzyme immunoassay) were calculated (Mu¨ller-Decker et al., 1994). These values correlated sufficiently with in vivo irritancy data derived from the literature, whereas the release kinetics did not show such a correlation. The striking differences in the release kinetics indicate
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
differences in the mechanisms of action of the test compounds which have not yet been elucidated. As compared with the Il-1a response the arachidonic acid response was the more sensitive endpoint for 11 out of 15 substances. Nevertheless, Il-1a release proved to be a valuable parameter, for instance for estimating the activities of SnCl2, ZnCl2, and acrylamide. While these compounds were equipotent in inducing arachidonic acid release they could be clearly distinguished by means of the Il-1a response. No clearcut correlation between arachidonic acid release and cytotoxicity was seen indicating that the liberation of arachidonic acid does not necessarily depend on irreversible cell damage. However, to induce the Il-1a release all test compounds had to be employed in cytotoxic concentrations. Do the in vitro data provide a reliable parameter of skin irritancy in man? On the first glance this question may be given an affirmative answer. A more detailed analysis of the corresponding literature showed, however, that the available data from animal experiments were entirely incoherent and insufficient, let alone the problem of species differences. Therefore, we had to validate our in vitro results by a placebo-controlled, open randomized study using healthy volunteers (Caucasian males, age 18– 45 years). Upon patch application of the test compound in four concentrations onto the skin of the inner forearm by means of a Duhring occlusion chamber (Frosch and Kligman, 1979) erythema development was followed up by visual scoring in order to determine individual erythema scores. In addition, transepidermal water loss (Agner and Serup, 1990) and the accumulation of Il-1a and various eicosanoids in suction blister fluid was determined. The Il-1a concentration was assayed in 50 ml aliquots, eicosanoid analysis was carried out with another 100 ml aliquot of suction blister fluid. The latter was accomplished by gas chromatography/ mass spectrometry using the negative ion ionization mode with [18O]2-labeled eicosanoids as internal standards (Lehmann et al., 1992). In contrast to the cell culture, arachidonic acid released in skin is not bound to serum albumin and thus protected from metabolization. Therefore, ei-
113
cosanoids such as the prostaglandin D2, E2, and F2a, 6-keto-PGF1a, leukotriene B4, and 5-, 12-, and 15-HETE were assayed in addition to arachidonic acid. Despite large interindividual differences, a high correlation between the in vivo and the in vitro results was observed when nine out of 16 compounds were tested, indicating that the in vitro endpoint of arachidonic acid release may indeed be useful to predict the response in humans. In more detail, the in vitro and in vivo comparison showed that compounds which were least active or inactive in the erythema test did not raise substantially eicosanoid levels in blister fluids and they were also the least active in the in vitro approach. On the other hand, compounds which induced erythemas stimulated arachidonic acid metabolism in vivo and belonged to the most active compounds in vitro as far as the SC50 values, the absolute values and the time courses of arachidonic acid release are concerned (Mu¨llerDecker et al., 1992, 1994; Marks et al., 1995; Mu¨ller-Decker et al., 1998a). Thus, arachidonic acid metabolism seems to be generally involved in erythema formation caused by various classes of chemicals, such as detergents, phenol, and alcohols, despite the possibility that different classes of compounds may induce irritation along different pathways.
3. Arachidonic acid metabolism as a target of cancer chemoprevention Disease prevention is accomplished in two ways, i.e. primary prevention aiming at an avoidance or inactivation of environmental biohazards, and secondary prevention trying to interrupt or reverse the process of disease development. Since the latter approach makes use of chemical agents, i.e. drugs, food additives etc., it is also called chemoprevention. Chemopreventive measures are particularly applicable if the disease develops via harmless preliminary stages over a long period of time and if the cellular and molecular mechanisms underlying this development and being the targets proper of chemopreventive agents are known.
114
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Cancer represents a striking example of a longterm pathological process which proceeds via several stages. The disease results from the progressive accumulation of genetic damage (Fearon and Vogelstein, 1990) and is caused by a combined effect of both genotoxic and nongenotoxic agents (Ames et al., 1995). The latter promote cancer development, in that they either facilitate the attack of genotoxic agents or increase the probability of additional genetic damage by stimulating cellullar proliferation (or inhibiting programmed cell death) and inducing the formation of endogenous genotoxic agents (Ohshima and Bartsch, 1994; Ames et al., 1995). The multistage process of cancer development and the synergistic interactions between genotoxic and non-genotoxic carcinogens have been most thoroughly investigated using rat liver and mouse skin as experimental models. In the mouse skin model tumorigenesis is initiated by genotoxic agents and promoted by a repeated induction of a regenerative hyperproliferative response accomplished either by mechanical wounding or chronic application of ‘tumor promoters’ such as the phorbol ester TPA (Marks and Fu¨rstenberger, 1995). TPA does, in fact, provide a most powerful tool for an investigation of the mechanisms of tumor development. Experimentally induced skin cancer development starts with reversible epidermal hyperplasia, followed by the appearance of clonal preneoplastic lesions (reversibly and irreversibly growing papillomas) and ending up with invasive and metastazising carcinomas. While papilloma development requires tumor-promoting treatment the progression of papillomas to carcinomas may occur spontaneously indicating an endogenous production of genotoxic factors (Fu¨rstenberger and Kopp-Schneider, 1995). The latter are generally believed to include active oxygen species, peroxides, free organic radicals, aldehydes etc. which originate, in particular, from oxidative lipid metabolism. In fact, antioxidants and radical scavengers have been found to inhibit skin carcinogenesis, i.e. to act as chemopreventive agents. The pathways of arachidonic acid metabolism are particularly rich sources of such genotoxic factors (Marnett, 1994). In fact, skin tumors ex-
hibit a dramatic accumulation of eicosanoids such as prostaglandins and 8- and 12-HETE and an abberrant expression of the corresponding enzymes, i.e. prostaglandin H synthase-2 (alias cyclooxygenase-2) and 8- and 12-lipoxygenase, indicating a pathological overactivation of arachidonic acid metabolism (Krieg et al., 1995; Mu¨llerDecker et al., 1995). This situation develops in the course of tumor promotion, when arachidonic acid metabolism and the corresponding enzyme expression—as typical responses to skin damage and irritation— are induced by each tumor promoting treatment. While this induction is transient in normal tissue it becomes constitutive in neoplastic lesions (Mu¨ller-Decker et al., 1995; Scholz et al., 1995). It is proposed that due to a ras-mutation or other genetic defects neoplastic epidermal cells become subject to a permanent autocrine stimulation by ‘wound hormones’ such as TGFa which among other effects induce the release of arachidonic acid from phospholipids (Kast et al., 1993) and the de novo synthesis of cyclooxygenase-2 (COX-2) and probably 8- and 12-lipoxygenase (unpublished results) (Fig. 2). The causal relationship between arachidonic acid metabolism and carcinogenesis is impressively demonstrated by the powerful chemopreventive effects of cyclooxygenaseand lipoxygenase inhibitors. When applied together with tumor promoters these agents almost completely prevent epidermal tumor development (Petrusevska et al., 1988; Fu¨rstenberger et al., 1989). As fas as cyclooxygenase inhibitors are concerned this inhibitory effect correlates with COX-2 induction and prostaglandin F2a (PGF2a ) accumulation and is specifically reversed by PGF2a indicating a key role of this factor in tumorigenesis (Fig. 3). The most prominent cyclooxygenase inhibitors are the non-steroidal antiinflammatory drugs (NSAIDs), such as aspirin, indomethacin, sulindac, diclofenac etc. There is now ample evidence that NSAIDs rank among the most powerful chemopreventive agents not only in animal experiments but also for man. The incidence of colorectal cancer, in particular, has been found to be substantially reduced in chronic aspirin users (Kune et al., 1988), and local
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
115
Fig. 2. Two pathways of COX-2 induction in the course of tumor development. In experimental mouse skin carcinogenesis the COX-2 gene is thought to be activated along the Ras-MAP kinase cascade (left). While tumor initiation results from oncogene activation, for instance of H-ras, tumor promotion by phorbol ester TPA probably activates the MAP kinase cascade by stimulating protein kinase C (which, in turn, activates the protein kinase Raf-1 positioned downstream of Ras). Skin tumor promotion by wounding is most probably mediated by TGFa which activates the cascade by interacting with the EGF receptor (EGFR). Since TGFa production is induced along the Ras-MAP kinase cascade, the tumor promoting effect may become autonomous due to autocrine feedback. In colon carcinogenesis, such as observed in FAP patients and Min mice, COX-2 expression has been proposed to be induced along the b-catenin/armadillo pathway of signal transduction (right). This signaling cascade is activated by the Fz transmembrane receptor with the Wnt growth factors as endogenous ligands, and negatively controlled by the APC tumor suppressor protein. The corresponding gene becomes deleted in early stages of colon carcinogenesis resulting in ‘self-promotion’ of tumor development (Kinzler and Vogelstein, 1996; Prescott and White, 1996). The exact mechanism of COX-2 induction along the b-catenin/armadillo pathway is not known yet.
NSAID treatment has been repeatedly shown to interrupt and reverse benign colon tumor development in patients suffering from familial adenomatous polyposis (FAP) (Winde et al., 1993). FAP patients as well as the corresponding animal model, the MIN mouse, carry a non-sense mutation of the APC tumor supperssor gene (Polakis, 1997). There is now strong evidence that this genetic defect results in a constitutive overexpression of cyclooxygenase-2 (Kargman et al., 1995; Prescott and White, 1996). Thus, it appears as if two different mutations, i.e. of H-ras in epidermis and APC in colon epithelium, result in an overactivation of arachidonic acid metabolism which in both tissues provides a critical event in tumor development (Fig. 2). Recently, enzyme inhibitors have been developed which in contrast to the conventional NSAIDs discriminate between the two cyclooxygenase isoenzymes, i.e. the constitutively expressed COX-1 and the inducible COX-2. Both
types of inhibitors suppress experimental mouse skin carcinogenesis indicating both isoenyzmes to be involved (Mu¨ller-Decker et al., 1998b). However, specific COX-2 inhibitors have the advantage to lack the side-effects of conventional NSAIDs such as gastrointestinal bleeding and other stomach problems which are thought to result from COX-1 inhibition (Fro¨hlich, 1997). This makes COX-2 inhibitors particularly attractive for cancer chemoprevention. Lipoxygenase inhibitors are also potential antineoplastic agents since in the mouse skin model unspecific lipoxygenase inhibition has been found to interrupt tumor development in a similar manner as cyclooxygenase inhibition (Petrusevska et al., 1988). Moreover, lipoxygenase-derived arachidonic acid metabolites have been shown to exhibit genotoxic efficacy (Fu¨rstenberger et al., 1990), to inhibit programmed cell death (Tang et al., 1996) and to facilitate metastasis in experimental systems (Honn et al., 1994). Unfortunately, in-
116
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Fig. 3. Inhibition of mouse skin tumor promotion by indomethacin. A three-stage tumorigenesis experiment with dimethylbenz[a]anthracene as an initiator, phorbol ester TPA as a first-stage, and phorbol ester RPA as a second-stage promoter was carried out and the tumor yield (average number of papillomas per animal, ordinate) was determined after 18 weeks. This positive control value is represented by the upper broken line labelled ‘no indomethacin’, while the lower broken line shows the negative control (no TPA treatment). The strong curve (black dots) shows the effect on tumor yield of indomethacin (550 nmol) locally applied before, together with or after TPA application as indicated on the abscissa. When the drug was given 4 h after TPA a complete suppression of tumor development was observed. This effect was completely abolished by simultaneous application of PGF2a (28 nmol), but not by PGE2 (columns on the right side; in this experiment indomethacin and prostaglandins were given 3 h after TPA treatment). The maximum of the anti-promoting effect of indomethacin coincided with a biphasic elevation of PGF2a (thin curve) and an expression of prostaglandin H-synthase 2 (PGHS-2 alias COX-2) in TPA-treated epidermis. For experimental details see Fu¨rstenberger et al. (1989) and Scholz et al. (1995).
hibitors specific for those lipoxygenases which are thought to be relevant for tumor development (8and 12-lipoxygenases) are not yet available either for experimental or for clinical applications. An important question is whether the chemopreventive effect of inhibitors of arachidonic acid
metabolism in man is restricted to colorectal carcinogenesis or may find a broader applicability. Indeed, evidence is accumulating from various animal tumor models as well as from analysis of cancer biopsies that cyclooxygenase-2 may also play a role in the development of prostate, breast,
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
117
Fig. 4. Expression of cyclooxygenases (COX) and lipoxygenases (LOX) in normal and neoplastic human epidermis. The photographs show Western blots obtained upon immunoprecipitation. For this purpose biopsy material (obtained from Professor Gross, Dermatological Clinic of the University of Rostock) was homogenized and the 4000 ×g supernatant was successively incubated with anti-mouse COX-1 antibody (Scholz et al., 1995), anti-human COX-2 antibody (SC 1745, Santa Cruz, Heidelberg, Germany), and anti-human 12-LOX antibody (pLOX, Alexis, Gru¨nberg, Germany). The Western blot was stained using the same antisera. N, normal epidermis; BCC, basal cell carcinoma; KA, keratoacanthoma; SCC, squameous cell carcinoma. As positive controls (M) for the immunoblot procedure microsomal protein from mouse epidermal cells (PDV) was used for the COX-blots, and protein of a 10000× g supernatant from human embryonic kidney cells (HEK-293) transfected with human platelet-type 12-lipoxygenase (pLOX) for the LOX-blot. Note the higher electrophoretic mobility of mouse COX-2 as compared with human COX-2.
lung, and non-melanoma skin cancers. In biopsies from actinic keratoses, the precursor lesions to human squamous cell carcinomas, as well as from squamous cell carcinomas and keratoacanthomas an overexpression of both COX-2 and 12-Lipoxygenase has been found (Fig. 4) (unpublished results). Therefore, COX-2 and LOX inhibitors may be effective for skin cancer prevention not only in experimental animals but also in man, for instance in high-risk populations such as immunosuppressed patients who have undergone organ transplantation.
References Agner, T., Serup, J., 1990. Sodium lauryl sulphate for irritant patch testing: a dose response study using bioengineering methods. J. Invest. Dermatol. 95, 543–547. Ames, B.N., Gold, L.S., Willett, W.C., 1995. The causes and prevention of cancer. Proc. Natl. Acad. Sci. USA 92, 5258 – 5265. Fearon, E.R., Vogelstein, B., 1990. A genetic model for colorectal tumorigenesis. Cell 61, 759–767. Fro¨hlich, J.C., 1997. A classification of NSAIDs according to the relative inhibition of cyclooxygenase isoenzymes. Trends Pharmacol. Sci. 18, 30–34. Frosch, P.J., Kligman, A.M., 1979. The Duhring Chamber test. Contact Dermat. 5, 73–81. Fu¨rstenberger, G., Kopp-Schneider, A., 1995. Malignant progression of papillomas induced by the initiation-promotion protocol in NMRI mouse skin. Carcinogenesis 16, 61–69.
Fu¨rstenberger, G., Gross, M., Marks, F., 1989. Eicosanoids and multistage carcinogenesis in NMRI mouse skin: role of prostaglandin E and F in conversion (first stage of tumor promotion) and promotion (second stage of tumor promotion). Carcinogenesis 10, 91 – 96. Fu¨rstenberger, G., Hagedorn, H., Schurich, B., Petrusevska, B., Fusenig, N.E., Marks, F., 1990. Arachidonic acid metabolites as endogenous mediators of chromosomal alterations and tumor development in mouse skin in vivo. In: Garner, R.C., Hradec, J. (Eds.), Biochem. Chem. Carc. Plenum Press, New York, p. 1990). Honn, K.V., Tang, D.G., Butovich, I.A., Lin, B., Timar, J., Hagmann, W., 1994. 12 – Lipoxygenases and 12(S)-HETE: role in cancer metastasis. Cancer Metastasis Rev. 13, 365 – 396. Kargman, S.L., O’Neill, G.P., Vickers, P.J., Evans, J.F., Mancini, J.A., Jothy, S., 1995. Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer. Cancer Res. 55, 2556 – 2559. Kast, R., Fu¨rstenberger, G., Marks, F., 1993. Activation of cytosolic phospholipase A2 by transforming growth factora in HEL-30 keratinocytes. J. Biol. Chem. 268, 16795 – 16807. Kinzler, K.W., Vogelstein, B., 1996. Lession learned from hereditary colorectal cancer. Cell 87, 159 – 170. Krieg, P., Kinzig, A., Ress-Lo¨schke, M., Vogel, S., VanLandingham, B., Stephan, M., Lehmann, W.D., Marks, F., Fu¨rstenberger, G., 1995. 12-Lipoxygenase isoenzymes in mouse skin tumor development. Mol. Carcinog. 14, 118 – 129. Kune, G.A., Kune, S., Watson, L.F., 1988. Colorectal cancer risk, chronic illnesses, operations and medications: casecontrol results from the Melbourne Colorectal Cancer Study. Cancer Res. 48, 4399 – 4404. Kupper, T.S., Groves, R.W., 1995. The interleukin-1 axis and cutaneous inflammation. J. Investig. Dermatol. 105, 62S – 66.
118
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Lehmann, W.D., Stephan, M., Fu¨rstenberger, G., 1992. Profiling assay for lipoxygenase products of linoleic and arachidonic acid by gas chromatography-mass spectometry. Anal. Biochem. 204, 158–170. Marks, F., Fu¨rstenberger, G., 1993. Proliferative responses of the skin to external stimuli. Environ. Health Perspect. 101, 95– 102. Marks, F., Fu¨rstenberger, G., 1995. Tumor promotion in skin. In: Arcos, J.C., et al. (Eds.), Chemical Induction of Cancer. Birkha¨user, Boston, pp. 123–160. Marks, F., Fu¨rstenberger, G., Heinzelmann, T., Mu¨llerDecker, K., 1995. Mechanisms in tumor promotion: guidance for risk assessment and cancer chemoprevention. Toxicol. Lett. 82/83, 907–917. Marnett, L.J., 1992. Aspirin and the potential role of prostaglandins in colon cancer. Cancer Res. 52, 5575– 5589. Marnett, L.J., 1994. Generation of mutagens during arachidonic acid metabolism. Cancer Metastasis Rev. 13, 303– 308. Mu¨ller-Decker, K., Fu¨rstenberger, G., Marks, F., 1992. Development of an in vitro alternative assay to the Draize skin irritancy test using human keratinocyte-derived proinflammatory key mediators and cell viability as test parameters. In Vitro Toxicol. 5, 191–209. Mu¨ller-Decker, K., Fu¨rstenberger, G., Marks, F., 1994. Keratinocyte-derived proinflammatory key mediators and cell viability as in vitro parametes of irritancy: a possible alternative to the Draize skin irritation test. Toxicol. Appl. Pharmacol. 127, 99 – 208. Mu¨ller-Decker, K., Scholz, K., Marks, F., Fu¨rstenberger, G., 1995. Differential expression of prostaglandin H synthase isoenzymes during multistage carcinogenesis in mouse epidermis. Mol. Carcinog. 12, 31–41. Mu¨ller-Decker, K., Heinzelmann, T., Fu¨rstenberger, G., Kecskes, A., Lehmann, W.D., Marks, F., 1998a. Arachi-
donic acid metabolism in primary irritant dermatitis produced by patch testing of human skin with surfactants. Toxicol. Appl. Pharmacol. (in press). Mu¨ller-Decker, K., Kopp-Schneider, A., Seibert, K., Marks, F., Fu¨rstenberger, G., 1998b. Localization of prostaglandin H synthase isoenzymes in murine epidermal tumors: suppression of skin tumor promotion by PGHS-2 inhibitors. Mol. Carcinog. (in press). Ohshima, H., Bartsch, H., 1994. Chronic infections and inflammatory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis. Mutat. Res. 305, 253 – 264. Petrusevska, R.T., Fu¨rstenberger, G., Marks, F., Fusenig, N.E., 1988. Cytogenetic effects caused by phorbol ester tumor promoters in pimary mouse keratinocyte cultures: correlation with the convertogenic activity of TPA in multistage skin carcinogenesis. Carcinogenesis 9, 1207 – 1215. Polakis, P., 1997. The adenomatous polyposis coli (ADC) tumor suppressor. Biochim. Biophys. Acta 1332, F127 – F147. Prescott, S.M., White, R.L., 1996. Self-Promotion? Intimate connections between APC and prostaglandin H synthase-2. Cell 87, 783 – 786. Scholz, K., Fu¨rstenberger, G., Mu¨ller-Decker, K., Marks, F., 1995. Differential expression of prostaglandin-H synthase isoenzymes in normal and activated keratinocytes in vivo and in vitro. Biochem. J. 309, 263 – 269. Stadnyk, A.W., 1994. Cytokine production by epithelial cells. FASEB J. 8, 1041 – 1047. Tang, D.G., Chen, Y.Q., Honn, K.V., 1996. Arachidonate lipoxygenases as essential regulators of cell survival and apoptosis. Proc. Natl. Acad. Sci. USA 93, 5241 – 5246. Winde, G., Gumbinger, H.G., Osswald, H., Kemper, F., Bunte, H., 1993. The NSAID sulindac reverses rectal adenomas in colectomized patients with familial adenomatous polyposis: Clinical results of a dose-finding study on rectal sulindac administration. Int. J. Colorectal Dis. 8, 13 – 17.
.
Arachidonic acid metabolism as a reporter of skin irritancy and target of cancer chemoprevention F. Marks *, G. Fu¨rstenberger, K. Mu¨ller-Decker Deutsches Krebsforschungszentrum (German Cancer Research Center), Forschungsschwerpunkt Tumorzellregulation, D-69120 -Heidelberg, Germany
Abstract Keratinocytes respond to skin irritation and injury by cytokine release and a rapid but transient activation of arachidonic acid metabolism along both the cyclooxygenase and lipoxygenase pathways. In the first part of this article results are reviewed indicating that the release of pro-inflammatory mediators such as eicosanoids and interleukin-1 from keratinocytes provides a suitable in vitro parameter of irritancy. Based on this response an assay system has been established which may partially replace animal tests such as the Draize test. A permanent overactivation of arachidonic acid metabolism appears to be a driving force of tumor development in both experimental animals and man. Inhibition of the enzymes involved (such as cyclooxygenases by nonsteroidal antiinflammatory drugs) provides, therefore, a powerful and promising measure of cancer chemoprevention. The state of the art in this rapidly developing field is briefly reviewed in the second part of this article. © 1998 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Skin irritancy; Eicosanoids; Interleukin-1; Nonsteroidal antiinflammatory drugs; Cancer; Cyclooxygenase
1. Introduction A major function of the skin is to protect the body from environmental hazards. This function is accomplished by both terminal differentiation of epidermal keratinocytes into a resistant and impermeable horny layer and a very rapid and characteristic tissue response to damage and irritation. This response, called hyperplastic transfor-
* Corresponding author.
mation (Marks and Fu¨rstenberger, 1993), includes defence reactions becoming apparent as inflammation, as well as protective and repair processes such as development of epidermal hyperplasia and wound healing. Since skin inflammation and epidermal hyperplasia in general constitute a combined response it may be proposed that both originate from a common molecular trigger mechanism despite the fact that the overall reaction results from highly sophisticated interactions between the different compartments and cell types of the skin.
0378-4274/98/$19.00 © 1998 Elsevier Science Ireland Ltd. All rights reserved. PII S0378-4274(98)00057-5
112
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
As the most exposed outpost of the skin the keratinocytes of the epidermis have a major control function in that upon irritation and damage they become rapidly activated to produce and release various signaling factors such as cytokines, growth factors, pro-inflammatory mediators etc. (Kupper and Groves, 1995). Some of these factors may act back on keratinocytes thereby inducing cellular proliferation, terminal differentiation, cytokine release etc., others recruit and control other cell types involved in inflammation and tissue repair. Considering its regulatory function, epidermis may be, therefore, looked upon as a highly effective signaling interface between the environment and the body. The cascades of intercellular signaling triggered by activated keratinocytes are probably started off by the release of a few key compounds which are stored by the cells in an inactive form or as precursor molecules. These key factors include interleukin-1a (Il-1a) and arachidonic acid. While Il-1a provides a major regulator of cytokine production (Stadnyk, 1994; Kupper and Groves, 1995), arachidonic acid is rapidly metabolized to a wide variety of highly active compounds, i.e. the eicosanoids such as prostaglandins, thromboxanes, leukotrienes etc. (Fig. 1), which act mainly as short-lived local mediators being involved in the control of cell proliferation and differentiation, progammed cell death, erythema and oedema formation, leukocyte invasion etc. Thus, both Il-1a and arachidonic acid may be consid-
Fig. 1. The major pathways of eicosanoid formation. Upon release from phospholipids (as catalyzed, for instance, by signal-controlled phospholipase A2) arachidonic acid is immediately transformed by various enzymes into a large number of biologically highly active compounds, i.e. the eicosanoids.
ered to represent bona fide pro-inflammatory key mediators. Provided the release of these factors is a general response of keratinocytes to irritation and damage, it would offer an attractive possibility to develop an in vitro test for skin irritants thus aiming at the replacement of animal experiments. Moreover, eicosanoid formation has been shown to be a critical event in tumor formation in skin and perhaps also in other tissues (Fu¨rstenberger et al., 1989; Marnett, 1992). Thus, arachidonic acid metabolism may provide a promising target of cancer chemoprevention.
2. The release from keratinocytes of pro-inflammatory key mediators as an in-vitro parameter of skin irritancy We have investigated the possibility to use the release of both Il-1a and arachidonic acid from cultivated keratinocytes as in vitro parameter of skin irritancy. To circumvent from the beginning the problem of species differences we have chosen the human keratinocyte line HPKII, which can be easily grown in vitro (Mu¨ller-Decker et al., 1992). To measure the release of arachidonic acid the cells were prelabeled for 18 h with [114 C]arachidonic acid. After treatment with the test substances, arachidonic acid and eicosanoids were extracted from the culture medium, separated by thin layer chromatography and quantitated by radiodensitometry. The release of Il-1a was determined by means of an enzyme immuno assay, and cell viability was monitored using the MTT test (Mu¨ller-Decker et al., 1992). Fifteen structurally unrelated chemicals with graded irritant activities were tested and the concentrations required for the induction of the halfmaximal arachidonic acid release (SC50 values) or a 10-fold stimulation of Il-1a release (ED10 values, SC50 values could not be determined because in higher concentrations the test substances interacted with the Il-1a enzyme immunoassay) were calculated (Mu¨ller-Decker et al., 1994). These values correlated sufficiently with in vivo irritancy data derived from the literature, whereas the release kinetics did not show such a correlation. The striking differences in the release kinetics indicate
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
differences in the mechanisms of action of the test compounds which have not yet been elucidated. As compared with the Il-1a response the arachidonic acid response was the more sensitive endpoint for 11 out of 15 substances. Nevertheless, Il-1a release proved to be a valuable parameter, for instance for estimating the activities of SnCl2, ZnCl2, and acrylamide. While these compounds were equipotent in inducing arachidonic acid release they could be clearly distinguished by means of the Il-1a response. No clearcut correlation between arachidonic acid release and cytotoxicity was seen indicating that the liberation of arachidonic acid does not necessarily depend on irreversible cell damage. However, to induce the Il-1a release all test compounds had to be employed in cytotoxic concentrations. Do the in vitro data provide a reliable parameter of skin irritancy in man? On the first glance this question may be given an affirmative answer. A more detailed analysis of the corresponding literature showed, however, that the available data from animal experiments were entirely incoherent and insufficient, let alone the problem of species differences. Therefore, we had to validate our in vitro results by a placebo-controlled, open randomized study using healthy volunteers (Caucasian males, age 18– 45 years). Upon patch application of the test compound in four concentrations onto the skin of the inner forearm by means of a Duhring occlusion chamber (Frosch and Kligman, 1979) erythema development was followed up by visual scoring in order to determine individual erythema scores. In addition, transepidermal water loss (Agner and Serup, 1990) and the accumulation of Il-1a and various eicosanoids in suction blister fluid was determined. The Il-1a concentration was assayed in 50 ml aliquots, eicosanoid analysis was carried out with another 100 ml aliquot of suction blister fluid. The latter was accomplished by gas chromatography/ mass spectrometry using the negative ion ionization mode with [18O]2-labeled eicosanoids as internal standards (Lehmann et al., 1992). In contrast to the cell culture, arachidonic acid released in skin is not bound to serum albumin and thus protected from metabolization. Therefore, ei-
113
cosanoids such as the prostaglandin D2, E2, and F2a, 6-keto-PGF1a, leukotriene B4, and 5-, 12-, and 15-HETE were assayed in addition to arachidonic acid. Despite large interindividual differences, a high correlation between the in vivo and the in vitro results was observed when nine out of 16 compounds were tested, indicating that the in vitro endpoint of arachidonic acid release may indeed be useful to predict the response in humans. In more detail, the in vitro and in vivo comparison showed that compounds which were least active or inactive in the erythema test did not raise substantially eicosanoid levels in blister fluids and they were also the least active in the in vitro approach. On the other hand, compounds which induced erythemas stimulated arachidonic acid metabolism in vivo and belonged to the most active compounds in vitro as far as the SC50 values, the absolute values and the time courses of arachidonic acid release are concerned (Mu¨llerDecker et al., 1992, 1994; Marks et al., 1995; Mu¨ller-Decker et al., 1998a). Thus, arachidonic acid metabolism seems to be generally involved in erythema formation caused by various classes of chemicals, such as detergents, phenol, and alcohols, despite the possibility that different classes of compounds may induce irritation along different pathways.
3. Arachidonic acid metabolism as a target of cancer chemoprevention Disease prevention is accomplished in two ways, i.e. primary prevention aiming at an avoidance or inactivation of environmental biohazards, and secondary prevention trying to interrupt or reverse the process of disease development. Since the latter approach makes use of chemical agents, i.e. drugs, food additives etc., it is also called chemoprevention. Chemopreventive measures are particularly applicable if the disease develops via harmless preliminary stages over a long period of time and if the cellular and molecular mechanisms underlying this development and being the targets proper of chemopreventive agents are known.
114
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Cancer represents a striking example of a longterm pathological process which proceeds via several stages. The disease results from the progressive accumulation of genetic damage (Fearon and Vogelstein, 1990) and is caused by a combined effect of both genotoxic and nongenotoxic agents (Ames et al., 1995). The latter promote cancer development, in that they either facilitate the attack of genotoxic agents or increase the probability of additional genetic damage by stimulating cellullar proliferation (or inhibiting programmed cell death) and inducing the formation of endogenous genotoxic agents (Ohshima and Bartsch, 1994; Ames et al., 1995). The multistage process of cancer development and the synergistic interactions between genotoxic and non-genotoxic carcinogens have been most thoroughly investigated using rat liver and mouse skin as experimental models. In the mouse skin model tumorigenesis is initiated by genotoxic agents and promoted by a repeated induction of a regenerative hyperproliferative response accomplished either by mechanical wounding or chronic application of ‘tumor promoters’ such as the phorbol ester TPA (Marks and Fu¨rstenberger, 1995). TPA does, in fact, provide a most powerful tool for an investigation of the mechanisms of tumor development. Experimentally induced skin cancer development starts with reversible epidermal hyperplasia, followed by the appearance of clonal preneoplastic lesions (reversibly and irreversibly growing papillomas) and ending up with invasive and metastazising carcinomas. While papilloma development requires tumor-promoting treatment the progression of papillomas to carcinomas may occur spontaneously indicating an endogenous production of genotoxic factors (Fu¨rstenberger and Kopp-Schneider, 1995). The latter are generally believed to include active oxygen species, peroxides, free organic radicals, aldehydes etc. which originate, in particular, from oxidative lipid metabolism. In fact, antioxidants and radical scavengers have been found to inhibit skin carcinogenesis, i.e. to act as chemopreventive agents. The pathways of arachidonic acid metabolism are particularly rich sources of such genotoxic factors (Marnett, 1994). In fact, skin tumors ex-
hibit a dramatic accumulation of eicosanoids such as prostaglandins and 8- and 12-HETE and an abberrant expression of the corresponding enzymes, i.e. prostaglandin H synthase-2 (alias cyclooxygenase-2) and 8- and 12-lipoxygenase, indicating a pathological overactivation of arachidonic acid metabolism (Krieg et al., 1995; Mu¨llerDecker et al., 1995). This situation develops in the course of tumor promotion, when arachidonic acid metabolism and the corresponding enzyme expression—as typical responses to skin damage and irritation— are induced by each tumor promoting treatment. While this induction is transient in normal tissue it becomes constitutive in neoplastic lesions (Mu¨ller-Decker et al., 1995; Scholz et al., 1995). It is proposed that due to a ras-mutation or other genetic defects neoplastic epidermal cells become subject to a permanent autocrine stimulation by ‘wound hormones’ such as TGFa which among other effects induce the release of arachidonic acid from phospholipids (Kast et al., 1993) and the de novo synthesis of cyclooxygenase-2 (COX-2) and probably 8- and 12-lipoxygenase (unpublished results) (Fig. 2). The causal relationship between arachidonic acid metabolism and carcinogenesis is impressively demonstrated by the powerful chemopreventive effects of cyclooxygenaseand lipoxygenase inhibitors. When applied together with tumor promoters these agents almost completely prevent epidermal tumor development (Petrusevska et al., 1988; Fu¨rstenberger et al., 1989). As fas as cyclooxygenase inhibitors are concerned this inhibitory effect correlates with COX-2 induction and prostaglandin F2a (PGF2a ) accumulation and is specifically reversed by PGF2a indicating a key role of this factor in tumorigenesis (Fig. 3). The most prominent cyclooxygenase inhibitors are the non-steroidal antiinflammatory drugs (NSAIDs), such as aspirin, indomethacin, sulindac, diclofenac etc. There is now ample evidence that NSAIDs rank among the most powerful chemopreventive agents not only in animal experiments but also for man. The incidence of colorectal cancer, in particular, has been found to be substantially reduced in chronic aspirin users (Kune et al., 1988), and local
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
115
Fig. 2. Two pathways of COX-2 induction in the course of tumor development. In experimental mouse skin carcinogenesis the COX-2 gene is thought to be activated along the Ras-MAP kinase cascade (left). While tumor initiation results from oncogene activation, for instance of H-ras, tumor promotion by phorbol ester TPA probably activates the MAP kinase cascade by stimulating protein kinase C (which, in turn, activates the protein kinase Raf-1 positioned downstream of Ras). Skin tumor promotion by wounding is most probably mediated by TGFa which activates the cascade by interacting with the EGF receptor (EGFR). Since TGFa production is induced along the Ras-MAP kinase cascade, the tumor promoting effect may become autonomous due to autocrine feedback. In colon carcinogenesis, such as observed in FAP patients and Min mice, COX-2 expression has been proposed to be induced along the b-catenin/armadillo pathway of signal transduction (right). This signaling cascade is activated by the Fz transmembrane receptor with the Wnt growth factors as endogenous ligands, and negatively controlled by the APC tumor suppressor protein. The corresponding gene becomes deleted in early stages of colon carcinogenesis resulting in ‘self-promotion’ of tumor development (Kinzler and Vogelstein, 1996; Prescott and White, 1996). The exact mechanism of COX-2 induction along the b-catenin/armadillo pathway is not known yet.
NSAID treatment has been repeatedly shown to interrupt and reverse benign colon tumor development in patients suffering from familial adenomatous polyposis (FAP) (Winde et al., 1993). FAP patients as well as the corresponding animal model, the MIN mouse, carry a non-sense mutation of the APC tumor supperssor gene (Polakis, 1997). There is now strong evidence that this genetic defect results in a constitutive overexpression of cyclooxygenase-2 (Kargman et al., 1995; Prescott and White, 1996). Thus, it appears as if two different mutations, i.e. of H-ras in epidermis and APC in colon epithelium, result in an overactivation of arachidonic acid metabolism which in both tissues provides a critical event in tumor development (Fig. 2). Recently, enzyme inhibitors have been developed which in contrast to the conventional NSAIDs discriminate between the two cyclooxygenase isoenzymes, i.e. the constitutively expressed COX-1 and the inducible COX-2. Both
types of inhibitors suppress experimental mouse skin carcinogenesis indicating both isoenyzmes to be involved (Mu¨ller-Decker et al., 1998b). However, specific COX-2 inhibitors have the advantage to lack the side-effects of conventional NSAIDs such as gastrointestinal bleeding and other stomach problems which are thought to result from COX-1 inhibition (Fro¨hlich, 1997). This makes COX-2 inhibitors particularly attractive for cancer chemoprevention. Lipoxygenase inhibitors are also potential antineoplastic agents since in the mouse skin model unspecific lipoxygenase inhibition has been found to interrupt tumor development in a similar manner as cyclooxygenase inhibition (Petrusevska et al., 1988). Moreover, lipoxygenase-derived arachidonic acid metabolites have been shown to exhibit genotoxic efficacy (Fu¨rstenberger et al., 1990), to inhibit programmed cell death (Tang et al., 1996) and to facilitate metastasis in experimental systems (Honn et al., 1994). Unfortunately, in-
116
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Fig. 3. Inhibition of mouse skin tumor promotion by indomethacin. A three-stage tumorigenesis experiment with dimethylbenz[a]anthracene as an initiator, phorbol ester TPA as a first-stage, and phorbol ester RPA as a second-stage promoter was carried out and the tumor yield (average number of papillomas per animal, ordinate) was determined after 18 weeks. This positive control value is represented by the upper broken line labelled ‘no indomethacin’, while the lower broken line shows the negative control (no TPA treatment). The strong curve (black dots) shows the effect on tumor yield of indomethacin (550 nmol) locally applied before, together with or after TPA application as indicated on the abscissa. When the drug was given 4 h after TPA a complete suppression of tumor development was observed. This effect was completely abolished by simultaneous application of PGF2a (28 nmol), but not by PGE2 (columns on the right side; in this experiment indomethacin and prostaglandins were given 3 h after TPA treatment). The maximum of the anti-promoting effect of indomethacin coincided with a biphasic elevation of PGF2a (thin curve) and an expression of prostaglandin H-synthase 2 (PGHS-2 alias COX-2) in TPA-treated epidermis. For experimental details see Fu¨rstenberger et al. (1989) and Scholz et al. (1995).
hibitors specific for those lipoxygenases which are thought to be relevant for tumor development (8and 12-lipoxygenases) are not yet available either for experimental or for clinical applications. An important question is whether the chemopreventive effect of inhibitors of arachidonic acid
metabolism in man is restricted to colorectal carcinogenesis or may find a broader applicability. Indeed, evidence is accumulating from various animal tumor models as well as from analysis of cancer biopsies that cyclooxygenase-2 may also play a role in the development of prostate, breast,
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
117
Fig. 4. Expression of cyclooxygenases (COX) and lipoxygenases (LOX) in normal and neoplastic human epidermis. The photographs show Western blots obtained upon immunoprecipitation. For this purpose biopsy material (obtained from Professor Gross, Dermatological Clinic of the University of Rostock) was homogenized and the 4000 ×g supernatant was successively incubated with anti-mouse COX-1 antibody (Scholz et al., 1995), anti-human COX-2 antibody (SC 1745, Santa Cruz, Heidelberg, Germany), and anti-human 12-LOX antibody (pLOX, Alexis, Gru¨nberg, Germany). The Western blot was stained using the same antisera. N, normal epidermis; BCC, basal cell carcinoma; KA, keratoacanthoma; SCC, squameous cell carcinoma. As positive controls (M) for the immunoblot procedure microsomal protein from mouse epidermal cells (PDV) was used for the COX-blots, and protein of a 10000× g supernatant from human embryonic kidney cells (HEK-293) transfected with human platelet-type 12-lipoxygenase (pLOX) for the LOX-blot. Note the higher electrophoretic mobility of mouse COX-2 as compared with human COX-2.
lung, and non-melanoma skin cancers. In biopsies from actinic keratoses, the precursor lesions to human squamous cell carcinomas, as well as from squamous cell carcinomas and keratoacanthomas an overexpression of both COX-2 and 12-Lipoxygenase has been found (Fig. 4) (unpublished results). Therefore, COX-2 and LOX inhibitors may be effective for skin cancer prevention not only in experimental animals but also in man, for instance in high-risk populations such as immunosuppressed patients who have undergone organ transplantation.
References Agner, T., Serup, J., 1990. Sodium lauryl sulphate for irritant patch testing: a dose response study using bioengineering methods. J. Invest. Dermatol. 95, 543–547. Ames, B.N., Gold, L.S., Willett, W.C., 1995. The causes and prevention of cancer. Proc. Natl. Acad. Sci. USA 92, 5258 – 5265. Fearon, E.R., Vogelstein, B., 1990. A genetic model for colorectal tumorigenesis. Cell 61, 759–767. Fro¨hlich, J.C., 1997. A classification of NSAIDs according to the relative inhibition of cyclooxygenase isoenzymes. Trends Pharmacol. Sci. 18, 30–34. Frosch, P.J., Kligman, A.M., 1979. The Duhring Chamber test. Contact Dermat. 5, 73–81. Fu¨rstenberger, G., Kopp-Schneider, A., 1995. Malignant progression of papillomas induced by the initiation-promotion protocol in NMRI mouse skin. Carcinogenesis 16, 61–69.
Fu¨rstenberger, G., Gross, M., Marks, F., 1989. Eicosanoids and multistage carcinogenesis in NMRI mouse skin: role of prostaglandin E and F in conversion (first stage of tumor promotion) and promotion (second stage of tumor promotion). Carcinogenesis 10, 91 – 96. Fu¨rstenberger, G., Hagedorn, H., Schurich, B., Petrusevska, B., Fusenig, N.E., Marks, F., 1990. Arachidonic acid metabolites as endogenous mediators of chromosomal alterations and tumor development in mouse skin in vivo. In: Garner, R.C., Hradec, J. (Eds.), Biochem. Chem. Carc. Plenum Press, New York, p. 1990). Honn, K.V., Tang, D.G., Butovich, I.A., Lin, B., Timar, J., Hagmann, W., 1994. 12 – Lipoxygenases and 12(S)-HETE: role in cancer metastasis. Cancer Metastasis Rev. 13, 365 – 396. Kargman, S.L., O’Neill, G.P., Vickers, P.J., Evans, J.F., Mancini, J.A., Jothy, S., 1995. Expression of prostaglandin G/H synthase-1 and -2 protein in human colon cancer. Cancer Res. 55, 2556 – 2559. Kast, R., Fu¨rstenberger, G., Marks, F., 1993. Activation of cytosolic phospholipase A2 by transforming growth factora in HEL-30 keratinocytes. J. Biol. Chem. 268, 16795 – 16807. Kinzler, K.W., Vogelstein, B., 1996. Lession learned from hereditary colorectal cancer. Cell 87, 159 – 170. Krieg, P., Kinzig, A., Ress-Lo¨schke, M., Vogel, S., VanLandingham, B., Stephan, M., Lehmann, W.D., Marks, F., Fu¨rstenberger, G., 1995. 12-Lipoxygenase isoenzymes in mouse skin tumor development. Mol. Carcinog. 14, 118 – 129. Kune, G.A., Kune, S., Watson, L.F., 1988. Colorectal cancer risk, chronic illnesses, operations and medications: casecontrol results from the Melbourne Colorectal Cancer Study. Cancer Res. 48, 4399 – 4404. Kupper, T.S., Groves, R.W., 1995. The interleukin-1 axis and cutaneous inflammation. J. Investig. Dermatol. 105, 62S – 66.
118
F. Marks et al. / Toxicology Letters 96,97 (1998) 111–118
Lehmann, W.D., Stephan, M., Fu¨rstenberger, G., 1992. Profiling assay for lipoxygenase products of linoleic and arachidonic acid by gas chromatography-mass spectometry. Anal. Biochem. 204, 158–170. Marks, F., Fu¨rstenberger, G., 1993. Proliferative responses of the skin to external stimuli. Environ. Health Perspect. 101, 95– 102. Marks, F., Fu¨rstenberger, G., 1995. Tumor promotion in skin. In: Arcos, J.C., et al. (Eds.), Chemical Induction of Cancer. Birkha¨user, Boston, pp. 123–160. Marks, F., Fu¨rstenberger, G., Heinzelmann, T., Mu¨llerDecker, K., 1995. Mechanisms in tumor promotion: guidance for risk assessment and cancer chemoprevention. Toxicol. Lett. 82/83, 907–917. Marnett, L.J., 1992. Aspirin and the potential role of prostaglandins in colon cancer. Cancer Res. 52, 5575– 5589. Marnett, L.J., 1994. Generation of mutagens during arachidonic acid metabolism. Cancer Metastasis Rev. 13, 303– 308. Mu¨ller-Decker, K., Fu¨rstenberger, G., Marks, F., 1992. Development of an in vitro alternative assay to the Draize skin irritancy test using human keratinocyte-derived proinflammatory key mediators and cell viability as test parameters. In Vitro Toxicol. 5, 191–209. Mu¨ller-Decker, K., Fu¨rstenberger, G., Marks, F., 1994. Keratinocyte-derived proinflammatory key mediators and cell viability as in vitro parametes of irritancy: a possible alternative to the Draize skin irritation test. Toxicol. Appl. Pharmacol. 127, 99 – 208. Mu¨ller-Decker, K., Scholz, K., Marks, F., Fu¨rstenberger, G., 1995. Differential expression of prostaglandin H synthase isoenzymes during multistage carcinogenesis in mouse epidermis. Mol. Carcinog. 12, 31–41. Mu¨ller-Decker, K., Heinzelmann, T., Fu¨rstenberger, G., Kecskes, A., Lehmann, W.D., Marks, F., 1998a. Arachi-
donic acid metabolism in primary irritant dermatitis produced by patch testing of human skin with surfactants. Toxicol. Appl. Pharmacol. (in press). Mu¨ller-Decker, K., Kopp-Schneider, A., Seibert, K., Marks, F., Fu¨rstenberger, G., 1998b. Localization of prostaglandin H synthase isoenzymes in murine epidermal tumors: suppression of skin tumor promotion by PGHS-2 inhibitors. Mol. Carcinog. (in press). Ohshima, H., Bartsch, H., 1994. Chronic infections and inflammatory processes as cancer risk factors: possible role of nitric oxide in carcinogenesis. Mutat. Res. 305, 253 – 264. Petrusevska, R.T., Fu¨rstenberger, G., Marks, F., Fusenig, N.E., 1988. Cytogenetic effects caused by phorbol ester tumor promoters in pimary mouse keratinocyte cultures: correlation with the convertogenic activity of TPA in multistage skin carcinogenesis. Carcinogenesis 9, 1207 – 1215. Polakis, P., 1997. The adenomatous polyposis coli (ADC) tumor suppressor. Biochim. Biophys. Acta 1332, F127 – F147. Prescott, S.M., White, R.L., 1996. Self-Promotion? Intimate connections between APC and prostaglandin H synthase-2. Cell 87, 783 – 786. Scholz, K., Fu¨rstenberger, G., Mu¨ller-Decker, K., Marks, F., 1995. Differential expression of prostaglandin-H synthase isoenzymes in normal and activated keratinocytes in vivo and in vitro. Biochem. J. 309, 263 – 269. Stadnyk, A.W., 1994. Cytokine production by epithelial cells. FASEB J. 8, 1041 – 1047. Tang, D.G., Chen, Y.Q., Honn, K.V., 1996. Arachidonate lipoxygenases as essential regulators of cell survival and apoptosis. Proc. Natl. Acad. Sci. USA 93, 5241 – 5246. Winde, G., Gumbinger, H.G., Osswald, H., Kemper, F., Bunte, H., 1993. The NSAID sulindac reverses rectal adenomas in colectomized patients with familial adenomatous polyposis: Clinical results of a dose-finding study on rectal sulindac administration. Int. J. Colorectal Dis. 8, 13 – 17.
.