Division Biochemistry of Tissue-specific Regulation (B0500)70 Research Program Tumor Cell Regulation Division Biochemistry of Tissue-specific Regulation DKFZ 2001: Research Report

69

Research ProgramTumor Cell Regulation

DivisionBiochemistry of Tissue-specific Regulation

DKFZ 2001: Research Report 1999/2000

Division Biochemistry of Tissue-specific Regulation (B0500)Head: Prof. Dr. rer. nat. Friedrich Marks

Scientists:Dr. rer. nat. Gerhard FürstenbergerDr. rer. nat. Michael GschwendtDr. rer. nat. Peter KriegDr. rer. nat. Karin Müller-DeckerDr. rer. nat. Michael RogersDr. rer. nat. Jürgen SchweizerDr. rer. nat. Hermelita WinterDr. rer. nat. Luise Stempka

Graduate students:Corinna Bähr Thorsten DahmSabine Häussermann Wolfgang HirschnerFelipe Jave-Suarez Karsten MüllerGitta Neufang Malte SiebertAnette Zagorski

Technical personnel:Ulrike Beckhaus (part time) Claudia EhmannWalter Kittstein (of absence) Dagmar Kucher (part time)Ina Kutschera Sandra PfrangAndrea Pohl-Arnold (part time) Gabriele RinckeBrigitte Steinbauer (part time) Ingeborg Vogt

Apprentices:Corinna Metzger Verona GschwendChristine Schmidt

Secretary: Wiltrud S. Bockemühl

Other technical services: Roswitha Epp

The division consists of three groups:Group 1: Eisosanoids and tumor development (Dr.Gerhard Fürstenberger)Group 2: Protein kinase C (Dr. Michael Gschwendt)Group 3. Normal and pathological epidermal differentiation(Dr. Jürgen Schweizer)

These three topics of research originate from the division’smany years of work on epidermal growth regulation and, inparticular, on chemical carcinogenesis using the mouseskin model of initiation-promotion. In this context the divi-sion had focussed its interest on tumor promotion and mo-lecular effects of skin tumor promoters such as the phorbolesters, in particular.As to the experimental approach, the division has widelyleft the more toxicologically and pharmacologically ori-ented work pursued in the past by turning to a moremechanistically oriented in-depth investigation of selectedaspects. For this purpose, techniques of molecular genet-ics and molecular biology have become the major tools ofresearch. As one ultimate goal cancer chemoprevention byinterrupting tumor development at a premalignant stage,i.e. by inhibiting tumor promotion, has crystallized out ofthese studies. In the last 5 years this development has ledto numerous collaborative projects with clinics and phar-maceutical companies.

1. Eicosanoids and Tumor DevelopmentHead of the group: Dr. Gerhard Fürstenberger

Eicosanoids (e.g. prostaglandins, thromboxanes, leuko-trienes etc.) are arachidonic acid- derived tissue hormones[1,28], which are critically involved in chronic and degen-erative diseases such as atherosclerosis, rheumatoidarthrithis, Alzheimer’s dementia, and cancer [27,29].Therefore, inhibitors of eicosanoid formation have beenproposed or already used for chemoprevention of thesediseases [2]. Using the animal model of multistage skincarcinogenesis we found antineoplastic effects of such in-hibitors including the so-called nonsteroidal anti-inflamma-tory drugs (NSAID) in the early eighties. This chemopre-ventive effect was found to be due to an inhibition of pros-taglandin synthesis in epidermal cells [3-5]. Especially forthe prostaglandin type F2� a tumor promoting effect couldbe demonstrated. In the early nineties these findings andcorresponding results from other laboratories met withclinical and epidemiological data demonstrating a pro-nounced chemopreventive effect of NSAIDs such as aspi-rin, sulindac, etc. in man, in particular against colorectalcancer [4] thus indicating a novel and promising measureof cancer chemoprevention [5].

More recently we have been following a twofold strategy:i.) to investigate mechanistic aspects of the role of arachi-

donic acid metabolism in carcinogenesis using themouse skin model and corresponding in vitro systems;

ii.) to extrapolate to the human situation.The final goal of these studies is to contribute to the devel-opment of novel mechanism-based cancer preventionstrategies by means of pharmacological intervention.

1.1 CyclooxygenasesK. Müller-Decker, G. Neufang, T. Dahm, F. Marks,G. FürstenbergerGuests: S. Loukanov (Univ. Sofia, Bulgaria), S. Charyalertsak(National Cancer Institute Bangkok, Thailand), R. Prashar (Univ.of Radjastan, Jaipur, India).

Collaboration with the dermatological clinics of the University ofRostock (Prof. Dr. G. Gross) and Mannheim (Dr. C. Bayerl), thesurgical clinic of the Univ. of Münster (Dr. G. Winde), the MahodiUniversity of Bangkok, Thailand (Prof. V. Sirikulchayanonta), theUniversity of California, Los Angeles, USA (Dr. H.R. Herschman)and the Searle Monsanto Comp. St. Louis, USA, the Bayer AGLeverkusen and the Falk Pharma Comp. Freiburg

Mouse skin model. Cyclooxygenases (COX) catalyse theformation of prostaglandins and thromboxanes fromarachidonic acid and are inhibited by nonsteroidal anti-in-flammatory drugs (NSAIDs). Presently 2 isoenzymes areknown, i.e. a ubiquitous COX-1 and a COX-2, which inmost tissues is expressed only temporarily, in particular inemergency situations such as wounding. inflammation, in-fections, etc. [6]. Our observations and results from otherlaboratories have shown, that a permanent overexpressionof COX-2 is an early event of carcinogenesis which might

70




be causally related to tumor development in many, possi-bly even in all epithelia [3-6].

Studies on the mouse skin model support this conclusionin that in normal epidermal cells only COX-1 is foundwhereas COX-2 appears only after wounding or treatmentwith a tumor promoting substance [7,8]. This COX-2 ex-pression correlates with the formation of prostaglandin F2�,which in mouse skin is directly involved in tumor promotion(see above). COX-2 was consistantly found in both benignand malignant tumors of mouse skin, and the treatment ofthe animals with selective COX-2 inhibitors suppressedboth prostaglandin formation and tumor development.

Human cancer. As in untreated mouse skin also in normalhuman skin only COX-1, but practically no COX-2 wasfound [9]. COX expression in basal cell carcinomas did notdiffer from that of normal skin. However, as in the mouseskin model upregulation of COX-2 was a consistant fea-ture of squamous cell carcinomas (SCCs) and, in addition,of actinic keratoses, i.e. benign preliminary stages of squa-mous cell carcinomas. The latter result indicates that theaberrant COX-2 expression is an early premalignant eventalso in the course of human skin cancer development.

These data may justify a clinical trial aimed at the preven-tion of skin cancer in high risk persons by local or systemicapplication of conventional NSAIDs or specific COX-2 in-hibitors. For this purpose studies addressing the skin per-meability of NSAIDs are being performed in collaborationwith the pharmaceutical industry.

A pathological overproduction of COX-2 was also found intissue samples from patients with Crohn’s Colitis andcolorectal tumors [10]. Crohn’s Colitis is a preliminarystage of large bowel cancer. However, the drug 5-ami-nosalicylic acid, which is most frequently used against in-flammatory bowel diseases turned out to be only a weakCOX-2 inhibitor.

In a collaborative study with the National Cancer Instituteof Thailand we observed a strong overexpression of COX-2 also in human gall bladder cancer (cholangiocarcinoma).This tumor is abundant in Southeast Asia being most prob-ably promoted by chronic tissue damage which is causedby a liver fluke. This situation corresponds to the tumorpromoting effects of chronic tissue damage in the mouseskin model. In contrast to neoplastic skin, gall bladder tu-mors overproduced both COX-enzymes, whereas in nor-mal gall bladder tissue no COX-2 was found [11] ( see alsoFig.1). Overexpression of COX-2 was also found in livercancer, which is the most frequent cancer in Thailand (un-published results). The goal of these studies is to find outwhether or not chemoprevention by NSAIDs of liver andgall bladder tumors will be feasible, at least for high riskpopulations in Southeast Asia.

Transgenic mouse models. To understand function ofaberrant COX-2 expression in skin transgenic mouse lineshave been established, expressing the gene in a tissue-and differentiation-specific manner in epidermis. For thispurpose, the COX-2 gene was fused with skin-specificgene promoters, such as those for keratin 5 and keratin 6.Mice which carry the COX-2 gene under the control of the

keratin 5 promoter exhibit pronounced alterations of skinmorphology such as a disturbancy of hair growth and hy-perplasia of the sebaceous glands [26]. The sensibility ofthese animals for chemical skin carcinogenesis as well asthe effect of the transgene on embryonic skin developmentare presently under investigation.

Proinflammatory mediator release as an in vitro pa-rameter of skin irritancy. In cooperation with theSchering AG we developed an in vitro test for skin irritantsbased on the release of arachidonic acid and interleukin-1� from a human keratinocyte line. This test was validatedby a placebo-controlled, open, randomized patch teststudy with volunteers and shown to provide a suitable sub-stitute for animal experiments [12-15]. The responsible sci-entist, Dr. Karin Müller-Decker, has been awarded theFelix Wankel Tierschutzpreis 2000.

Fig.1. This microscopic picture shows the excessive formation ofthe enzyme cyclooxygenase-2 (COX-2) in a human bile duct car-cinoma from Thailand. Note the staining of the cancer tissue witha COX-2-specific antibody (black arrow). In contrast, normal bileduct tissue is devoid of COX-2 (white arrow). Such findings nour-ish the hope in chemoprevention with COX inhibitors (such as as-pirin) of bile duct cancer which is particularly frequent in South-east Asia.A similar overproduction of COX-2 was found in carcinomas andprecancerous tissues of many other organs.

2.1 LipoxygenasesP. Krieg, M. Heidt, F. Bürger, M. Siebert, K. Müller,A. Zagorski, F. Marks, G. FürstenbergerGuest: N. Qiao

Cooperation with: N. Nair, H. Bartsch, W. Lehmann, C-W. von derLieth (DKFZ), B. Marian (University of Vienna, Austria), A. Brash(Vanderbilt University, Nashville, USA), K. Honn (Wayne StateUniversity, Detroit, USA), P.L: Grover (Haddow Laboratories,Sutton, UK) and the L’Oreal Company,Paris, France

Lipoxygenases (LOX) represent a widespread family oflipid-peroxidizing enzymes catalyzing the regioselectiveand stereoselective dioxygenation of free and esterifiedpolyenic fatty acids to the corresponding hydroperoxy de-rivatives, which may be metabolized to various bioactivelipid mediators, including leukotrienes, lipoxins, hydroxy-eicosatetraenoic acids (HETEs), and hepoxilins. LOX me-tabolites are thought to play an important role in the matu-ration of certain cell types such as erythrocytes and kerati-

71




nocytes, in inflammatory processes, blood clotting, leuko-cyte chemotaxis, in certain inflammatory skin diseases likepsoriasis, as well as in atherosclerosis and cancer.

Using the mouse skin carcinogenesis model we previouslyhave provided evidence that LOX-catalyzed eicosanoidmetabolism plays an important role in epithelial tumor de-velopment. In an effort to analyze the epidermal LOX-cata-lyzed arachidonic acid metabolism in more detail we havecloned six distinct LOX isoforms from mouse skin by usingcDNA libraries originating from normal and TPA-treatedmouse skin and from mouse skin tumors as well. In addi-tion to the human orthologs of the platelet-type and leuko-cyte-type 12S-LOX a third 12S-LOX isoenzyme was identi-fied which upon its specific expression in epidermis wascalled epidermis-type 12S-LOX. Furthermore, by using de-generate PCR technique we were able to clone threemembers of a novel distinct epidermis-type subclass withinthe LOX multigene family including 8S-LOX, 12R-LOX andepidermis-type LOX-3 [16,17].

12R-LOX represents the first mammalian LOX isoenzymewith R-chirality, while e-LOX-3 possesses an of-as-yet un-known enzymatic specificity. Their unusual size (701 and711 amino acid residues) among mammalian LOX hasbeen attributed to an extra segment of 31 and 41 aminoacids, respectively, which can be located beyond the N-terminal �-barrel domain at the surface of the C-terminalcatalytic domain. The corresponding genes, unlike thegenes of all other mammalian LOX, are split into 15 exonsand 14 introns indicating that they represent members of anovel structural class of mammalian LOX. Together withthe gene encoding 8-LOX and its human ortholog 15-LOX-2, respectively, they were found tightly clustered at chro-mosome 11 and at the syntenic human region 17p13.1[18,19].

Epidermis-type LOX were shown to exhibit unusual enzy-matic properties suggesting that these enzymes preferen-tially utilize more complex lipids or that either activating co-factors or protein modifications are required for full enzy-matic activity [20,21]. These unusual enzymatic propertiesmay be attributed to distinct functions of epidermal LOX for

the structural and functional integrity of stratifying epithe-lia.

The six LOX isoenzymes found in epidermis exhibited adifferentiation-dependent mRNA expression pattern, indi-cating that the individual isoenzymes may be associatedwith processes of proliferation and terminal differentiation[22]. Evidence for an involvement of distinct LOX in epider-mal tumor development is provided by the observation thatthe platelet-type 12S-LOX and the 8S-LOX were found tobe overexpressed in tumors and the concomitant accumu-lation of the corresponding 8- and 12-HETE [23]. Both in-duce chromosomal damage in keratinocytes in vitro andare thought - together with other products of arachidonicacid metabolism - to mount an endogenous genotoxic po-tential, which may be involved in the accumulation of ge-netic alterations occurring in the course of malignant pro-gression from papillomas and carcinomas [24].

Publications (* = external coauthors)[1] Marks, F. (1999) Arachidonic acid and companions: an abun-dant source of biological signals. In F. Marks and G. Fürstenber-ger (eds) Prostaglandins, Leukotrienes, and Other Eicosanoids.From Biogenesis to Clinical Applications, Wiley/VCh, Weinheimpp. 1-45.[2] Marks, F., Fürstenberger, G. (1999) Krebsprävention mitSchmerzmitteln. Spektrum der Wissenschaft (German issue ofScientific American) 2, 52-60.[3] Marks, F., Müller-Decker, K., Fürstenberger, G. (1998)Eicosanoids as endogenous mediators of carcinogenesis and re-porters of skin toxicity: cancer chemoprevention by inhibitors ofarachidonic acid metabolism. Advances in Mol. Toxicol. (Reiss, C.,Parvez, S, Lobbe, G., Parvez, H. eds) pp. 247-267.[4] Marks, F., Fürstenberger, G. (1999) Eicosanoids and Cancer.In F. Marks and G. Fürstenberger (eds) Prostaglandins,Leukotrienes, and Other Eicosanoids. From Biogenesis to ClinicalApplications, Wiley/VCh, Weinheim pp. 305-332.[5] Marks, F., Fürstenberger, G., Müller-Decker, K. (1998) Meta-bolic targets of cancer chemoprevention: interruption of tumor de-velopment by inhibitors of arachidonic acid metabolism. In: H.J.Senn, A. Costa, C. Jordan (Herausg.): ‘Cancer Chemoprevention’,Rec. Results in Cancer Res. 151, 45-66. Springer, Berlin 1998.[6] Müller-Decker, K. (1999) Prostaglandin H isoenzymes. In F.Marks and G. Fürstenberger (eds) Prostaglandins, Leukotrienes,and Other Eicosanoids. From Biogenesis to Clinical Applications,Wiley/VCh, Weinheim.

Fig.2. High tissue levels of 8- and 12-hydroxyeicosatetraenic acid (HETE) in mouse skin tu-mors as a symptom of an impaired regulationof lipoxygenase-controlled arachidonic acidmetabolism. Note that this aberration is foundparticularly in benign lesions (papillomas).8- and 12-HETE are chromosome-damagingagents and their accumulation in tumor tissuescorrelates with an increase of DNA-damage. Adysregulated arachidonic acid metabolism maythus contribute to the so-called genetic insta-bility of tumor cells, which is thought to facili-tate the conversion of the benign into the ma-lignant state. Control animals were treated ei-ther with acetone or the tumor promoter TPA.

1 1 2 3 1 2 3 4 5 6 1 2 3 4 5 6 10

1000

2000

3000

4000

5000

12-HETE 8-HETE

HETE [ng/g tissue]

control control reversible irreversible carcinomas(acetone) (TPA) papillomas papillomas

72




[7] Müller-Decker, K., Kopp-Schneider, A., Marks, F., *Seibert, K.,Fürstenberger, G. (1998) Localization of prostaglandin H synthaseisoenzymes in murine epidermal tumors: suppression of skin tu-mor promotion by inhibition of prostaglandin H synthase-2. Molec.Carcinogenesis 23: 36-44.[8] Müller-Decker, K., Scholz, K., Neufang, G., Marks, F.,Fürstenberger, G. (1998) Localization of prostaglandin-H syn-thase-1 and -2 in mouse skin: implication for cutaneous function:Exp. Cell Res. 242: 84-91.[9] Müller-Decker, K., Reinerth, G., Krieg, P., Marks, F.,*Zimmermann, G., *Heise, U., *Bayerl, C., Fürstenberger, G.(1999) Expression of prostaglandin-H synthase isozymes in nor-mal human skin and epidermal neoplasias. Int. J. Cancer 80: 648-656.[10] Müller-Decker, K., Albert, C., *Lukanov, T., *Winde, G., Marks,F., Fürstenberger, G. (1999) Cellular localization of cyclooxygen-ase isozymes in Crohn´s colitis and colorectal cancer. Int. J.Colorectal Dis. 14: 212-218.[11] *Chariyalertsak, S., *Sirikulchayanonta, V., Mayer, D., Kopp-Schneider, A., Fürstenberger, G., Marks, F., Müller-Decker, K(2001). Aberrant cyclooxygenase isozyme expression in humanintrahepatic cholangiocarcinoma. Gut, 48: 80-86.[12] Müller-Decker, K., Heinzelmann, T., Fürstenberger, G.,Marks, F. (1998) Entzündungsmediatoren aus einer humanenKeratinozytenlinie als Indikatoren für Hautreizung: Entwicklungeiner Ersatzmethode für den Draize-Test. Krebsforschung heute,Steinkopf-Verlag, Darmstadt, pp 47-51.[13] Müller-Decker, K., Heinzelmann, T., Fürstenberger, G.,*Kecskes, A., Lehmann , W.D., Marks, F. (1998) Arachidonic acidmetabolism in primary irritant dermatitis produced by patch testingof human skin with surfactants. Toxicol. Appl. Pharmacol. 153: 59-67.[14] Marks, F., Fürstenberger, G., Müller-Decker, K. (1998) Arachi-donic acid metabolism as a reporter of skin irritancy and target ofcancer chemoprevention. Toxicol. Lett. 96: 111-118.[15] Müller-Decker, K., Heinzelmann, T., Fürstenberger, G.,Marks, F. (1998) Entzündungs-mediatoren aus einer humanenKeratinocytenlinie als Indikatoren für Hautreizung: Entwicklungeiner Ersatzmethode für den Draize-Test. In: Schöffl et al. (eds.)‘Forschung ohne Tierversuche’. Springer Wien etc. (1998).[16] Krieg, P., Kinzig, A., Heidt, M., Marks, F., Fürstenberger, G.(1998). cDNA cloning of a 8-lipoxygenase and a novel epidermis-type lipoxygenase from phorbol ester-treated mouse skin.Biochim. Biophys. Acta 1391: 7-12.[17] Krieg, P., Siebert, M., Kinzig, A., Bettenhausen, R., Marks, F.,Fürstenberger, G. (1999) Murine 12(R)-Lipoxygenase: Functionalexpression, genomic structure and chromosomal localization.FEBS Lett 446:142-148.[18] Krieg, P., Marks, F., and Fürstenberger, G. A gene cluster en-coding human epidermis-type lipoxygenases at chromosome17p13.1: Cloning, physical mapping, and expression. Genomics,in press.[19] Kinzig, A, Heidt M., Fürstenberger G, Marks F., and Krieg P.(1999) cDNA cloning, genomic structure and chromosomal local-ization of a novel murine epidermis-type lipoxygenase. Genomics,58, 158 – 164.[20] Bürger, F., Krieg, P., Marks, F., and Fürstenberger G. (2000)Positional- and stereo-selectivity of fatty acid oxygenationcatalysed by mouse (12S)-lipoxygenase isoenzymes. Biochem. J.348, 329-335.[21] Siebert, M., Krieg, P., Lehmann, W.D., Marks, F., andFürstenberger, G. (2001) Enzymatic characterization of epidermis-derived 12-lipoxygenase isoenzymes. Biochem. J. 355:97-104.[22] Heidt, M., Fürstenberger, G., Vogel, S., Marks, F., Krieg, P.(2000) Diversity of murine lipoxygenases: identification of a sub-family of epidermal isozymes exhibiting a differentiation-depen-dent mRNA expression pattern, Lipids 35: 701-707.

[23] Bürger, F., Krieg, P., Kinzig, A., Schurich, B., Marks, F.,Fürstenberger, G. (1999) Constitutive expression of 8-lipoxy-genase-derived arachidonic acid metabolites in keratinocytes.Molec. Carcinogenesis 24: 108-118.[24] Nair, J., Fürstenberger, G., Bürger, F., Marks, F., Bartsch, H.(2000) Promutagenic etheno-DNA adducts in multistage mouseskin carcinogenesis: correlation with lipoxygenase-catalyzedarachidonic acid metabolism. Chem. Res. Toxicol. 13: 703-709.[25] Marks, F., Fürstenberger, G. (2000) Cancer chemopreventionthrough interruption of multistage carcinogenesis: The lessonslearned by comparing mouse skin carcinogenesis and humanlarge bowel cancer. Europ. J. Cancer 36: 314-329.[26] Neufang, G., Fürstenberger, G., Marks, F., Müller-Decker, K.(2001) Abnormal differentiation of epidermis in transgenic miceconstitutively expressing cyclooxygenase-2 in skin. Proc. Natl.Acad. Sci. USA, 98: 7629-7634.[27] Marks, F., Müller-Decker, K., Fürstenberger, G. (2000) Acausal relationship between unscheduled eicosanoid signalingand tumor development: cancer chemoprevention by inhibitors ofarachidonic acid metabolism. Toxicology 153: 11-26.[28] Marks, F. (2000) Der Stoffwechsel der Arachidonsäure.Biologie in unserer Zeit 30: 342-353.[29] Marks, F. (2001) Krebs- und Alzheimer-Prävention mit nicht-steroidalen Entzündungshemmern. Deutsche Med. Wochenschr.126: 308-313.

2. Protein Kinase CHead of the group: Dr. Michael Gschwendt

Protein kinase C (PKC) covers a family of 12 isoenzymeswith phospholipid-dependent serine/threonine kinase ac-tivity [1]. These kinases occupy a key position in intracellu-lar signal-processing whereby the individual isoenzymesseem to exhibit specific functions and to be regulated se-lectively in a cell type-dependent manner. In experimentalcancer research the c- and n-type PKCs have gained con-siderable interest as cellular receptors of phorbol ester-type skin tumor promotors such as TPA. In the course ourstudies on skin tumor promotion we had for the first timeisolated and characterized an n-type PKC, i.e. PKC�, in anative form [reviewed in ref. 2]. PKC� was found to beubiquitously expressed and is thought to play an importantrole in intracellular signaling. In the meantime various au-thors have related this isoenzyme to the control of differen-tiation and proliferation in epidermis and to apoptosis, i.e.processes critical for tumor promotion.In the past 2 years most of our work has been concernedwith the posttranslational regulation of PKC� activity andlocalization as well as with the identification of interactingproteins.

2.1 Studies on PKC�

2.1.1 Posttranslational modification byphosphorylationL. Stempka, M. Gschwendt, F. Marks

Cooperation with various groups at the DKFZ

Activation loop phosphorylation. Many protein kinases,including PKC� and �, require a specific phosphorylationin the so-called activation loop of the active center in orderto acquire a catalytically competent conformation. Wecould demonstrate that this is not true for PKC� [3,4]. Incontrast, for instance, to PKC� the wild type of PKC� as

73




well as a mutant lacking the cognate phosphorylation sitein the activation loop could be expressed in bacteria in acatalytically competent form, i.e., were activatable byphospholipid and the phorbol ester TPA [3]. By means ofsite-directed mutagenesis we showed that a glutamic acidresidue in the activation loop of PKC� (position 500) takesover the role of the phosphate group [4]. Thus, contrary toother PKC isoenzymes PKC� does not require activationloop phosphorylation by a PKC-kinase such as phospho-lipid-dependent kinase I (PDKI).

Autophosphorylation. In addition to activation loop phos-phorylation, some PKC isoforms require autophosphoryl-ation of specific sites for a stabilization of the enzymati-cally active form. PKC� autophosphorylates with ATP or,even more effectively, with GTP as a phosphate donor [2].By means of MALDI mass spectrometry of tryptic peptideswe have identified serine 643 as an in vitro -and in vivo-autophosphorylation site of PKC� [4]. Site-directed muta-genesis indicated, however, that phosphorylation of thissite is not essential for enzymatic activity but may serveanother as yet unidentified purpose. Moreover, PKC� con-tains additional (auto)phosphorylation sites still to be iden-tified.

Tyrosine phosphorylation. We have found PKC� to bephosphorylated on tyrosine by the tyrosine kinase Src invitro and upon stimulation of cells with various mitogenicagents [2]. It is likely that the tyrosine phosphorylation invivo is also mediated by Src. By MALDI mass spectrom-etry we identified Tyr 311 as one of the sites that are phos-phorylated by Src in vitro (unpublished results). While, ingeneral, the function of tyrosine phosphorylation of PKC isnot well understood, our results indicate that in the case ofPKC�, in particular, it may regulate substrate specificity [2].

2.1.2 Interacting proteinsC. Bähr, L. Stempka, A. Rohwer, M. Gschwendt,F. Marks.

Cooperation with Dr. K. Bomsztyk, University of Washington, Se-attle, USA.

K protein. There is an increasing body of evidence indicat-ing that scaffold or anchoring proteins critically determineboth the specific subcellular localization and the selectiveinteraction of signal transducing proteins such as proteinkinases.

In a cooperation with K. Bomsztyk we found that the so-called K protein behaves like an anchoring protein forPKC� [5,6] The K protein, a constituent of the heterogen-ous nuclear ribonucleoprotein particle, shuttles betweenthe cytosol and the nucleus, binds nucleic acids and vari-ous signal-transducing proteins, including tyrosine kinasesof the Src family, and may play a role in the control of tran-scription and translation. In vitro and in vivo PKC� bindswith high affinity to the same region of the K protein (KI do-main) as Src or Lck. The K protein is phosphorylated bythese tyrosine kinases as well as by PKC� in vitro and invivo. The phosphorylation sites Tyr 236 and Ser 302, re-spectively, were identified by mutational analysis andMALDI mass spectrometry. Phosphorylation of Tyr 236 in

vitro and in vivo induces a dramatic increase in the bindingof Src (Lck) and PKC� to K protein thereby stimulating thephosphorylation of Ser 302 by PKC�. Phosphorylation ofSer 302 in turn induces release of PKC� from K protein.We conclude, that binding of PKC� and Src (Lck) to K pro-tein is tightly regulated in response to extracellular signalsand might facilitate and control phosphorylation of PKC�

by the tyrosine kinases of the Src family and -vice versa-phosphorylation and probably inactivation of Src by PKC�.Whether the K protein also mediates a selective phospho-rylation of other PKC� substrate proteins remains to beshown.

Studies using the yeast two-hybrid system (YTHS).More recently, we have applied the yeast two-hybrid sys-tem (YTHS) for cDNA cloning of PKC-interacting proteins(including substrates) from cDNA libraries. Since relatedinvestigations on PKC� had already been published byother groups, we began our studies using this isoenzyme,then extending the work on PKC� and just recently also onPKCµ. As baits we used the catalytic and regulatory do-mains of PKC�, PKC�, and PKCµ as well as the completeenzymes. Several clones were isolated. Their character-ization is in progress. One clone fished with the catalyticdomain of PKC� as a bait was identified as a novel A6-re-lated protein [A6rp; ref. 7]. A6rp was found to be ex-pressed ubiquitously. It binds ATP and is phosphorylatedby PKC� rather than by other PKCs. Recently, we haveisolated a novel PKC�-binding protein containing structuraldeterminants of a nuclear protein kinase that was termedDIK (Delta interacting kinase).

DIK, a novel protein kinse that interacts with PKC� [8].A novel serine/threonine kinase, termed DIK, was clonedusing the yeast two-hybrid system to screen a cDNA li-brary from the human keratinocyte cell line HaCaT with thecatalytic domain of rat protein kinase C� (PKC�cat) cDNAas bait. The predicted 784-amino acid polypeptide with acalculated molecular mass of 86 kDa contains a catalytickinase domain and a putative regulatory domain withankyrin-like repeats and a nuclear localization signal. Ex-pression of DIK at the mRNA and protein level could bedemonstrated in several cell lines. The dik gene is locatedon chromosome 21q22.3 and possesses 8 exons and 7introns. DIK was synthesized in an in vitro transcription/translation system and expressed as recombinant proteinin bacteria, HEK, COS-7, and baculovirus-infected insectcells. In the in vitro system and in cells, but not in bacteria,various posttranslationally modified forms of DIK were pro-duced. DIK was shown to exhibit protein kinase activity to-wards autophosphorylation and substrate phosphorylation.The interaction of PKC�cat and PKC� with DIK was con-firmed by co- immunoprecipitation of the proteins fromHEK cells transiently transfected with PKC�cat or PKC� andDIK expression constructs.

2.2 Studies on PKCµS. Häussermann, M. Gschwendt, F. Marks

Cooperation with Dr. F.-J. Johannes, University of Stuttgart.

74




PKCµ is more distantly related to the PKC family. Beingalso activated by phospholipid and diacylglycerol or TPA itdiffers from other PKCs by a pleckstrin homology domainand lack of a pseudosubstrate region. In a cooperationwith Dr. Johannes (University of Stuttgart) we studied therather unusual properties of PKCµ, as compared to PKC�

and other PKCs. PKCµ is an ubiquitous enzyme that is ac-tivated and phosphorylated but, contrary to other PKCs,not downregulated upon treatment of cells with TPA. Theefficient and rather specific PKC inhibitor Gö 6983 wasfound to be widely ineffective in suppressing PKCµ, thusallowing for a differentiation between PKCµ and the otherPKCs in vitro and in vivo. The identification of PKCµ sub-strates represents an unsolved problem. With the excep-tion of some synthetic peptides typical PKC substrates arephosphorylated only very weakly if at all by PKCµ. Whilewe identified aldolase as a rather good in vitro substrate ofPKCµ no in vivo substrates could be detected as yet. Thelack of a pseudosubstrate domain indicates that PKCµ dif-fers from other PKC isoenzymes also in the regulation ofenzyme activity. We found PKCµ to be inhibited by basicpeptides and activated by heparin, suramin, and dextransulfate to a much higher extent than by TPA/phosphatidyl-serine. Based on these results, we have postulated that anacidic domain (amino acids 336-391) is involved in theregulation of PKCµ activity.

In an attempt to elucidate a putative role of PKCµ in theregulation of the programmed cell death we found thatPKCµ is cleaved proteolytically upon induction ofapoptosis [9].

Treatment of U937 cells with various apoptosis-inducingagents, such as TNF� and �-D-arabinofuransylcytosine(ara-C) alone or in combination with the phorbol esterTPA, bryostatin 1 or cycloheximide, causes proteolyticcleavage of protein kinase Cµ (PKCµ) in the region be-tween regulatory and catalytic domain, generating a 62kDa fragment of the kinase. The formation of this fragmentis effectively suppressed by the caspase-3 inhibitor Z-DEVD-FMK. In accordance with these in vivo data, treat-ment of recombinant PKCµ with caspase-3 in vitro resultsalso in the generation of a 62 kDa fragment (p62). Treat-ment of several aspartic acid to alanine mutants of PKCµwith caspase-3 resulted in an unexpected finding. PKCµ isnot cleaved at one of the typical cleavage sites containingthe motif DXXD but at the atypical site CQND378/S379. Therespective fragment (amino acids 379-912) was expressedin bacteria as a GST fusion protein (GST-p62) and partiallypurified. In contrast to the intact kinase, the fragment doesnot respond to the activating cofactors TPA and phospha-tidylserine and thus is unable to phosphorylate substateseffectively.

Publications (* = external co-authors)[1] Marks, F., Gschwendt, M. (1996). In: ‚Protein Phosphorylation‘(F. Marks, ed.), p.81-116. VCH, Weinheim[2] Gschwendt, M. (1999) Protein kinase C�. Eur. J. Biochem.259: 555-564.

[3] Stempka, L., Girod, A., Müller, H.-J., Rincke, G., Marks, F.,Gschwendt, M., Bossemeyer, D.: Phosphorylation of protein ki-nase C� at threonine 505 is not prerequisite for enzymatic activity.Expression of rat protein kinase C� and an alanine 505 mutant inbacteria in a functional form. J. Biol. Chem. 272, 6805-6811(1997).[4] Stempka, L., Schnölzer, M., Radke, S., Rincke, G., Marks, F.,Gschwendt, M. (1999) Requirements of protein kinase C� forcatalytic function. J. Biol. Chem. 274: 8886-8892.[5] *Schullery, D.S., *Ostrowski, J., *Denisenko, O.N., Stempka,L., *Shnyreva, M., *Suzuki, H., Gschwendt, M., *Bomsztyk, K.(1999) Regulated interaction of protein kinase C� with the heter-ogenous nuclear ribonucleoprotein K protein. J. Biol. Chem. 274:15101-15109.[6] Ostrowski, J., Schullery, D., Denisenko, O.N., Gschwendt, M.,Bomsztyk, K. (2000) Role of tyrosine phosphorylation in the regu-lation of the interaction of hnRNP K protein with its protein andRNA partners. J. Biol. Chem. 275: 3619-3628.[7] Rohwer, A., Kittstein, W., Marks, F., Gschwendt, M. (1999)Cloning, expression and characterization of an A6-related protein.Eur. J. Biochem. 263: 1-9.[8] Bähr, C., Rohwer, A., Stempka, L., Rincke, G., Marks, F.,Gschwendt, M. (2000) DIK, a novel protein kinase that interactswith protein kinase C�. J. Biol. Chem. 275: 36350-36357.[9] Häussermann, S., Kittstein, W., Rincke, G., *Johannes, F.-J.,Marks, F., Gschwendt, M. (1999) Proteolytic cleavage of proteinkinase Cµ upon induction of apoptosis in U937 cells. FEBS Let-ters 462: 442-446.

3. Differentiation of Normal and NeoplasticEpidermisHead of the group: Dr. Jürgen Schweizer

The group investigates the normal and pathological differ-entiation of the hair follicle, i.e. the largest adnexal organof the epidermis. There studies have been focussed on acharacterization of hair keratin genes, on the expressionpattern of these genes and on the relationship betweenkerratin gene mutations and heritable hair anomalies.

3.1 Elucidation of the human hair keratinmultigene familyJ. Schweizer, M.A. Rogers, H. Winter

In cooperation with L. Langbein, DKFZ

The hair follicle is a morphologically complex organ,whose structures are determined by both cytokeratins (soft�-keratins) and hair keratins (hard �-keratins). Using bothcDNA and genomic libraries we were able to characterizethe type I and type II hair keratin gene loci on chromo-somes 17121.2 and 12q13, respectively. The type I locuscomprises 10 genes, including one transcribed pseudo-gene. The type II locus consists also consists of 10 genes,including 3 pseudogenes and one transcribed pseudogene[2,7]. All genes were sequenced.

By means of both specific cDNA probes and specific anti-bodies we determined the sequential hair keratin expres-sion patterns in the hair follicle [2,7]. Using the hair keratinantibodies for 2DE Western blots, we established a cata-log of the type I hair keratins [4]. The corresponding cata-log of the type II hair keratins is currently in progress. Inthe course of these studies, we were able to identify anovel type II cytokeratin, K6hf, which is specifically ex-

75




pressed in the companion layer of the hair follicle [3]. Thisshowed that, besides the outer and inner root sheath, thecompanion layer represents an independent epithelialcompartment of the hair follicle [ 4]. Currently, we are char-acterizing another multigene family, whose members arespecifically expressed in the hair follicle. These are the socalled “keratin associated proteins”, KAP, which generatethe matrix for the hair keratin intermediate filaments on thecortex and the cuticle of the hair shaft.(These studies aresupported by the DFG).

3.2 Hereditary hair anomalies.J. Schweizer, M.A. Rogers, H. Winter

In cooperation with Dr. A Taieb, Bordeaux, France; Dr. P.Vabres,Poitiers, France; Dr. R.D Clark, Chicago, USA; Dr. D.J. Schissel,US Army Hospital, Heidelberg

In a clinical project with various Departments of Dermatol-ogy, we investigate whether mutations in hair keratinsmay be the cause for a variety of hereditary hair disorders,including monilethrix, the loose anagen hair syndrome andpseudofollicultis barbae. We were able to show that moni-lethrix is caused by mutations in two hair keratins. The mu-tations are located at the beginning and the end of thecentral rod domain, which is necessary for the assembly ofhair keratins into intermediate filaments. A genotype-phe-notype correlation for monilethrix could not be determined[1,5,6]. Currently we investigate whether mutations in hairkeratins are also involved in the etiology of the looseanagen hair syndrome, LAH and in pseudofolliculitis bar-bae, PFB. Clinically, LAH patients suffer from easily re-movable anagen hairs, and PFB patients exhibit the for-mation of inflammatory pustules and ingrown hairs aftershaving. Since PFB occurs preferentially in negroid males,these studies are performed in cooperation with the US

Army. There is preliminary evidence that, besides a ge-netic predisposition in individuals with curled hairs, a mu-tation in K6hf may be causally involved in PFB (Thesestudies are supported by the DFG).

3.3 Regulation of hair keratin expressionL.F. Jave-Suarez, H. Winter, J. Schweizer

In cooperation with Dr. B. Cribier, Strasbourg, France

The elucidation of the human hair keratin family was aprerequisite for studies on the regulation of human hairkeratin expression. In the framework of a Ph.D. thesis, wecurrently investigate the role of various transcriptions fac-tors in the expression control of hair keratin genes. Thesefactors are the lymphocyte enhancer factor 1, LEF 1, �-catenin, Hoxc13 and the androgen receptor. In close asso-ciation with regulation studies, we investigate a variety ofhuman tumors which are thought to originate from the hairfollicle. Preliminary data suggest mutations in the ß-cate-nin gene to be involved in the formation of pilomatricomasand trichepitheliomas. Using the antibodies against the in-dividual hair keratins, we also envisage a better classifica-tion of these tumor types.

Publications (* = external co-authors)[1] Winter H, *Labrèze C, *Chapalain V, *Surlève-Bazeille JE,*Mercier M, Rogers MA, *Taieb A, Schweizer J. (1998) A variablemonilethrix phenotype associated with a novel mutationGlu402Lys in the helix termination motif of the type II hair keratinhHb1. J. Invest. Dermatol. 111:169-172[2] Rogers MA, Winter H, Wolf C, Jacobs M, Schweizer J. (1998)Characterization of a 190 kbp domain of human type I hair keratingenes. J. Biol. Chem. 273: 26683-26691[3] Winter H, Langbein L, Praetzel S, Jacobs, M, Rogers MA,Leigh IM, Tidman N, Schweizer J. (1998) A novel human type IIepithelial keratin,K6hf, specifically expressed in the companionlayer of the hair follicle. J. Invest. Dermatol. 111: 955-962.[4] Langbein L, Rogers MA, Praetzel S, Beckhaus U, RackwitzHR, Winter H, Schweizer J. (1999) The catalog of human hairkeratin genes. I. Expression of the nine type I hair keratins in thehair follicle. J. Biol. Chem. 274: 19874-19884[5] Winter H, *Clark,RD, *Tarras-Wahlberg C, Rogers MA,Schweizer J. (1999) Monilethrix: A novel mutation, Glu402 Lys, inthe helix termination motif and the first causative mutation,Asn114Asp, in the helix initiation motif of the type II hair keratinhHb6. J. Invest. Dermatol. 113: 263-266[6] Winter H, Rogers MA, Vabres P, Larrègue M, Schweizer J.(2000) A novel missense mutation, A118E, in the the helix initia-tion motif of the type II cortex keratin hHb6 causing monilethrix.Hum. Heredity 50: 322-324[7] Rogers MA, Winter H, Langbein L, Schweizer J. (2000) Char-acterization of a 300 kpb region of human DNA containing thetype II hair keratin gene locus. J Invest. Dermatol. 114: 464-472

Fig.3. Monilethrix patient. i)2 years old, with massivehairloss on the back of thehead, ii) 6 years old show-ing regrowth of fragilehairs, iii) microscopic pic-ture of Moniletrix hairs:note the constrictions werethe hairs easily break off.

Documents

Division Biochemistry of Tissue-specific Regulation (B0500)70 Research Program Tumor Cell Regulation Division Biochemistry of Tissue-specific Regulation DKFZ 2001: Research Report