Biochimica et Biophysica Acta, 1160 (1992) 127-133 © 1992 Elsevier Science Publishers B.V. All rights reserved 0167-4838/92/$05.00

127

BBAPRO 34360

Calcium transport and regulation in human primary and metastatic melanoma Karin U. Schallreuter, John M. Wood, Christine Ehrke and Regina Lemke Department of Dermatology, Universityof Hamburg, Hamburg (Germany)

Key words: Calcium; Beta-2-adrenoceptor; Human melanoma; Epinephrine Thioredoxin reductase (TR) activity on primary melanomas and in surrounding skin is regulated by calcium and, therefore, TR activity can be used to measure the flux of calcium between primary tumors and their surrounding epidermis. Calcium uptake in human melanotic melanoma cell lines SKmel-23 (metastatic) and BC-PT-1 (primary) is related to the density of/3-2-adrenoceptors. The non-pigmented cell line HT-144 (metastatic), did not express/3-2-adrenoceptors, yielding a slow rate of calcium uptake compared to SKmel-23 and BC-PT-1. Cell extracts from melanotic and amelanotic melanoma tissues did not contain a phenylethanolamine-N-methyltransferase (PNMT) for the biosynthesis of epinephrine from norepinephrine and S-adenosylmethionine. However, human full-thickness skin, epidermis and cell cultures of human keratinocytes contained significant PNMT activities. Taken together, these results indicate that (a), TR can be used to monitor calcium flux between primary melanomas and their surrounding skin and vice versa and (b), calcium uptake may be regulated by stimulation of/3-2-adrenoceptors on melanotic melanomas by epinephrine synthesized in the surrounding skin.

Introduction The regulation of human metastatic melanoma thioredoxin reductase (TR) by Ca 2+ was realized after this enzyme was purified from pigmented and depigmented melanoma tissues [1]. A kinetic analysis of the reduction of DTNB, using N A D P H as the electron donor, revealed that enzyme from melanotic melanoma yielded sigmoidal kinetics in V vs. S plots; meanwhile, the same enzyme from amelanotic melanoma showed normal saturation with substrate [1,2]. Furthermore, the two enzymes differed in their chromatographic properties on FPLC. The allosteric regulated enzyme eluted at 0.105 M NaC1 on a Mono Q H R 5 / 5 column, whereas the enzyme showing normal Michaelis-Menten kinetics eluted at 0.12 M NaC1 [3]. This result suggested that these two enzymes had a different surface charge [3]. When the enzyme from amelanotic melanoma was preincubated with radiolabelled 45Ca2+, it was converted to the allosteric form, showing sig-

Correspondence to: K.U. Schallreuter, Department of Dermatology, University of Hamburg, Martinistral3e 52, Germany. Abbreviations: TR, thioredoxin reductase; PNMT, phenylethanolamine-N-methyltransferase; DTNB, 5.5'-dithiobisnitrobenzoic acid; NADPH, reduced nicotinamide adenine dinucleotide phosphate; FPLC, fast-protein liquid chromatography;TCA, trichloroacetic acid.

moidal kinetics and it eluted as a single radioactive peak at 0.105 M NaCI upon FPLC analysis [1-3]. Furthermore, melanoma T R has been shown to bind one 45Ca2+ per enzyme molecule with 1:1 stoichiometry at a concentration of 10 -6 M Ca 2+. Because the sequence of mammalian T R had not been determined, we examined Ca 2+ binding on the enzyme purified from Escherichia coli, where the primary sequence has been established [4]. T R from E. coli bound one 45Ca2+ per enzyme molecule [1]. An Intelligenetics computer analysis of the E. coli T R sequence compared to other Ca2+-binding proteins, with an EF-hand domain, allowed the identification of a single Ca 2+-binding site [1,3]. Based on oxygen donor ligands, and conserved hydrophobic/hydrophilic residues, there is a 68% sequence homology in 25 residues when compared to the first EF-hand site on calmodulin [1]. These results suggested that the electron transfer from T R to its natural electron acceptor thioredoxin depends on the Ca 2+ concentration. This new discovery implicated Ca 2+ in the regulation of at least three important processes in melanoma cells and tissues; (a), the regulation of intracellular redox [3-5]; (b), the regulation of deoxyribonucleotide biosynthesis via ribonucleotide reductase (i.e., D N A synthesis/cell division) [6-9] and (c), control of pigmentation through inhibition of tyrosinase by reduced thioredoxin [10,11]. Fig. 1 summarizes the central position of reduced

128 SOME IMPORTANTMETABOLICFUNCTIONSFOR REDUCED THIOREDOXlNIN THE HUMAN EPIDERMISCONTROLLED BY CALCIUM (POINTSOF INHIBITIONt ) MELANIN L-TYROSI~

®]

DEOXYRIB£~OTIDES

OALC,OM

NADPH NADP+ /SH Fig. ]. The production of reduced thioredoxin (T~. ) is regulated

by binding Ca2+ to TR (a). Reduced thioredoxinSH is an electron donor to ribonucleotide reductase (RR) (b) for the biosynthesis of deoxyribonucleotides(i.e., DNA synthesis) (c). Reduced thioredoxin functions as an allosteric inhibitor of tyrosinase to regulate melanin biosynthesis, and protects cells against oxidativedamage by hydrogen peroxide (d).

thioredoxin (11.4 kDa) in antioxidant defense, DNA synthesis (cell division) and pigmentation in human melanoma. Thioredoxin is an important anti-oxidant molecule and this thioprotein has a number of metabolic functions [12]. The allosteric regulation of thioredoxin reduction by Ca 2+ is reversible. Using isotopic 45Ca2+, this ion has been readily exchanged from T R by high concentrations of oxidized thioredoxin [1-3]. Therefore, while Ca 2+ regulates the reduction of oxidized thioredoxin by TR, high concentrations of this thioprotein also regulate Ca 2÷ exchange from the enzyme, thus, ensuring that the intracellular redox conditions do not become too oxidizing [2,12,13]. As a consequence, the equilibrium between oxidized and reduced thioredoxin is controlled by Ca 2÷. Recently, we have discovered that the Ca 2÷ status in primary melanomas and their surrounding epidermis can be determined using the analysis of membrane associated T R activities on tissue biopsies [14]. In these experiments, plasma membrane-associated T R activities were measured both in the presence and absence of calcium on four primary melanomas and in the surrounding epidermis in four directions around each tumor [14]. Since calcium concentration was the only controlled variable in these experiments, it was concluded that T R activity may be used directly as a measure of calcium status in the human epidermis. A

similar calcium regulation of membrane-associated TR activity has also been found on mitochondrial membranes [40]. This tissue assay yields four times more sensitivity as a prognostic indicator than the standard histologic procedure of measuring tumor thickness (i.e., Breslow level), [14,15]. In addition, it has been shown that the Ca2+/redox status of malignant melanoma metastases is very important to the activity of alkylating anti-tumor drugs, such as the chloroethylnitrosoureas [2,7,8]. Ca 2+ homeostasis in the epidermis is related to signal transduction mechanisms which control cAMPdependent processes [16]. The /3-2-adrenoceptors play an important role in the regulation of the Ca 2+ transport in human keratinocytes and in human epidermis [17-19]. Furthermore, epinephrine biosynthesis has been verified in rat skin [20]. This biosynthesis of epinephrine from norepinephrine and S-adenosylmethionine in rat skin involves an N-methyltransferase which differs from the adrenal and brain stem enzyme, since skin-PNMT is not inhibited by the specific inhibitor SKF-29661 [20]. The ultimate aim of this study was to determine how Ca 2+ flux may influence melanoma invasion and prognosis, since it has been shown that cell cultures of melanoma cells expressed /3-2-adrenoceptors [21]. Herein, we have examined the relationship between epinephrine biosynthesis, beta-2adrenoceptor expression and 45Ca2+ uptake in the human epidermis, in melanoma tissues and in cultures of human melanoma cells and keratinocytes. Materials and Methods

Enzyme assays Soluble thioredoxin reductase. Soluble TR was assayed by the method of Luthman and Holmgren [22], except that in experiments requiring the addition of calcium, E D T A was excluded from the reaction mixtures [23]. Membrane-associated thioredoxin reductase. Membrane-associated T R was assayed on 3-mm tissue punch biopsies by using a spin-labelled quaternary ammonium substrate (acetamido-2,2,6,6 tetramethylpiperidine-Noxyldimethyl ammonium bromide) which was shown to be specific for membrane-associated T R (Schallreuter and Wood, U.S. Patent 4849346 (1989)). The surfactant properties of this spin-label allows diffusion through the stratum corneum of the skin followed by a reaction at the plasma membrane outer surface without penetration into or through the membrane itself [24,25]. A small fraction of the spin-label is weakly immobilized at the plasma membrane by forming stable 'ion pairs' with negatively-charged groups. This bound fraction cannot be reduced by membrane-associated T R and, therefore, only free spin-label is reduced by the enzyme. The reduction of free nitroxide

129 radical substrate was followed in quartz tissue cells in a Bruker D-200 EPR spectrometer at 25°C (fieldset 3470 G, scan range 200 G, microwave frequency 9.71 GHz, microwave power 16 dB). Enzyme activities were determined, in the presence and absence of 6-10 -3 M Ca 2÷, on 3-mm punch biopsies. All reactions yielded linear kinetics (S.D. + 5%) under conditions of saturating free radical substrate (3.5 • 10 -3 M). Specific activities were determined as the decrease in amplitude of the nitroxide radical signal/3 mm biopsy per 10 min. Phenylethanolamine-N-methyltransferase assay. Reactions contained 100 /zl of membrane-free cell extract, 100 /xl of Tris-HC1 buffer (0.05 M (pH 7.5)) containing 5 /zmol of norepinephrine, 10 /xl of 3H Methyl-labelled S-adenosylmethionine (1.06 mCi/ mmol). Reactions were incubated at 25°C and stopped by the addition of 100 /zl of 5% trichloroacetic acid (TCA). After centrifugation at 5000 rpm for 10 min, 20 /zl of supernatant was applied to silica gel GF plates (1000 microns) and chromatographed in isopropanol/ formic acid/water (20: 1:5). Norepinephrine ( R f = 0.80), epinephrine (Rf 0.70) and S-adenosylmethionine (Rf 0.50) were well separated. 10 /zl of unlabelled carrier epinephrine (10 -3 M) was added to each spot in order to visualize [3H]methylepinephrine by UVlight. The marked epinephrine spots were removed from the plates sequentially, washed with 1.0 ml of 5% TCA and centrifuged. 500 /zl of supernatant was counted in 10 ml of Ecolite scintillation fluid in a Beckman LS 3800 counter (3H channel). Specific activities were determined as [3H]methylepinephrine (cpm)/mg protein per 5 min.

Binding assays A non-selective /3-adrenoceptor antagonist ( - ) [3H]CGP 12177 [(-)-(3t-butylamino-2-hydroxy propoxy)-[5,7-3H]benzimidazole-2-one) was used as the radioligand for all binding experiments [26] on cell cultures from human primary and metastatic melanoma and human keratinocytes. The labelled antagonist is hydrophilic, and as a consequence, circumvents cell penetration to allow binding to plasma membrane /3adrenoceptors in viable cells only [27]. All experiments were carried out in duplicate. Specific binding of ( - ) [3H]CGP 12177 to /3-2-adrenoceptors was determined by subtracting non-specific binding in the presence of the non-specific antagonist propranolol (5.0 ~mol) from the total binding in experiments with radioligand alone. Cells were incubated in a total volume of 1.0 ml culture medium (RPMI-1640) for 60 min with the radioligand to ensure saturation equilibrium. Keratinocytes were established from suction blister roofs of normal healthy skin (photo-skin type III). Preliminary kinetic experiments established that saturation equilibrium for both specific and non-specific binding (4.10 -9 M (-)-[3H]CGP 12177) occurs after

10 min incubation at 37°C and remains constant at 60 min. Reactions were terminated at 60 min by removal of the culture medium followed by two washes with 1.0 ml of 0.05 M Tris buffer containing 0.01 M MgCI 2 (pH 7.4). Cells were harvested in 1.0 ml of buffer with a rubber policeman followed by trypsination. Protein concentrations were determined against a trypsin blank by the method of Kalb and Bernlohr [28]. Aliquots of 0.2 ml were taken from each assay and counted in 10.0 ml Ecolite scintillation fluid in a Beckman LS 3800 counter on the 3H-channel. Similar experiments were conducted with human keratinocytes grown in serum free MCDB medium (0.1 • 10 -3 M Ca2÷).

Isotopic 45Ca uptake experiments Radiolabelled 45Ca2+ uptake was assayed in human melanoma cells SKmel 23 (from the Sloan-Kettering Laboratory), BC-PT-1 (primary melanoma cell line established by C. Ehrke) and amelanotic melanoma strain HT-144 (from ATCC) grown in RPMI-culture medium with 0.15- 10 -3 M calcium. The primary cell line (BCPT-1) was partially pigmented, whereas the metastatic lines were amelanotic (HT-144) and strongly pigmented (SKmel 23). Kinetics for 45Ca2+ uptake were compared as determined by cpm/mg of cell protein by the method previously reported by Schallreuter and Pittelkow [29].

Materials (-)-[3H]CGP 12177 (specific activity 55 mCi/mM) was purchased from Amersham, Buchler, Braunschweig, Germany; 45CAC12 (14.8 mCi/mg) and [3H] methyl-S-adenosylmethionine (1.06 mCi/mM) were obtained from ICN Biomedicals, Irvine, CA, USA. All other reagents were from Sigma, St. Louis, MO, USA. Results

Earlier studies established that membrane-associated TR activity is related to each individual's photoskin type and pigmentation in the normal healthy population with skin type (I) = 6.5 + 3.5, (II) = 14.6 + 3, (Ill/IV) = 20.4 + 3, (V) = 25 and (VI) = 30.6 [30]. Skin type is classified according to Fitzpatrick [31] with type I being the fairest and type VI being the darkest skin. In addition, it has been shown that highly pigmented melanotic melanomas showed a 2-10-fold increase in membrane-associated TR activity compared to normal skin from the same patients [22]. Membrane-associated TR has been assayed on primary melanomas and surrounding skin of 29 patients [14]. 24 Patients showed high enzyme activity as compared to normal skin, meanwhile five tumors showed low activity. These differences in enzyme activity were highly significant (P = 0.0001). A regression analysis of membrane-associated TR activity revealed that a change of four units of

130 enzyme activity corresponded to one unit tumor thickness according to Breslow [2,22]. These differences in membrane-associated TR activity in primary melanomas and their surrounding epidermis were shown to be the result of the influence of extracellular Ca z+ concentrations and indicated the existence of Ca 2÷ gradients on primary melanomas and their surrounding epidermis. As a consequence, membrane-associated TR activity, in the presence and absence of Ca 2÷, can be used to examine directly the influence of primary melanomas on metabolic changes in the surrounding skin of the tumor. Fig. 2 presents the analysis of TR activities measured on an 'enbloc' section over a distance of 15 cm from a primary melanoma to a lymph-node metastasis (Fig. 2a) together with parallel biopsies assayed in the presence of 6-10 -3 M Ca 2÷ (Fig. 2b). Both the primary melanoma and the lymph node metastasis yielded high enzyme activities, indicating localized low extracellular Ca 2+ concentrations around these active tumors. TR activity falls to normal skin type activity 4-5 cm away from each metastasis. Fig. 2c shows a more defined profile for Ca 2+ flux in the epidermis, by comparing the percentage of inhibition of TR by Ca 2÷. Apparent decreases in the influence of Ca 2÷ on enzyme activities directly on the tumors themselves (Fig. 2c) are probably due to less availability of free Ca 2÷ through its complexation to melanin, resulting in a lower available concentration for enzyme inhibition. Previous results showed that the specific binding of (-)-[3H]CGP 12177 to human melanoma cells is rapid with saturation of /3-2-adrenoceptors as a function of radioligand concentration [21]. Fig. 3a presents a comparative study of /3-2-adrenoceptor densities on three melanoma cell lines, two pigmented (i.e., SKmel 23 and BC-PT-1) and one non-pigmented (i.e., HT-144). These values are compared to /3-2-adrenoceptor densities on human keratinocytes (photo-skin type III). Pigmented melanoma cells express a significant number of receptors, but as shown previously, the density of/3-2-adrenoceptors on keratinocytes is considerably higher [17,18] (i.e., approx. 7000 receptors/cell). Since beta-2-adrenoceptor expression has been shown to influence intracellular Ca 2+ concentrations in human keratinocytes upon stimulation with epinephrine [19,32], 45Ca2+ has been used to compare the kinetics for transport of this ion in these three human melanoma cell lines. SKmel 23 and BC-PT-1 revealed a rapid uptake of 45CaZ+ compared to HT-144 and pigmented cells showed the highest intracellular concentration of this ion before efflux occurred (Fig. 3b). Rat skin contains an extraneuronal phenylethanolamine-N-methyltransferase (PNMT) with different properties to the neuronal enzyme [20]. The sympathetic denervation of rats by superior ganglioectomy did not influence rat skin PNMT activities [20]. Cell-

free extracts from full thickness human skin showed a rapid biosynthesis of [3H]methylepinephrine from norepinephrine and [3H]methyl-SAM. However, a slow degradation of this 'de novo' synthesized epinephrine occurred, reaching half-maximal concentration at 40 min under these experimental conditions (Fig. 4). Cell-free extracts from suction blister roofs of human

Fig. 2a

40

10

0

Fig. 2 b

i

N

m

m

m

m

40

~,30 O.

2o-

10-

I

I

I

0

I

I

2

r

I

I

I

I

4 6 8 10 distance from tumor, cm

12

14

Fig. 2c

-30 EIO0.::=

880-

-20

>

@

6o(1) t~

t-t-

"i" 4 0 -

,o

I

i

20I

I

2

I

I

4

I

~

I

i

8

I

i

10

i

i

12

i

[

I

14

Fig. 2. (a), Membrane-associated T R activity measured on 3-mm punch biopsies from a primary m e l a n o m a (0) over 15 cm to a lymph • node metastasis [15]. (b), Membrane-associated T R activity m e a s u r e d after preincubation for 10 min with 6.10 -3 M Ca 2÷. (c), The percentage inhibition of T R by 6.10 -3 M Ca 2+ compared to T R specific activity, showing the influence of Ca 2+ up to a distance of 4 cm from the primary tumor and lymph node metastasis.

131 3a 100

i

c

1

X

b

o)

b×

IA,-

v

,1:o. I

$ .o

°

0

,~.~e'~, *¢ ¢0~"; ~ J

~"

If: II{

I"-,

0

10 time (mJn)20

x

~'¢-.x

30

Fig. 3. (a),/3-2-adrenoceptor densities measured by (-)-[3H]CGP-12177 binding to human differentiated keratinocytes (K) metastatic melanoma cells (SKmel 23), primary melanoma cells (BC-PT-1) and amelanotic metastatic melanoma cells (HT-144) (binding in cpm/mg cell protein). (b), Kinetics for 45Ca2+ uptake and efflux by human melanoma cells BC-PT-1 ( x x ), SKmel 23 ( , x ,x ) and HT-144 (0 0) (measured as intracellular 45Ca2 +/mg cell protein).

MCDB medium with 0.1 • 1 0 - 3 M C a 2 + w e r e used for determination of PNMT activity [33]. Table I shows the specific activities of epidermal PNMT measured in

healthy epidermis (n = 2) (skin type III, Fitzpatrick classification) [31] and from pure human keratinocytes established from skin type III blisters in serum-free

E Q. 0

o3 !

06 X

.~_ ~4 c t-i e-

O

~/ E

~

0L

2

'~ 0 0

I

{

I

I

I

5

10

15

20

25

30

I

I

I

I

35

40

45

50

time (min) Fig. 4. Biosynthesis of epinephrine from norepinephrine and [3H]methyl S-adenosylmethionine by crude cell extracts (10.2 mg/ml) of human full thickness skin. In crude extracts, the synthesis of epinephrine is followed by its slow metabolism (assay conditions described in Materials and Methods).

132 TABLE I

PNMT-activity in human total epidermis and keratinocytes

THE BIOSYNTHETICPATHWAYFOR EPINEPHRINEINDICATINGTHE IMPORTANT POINTS OF REGULATION( , )

PNMT-activity

NH2-CH-COOH

Human total epidermis Patient 1 Patient 2

760 695

Human keratinocytes

822

Specific activity = [3H]methyl-labelled epinephrine formed (cpm)/mg of proteins per 5 min.

NH2-CH-COOH I OH2 C02

I

OH (H4-BIOPTERIN) OH L-tyrosine L-dopa

OH

dopamine

@~o2

C H 2 - N ~H I

cell-free extracts from epidermal blisters and from human keratinocyte cell cultures. PNMT activity could not be demonstrated in cell-free extracts from human melanotic and amelanotic metastatic melanoma tissues.

H-C-OH ~OCH OH

~t13

CH2-NH

3 ~®

.c H IH2-NC'-,H3

3-O-rnethyl- ~

epinephrine W.. ~, H-C-OH

Discussion

* Proceedings of a conference on diagnosis, and treatment of malignant melanoma. National Cancer Institute, Bethesda, MD, Feb. 1992.

H-C-OH J .

(~ [/0"~1

SAH SAM

(~)

The sensitivity of membrane-associated TR to allosteric inhibition by Ca 2÷ has been utilized to assess the lateral flux of calcium between cutaneous melanomas and their surrounding epidermis [14]. Ca 2÷ gradients were observed within a radius of 4 cm from primary tumors. At the present time, both the diagnosis and prognosis of primary melanomas are based on histologic methods by measuring tumor thickness [15] and the depth of invasion from the epidermis to the dermis [34]. Tumor histology has been used to define the surgical excision radius of primary melanoma as 0.5 cm; as a standard surgical practice in the USA *. Biochemical analysis using tissue TR activity is not only 4-fold more sensitive than current histologic procedures [14], but this assay also has shown the significant influence of primary melanomas on metabolic events of this tumor in the surrounding skin [14]. Until recently, an understanding of Ca 2÷ homeostasis in the human epidermis had escaped definition. The role of catecholamines in Ca 2÷ transport in skin emerged when (a), human keratinocytes and split thickness skin grafts were shown to express a high density of/3-2-adrenoceptors (i.e., approx. 7000 receptors;/keratinocyte) [17,18]; (b), catecholamines (e.g., epinephrine) caused an increase in intracellular calcium in human keratinocytes, and propranolol (a nonspecific /3-adrenergic antagonist) inhibited this intracellular calcium increase [32] and (c), epinephrine biosynthesis was discovered in rat skin via extraneuronal PNMT activity [20]. In this report, preliminary evidence for epinephrine biosynthesis in human full skin, human epidermis and in cell cultures of human keratinocytes has been

CH2-NH2 I OH2

CHO H-I_oH [~0

H

OH NH4÷

SAH SAM OH norepinephrine

epinephrine

H OH 3,4 di- OH-phenylglycolaldehyde Scheme I. The biosynthetic pathway for epinephrine from L-tyrosine. Epinephrine biosynthesis can be controlled by the regulation of tetrahydrobiopterin synthesis, an essential coenzyme/electron donor for tyrosine hydroxylase (a) and by the metabolism of norepinephrine by catechol-o-methyl transferase (e) or by monoamine oxidase (f).

demonstrated. Recent results from our laboratory have established that both PNMT and biopterin-dependent tyrosine hydroxylase activities are present in human keratinocyte cell extracts and also in extracts from human epidermal suction blisters. Taken together, these results indicate that human keratinocytes in the epidermis contain the key enzymes for epinephrine biosynthesis (Scheme I). PNMT activity could not be verified in human melanotic or amelanotic metastatic melanoma extracts, although a low expression of /3-2adrenoceptors is present on pigmented melanoma cells cultured on RPMI-1640 medium containing 10% fetal calf serum. It has been well-established that Ca 2+ represents an important second messenger in the human epidermis by controlling cell differentiation [36,37], proliferation [38], desmosome formation [38] and pigmentation [11]. The preliminary data reported herein suggest that epinephrine synthesized by keratinocytes may influence/3-2-adrenoceptor expression in the epidermis leading to increases in intracellular Ca 2+. The question remains why do pigmented melanonoma cells express /3-2-adrenoceptors but obviously do not produce epinephrine themselves? Do the surrounding cells, keratinocytes and fibroblasts, which also express /3-2adrenoceptors activity [39], regulate their own intra-

133 cellular Ca 2 +-concentrations? In the case of melanoma cells, an increase in intracellular Ca 2+ correlates with the activation of tyrosinase, the key enzyme for melanin biosynthesis [1,2,10,11]. It should be noted that L-tyrosine represents a common substrate for both the biosynthesis of the melanins and the catecholamines (Scheme I). Clearly, both of these biosynthetic pathways must be subject to fine controls in the human epidermis. A detailed understanding of the role of Ca 2+ in regulation of these two important processes needs further research. References 1 Schallreuter, K.U. and Wood, J.M. (1989) Biochim. Biophys. Acta 997, 242-247. 2 Schallreuter, K.U. and Wood, J.M. (1991) Melanoma Res. 1, 159-167. 3 Schallreuter, K.U. and Wood, J.M. (1991) in Novel Calcium Binding Proteins (Heizmann, C.W., ed.), pp. 339-360, Springer, Berlin. 4 Schallreuter, K.U., Pittelkow, M.R. and Wood, J.M. (1989) Biochim. Biophys. Res. Commun. 162, 1311-1316. 5 Schallreuter, K.U. and Wood, J.M. (1989) Free Rad. Biol. Med. 6, 519-532. 6 Holmgren, A. (1985) Annu. Rev. Biochem. 54, 237-271. 7 Schallreuter, K.U., Gleason, F.K. and Wood, J.M. (1990) Biochim. Biophys. Acta 1054, 14-20. 8 Schallreuter, K.U. and Wood, J.M. (1991) Biochim. Biophys. Acta 1096, 277-283. 9 Schallreuter, K.U. and Wood, J.M. (1989) Thioredoxin reductase in control of the pigmentary system. Clinics in Dermatology, pp. 92-105, J.P. Lippincott, Philadelphia. 10 Wood, J.M. and Schallreuter, K.U. (1988) Inorg. Chim. Acta 151, 7. 11 Wood, J.M. and Schallreuter, K.U. (1991) Biochim. Biophys. Acta, 1074, 378-385. 12 Schallreuter, K.U. and Witkop, C.J., Jr. (1988) J. Invest. Dermatol. 90, 372-377. 13 Schallreuter, K.U. and Wood, J.M. (1989) in Microbial Metabolism and the Carbon Cycle, pp. 303-323, Harwood, New York. 14 Schallreuter, K.U., J~inner, M., Mensing, H., Breitbart, E., Wood, J.M. and Berger, J. (1991) Int. J. Cancer 48, 15-19. 15 Breslow, A. (1970) Ann. Surg. 72, 902-908. 16 Carafoli, E. and Penniston, J.T. (1985) Sci. Am. 253, 70-78.

17 Steinkraus, V., K6rner, C., Steinfath, M. and Mensing, H. (1991). Arch. Dermatol. Res. 283, 328-332. 18 Steinkraus, V., Steinfath, M., K6rner, C. and Mensing, H. (1992) J. Invest. Dermatol. 98, 475-480. 19 Schallreuter, K.U., Pittelkow, M.R., Swanson, N.N. and Steinkraus, V. (1992) Arch. Dermatol. Res., in press. 20 Elayan, H., Kennedy, B., and Ziegler, M.G. (1990) Arch. Dermatol. Res. 282, 194-197. 21 Steinkraus, V., Nose, M., Mensing, H. and K6rner, C. (1990) Br. J. Dermatol. 123, 163-170. 22 Luthman, M. and Holmgren, A. (1982) Biochemistry 21, 66286633. 23 Schallreuter, K.U. and Wood, J.M. (1988) Biochim. Biophys. Acta 967, 103-109. 24 Schallreuter, K.U. and Wood, J.M. (1986) Biochim. Biophys. Res. Commun. 136, 630-637. 25 Schallreuter, K.U., Schulz, K.H. and Wood, J.M. (1986) Env. Health Perspectiv. 70, 229-237. 26 Affolter, H., Hertel, C., Jaeggi, K., Portenier, M. and Staehelin, M. (1985) Proc. Natl. Acad. Sci. USA 82, 925-929. 27 Staehelin, M., Simons, P., Jaeggi, K. and Wigger, N. (1983) J. Biol. Chem. 258, 3496-3502. 28 Kalb, V.J., Jr. and Bernlohr, R.W. (1977) Anal. Biochem. 82, 362-371. 29 Schallreuter, K.U. and Pittelkow, M.R. (1988) Arch. Dermatol. Res. 280, 137-139. 30 Schallreuter, K.U., Hordinsky, M.K. and Wood, J.M. (1987) Arch. Dermatol. 123, 615-619. 31 Fitzpatrick, T.B., Szabo, G., Seiji, M. and Quevedo, W.C. (1979) in Dermatology in General Medicine, pp. 131-163, McGraw-Hill, New York. 32 Koizumi, H., Zasui, C., Fukaya, T., Ohkawara, A. and Ueda, T. (1991) J. Invest. Dermatol. 96, 234-237. 33 Pittelkow, M.R., Zinsmeister, A. and Steinmuller, D. (1983) in Recent Advances in Bone Marrow Transplantation, pp. 227-241, Alan Liss, New York. 34 Clark, W.H., Jr., From, L., Bernardino, A. and Mihm, M.C. (1969) Cancer Res. 29, 705-712. 35 Boyce, S.T. and Harm, R.G. (1983) J. Invest. Dermatol. 81, 335-405. 36 Hennings, H. and Holbrook, K.A. (1983) Exp. Cell Res. 143, 127-142. 37 Durham, A.C. and Walton, J.M. (1982) Biosci. Rep. 2, 15-30. 38 Watt, F.M., Maney, D.C. and Garrod, D.R. (1984) J. Cell. Biol. 99, 2211-2215. 39 Eedy, D.J., Canavan, J.P., Shaw, C. and Trimble, E.R. (1990) Br. J. Dermatol. 122, 477-483. 40 Lenartowicz, E. (1992) Biochim. Biophys. Res. Communs. 184:2, 1088-1092.