High plasma Level of a Eumelanin Precursor, 6-Hydroxy-5-Methoxyindole-2-Carboxylic Acid as a Prognostic Marker for Malignant Melanoma

Hiroyuki Hara, Nicholas Walsh, Koji Yam.ada, and Kowichi Jimbow The Division of Dermatology and Cutaneous Sciences, Faculty of Medicine, University of Alberta, Canada

Melanin synthesis is a biologic property unique to the mela­nocyte. It is highly elevated in malignant melanoma with the production of both cumelanin (brown/black pigment) and pheomelanin (yellow/red pigment), dihydroxyindole (DHI) and cysteinyldopa (CD), respectively, being major precur­sors. Melanin metabolites are often released in the urine of patients with disseminated melanoma metastasis (melanuria). To establish a better method for the detection of occult mela­noma this study compares the plasma levels of a pheomelanin metabolite, S-S-CD, and a eumelanin metabolite, 6-hy­droxy-S-methoxyindole-2-carboxylic acid (6HSMI2C), in melanoma and non-melanoma patients and correlates them with tumor thickness and melanoma metastasis. We found a) that the normal plasma levels ofS-S-CD and 6HSM12C are less than 2.22 ng/ml and 1.04 ng/ml, respectively; b) that the group with the normal6HSMI2C plasma level does not have

The formation of melanin pigments (i.e., either the brown/black pigment of eumelanin or the yellow/ red pigment of pheomelanin) is a unique metabolic pathway for the melanocyte and is highly elevated in malignant melanoma (Fig 1) [1] . A significant num­

ber of reports have accumulated, indicating that the quantitation of melanin precursors/metabolites, especially a pheomelanin precur­sor, 5-S-cysteinyldopa, in the urine and plasma (serum) may be of value in the detection of melanoma metastases [2-7]. The major obstacle for measurement of 5-S-CD, either in the plasma or urine form, is that it can also be released in normal healthy subjects and after UV exposure [8-15]. Plasma dopa may be increased in some patients with advanced melanoma metastases [16]. Dopa itself has, however, a very low renal plasma clearance, because it is decarboxy­lated in the kidney, and determination of dopa in the urine is there­fore of value only in advanced stages of melanoma [17] . Instead, quantitation of the urinary excretion of the dopa metabolites, dopa­mine and 3, 4-dihydroxyphenyl-acetic acid (DOPAC) and their

Manuscript received July 12, 1993; accepted for publication December 1, 1993.

Reprint requests to~ Dr. Kowichi Jimbow, 260G Heritage Medical Re­search Centre, University of Alberta, Edmonton, Alberta, Canada T6G 252.

Abbreviations: 5-S-CD, 5-S-cysteinyldopa; DHI, dihydroxyindole; DHICA, S,6-dihydroxyindole-2-carboxylic acid; DOPA, 3,4-dihydroxy­phenylalanine; DO PAC, 3,4-dihydroxyphenyl-acetic acid; 5H6MI, 5-hy­droxy-6-methoxyindolc; 6HSMI, 6-hydroxy-5-methoxyindole; SH6MI2C, 5-hydroxy-6-methoxyindole-2-carboxylic acid; 6H5MI2C, 6-hydroxy-5-methoxyindole-2-carboxylic acid; HVA, homovanillic acid; VMA, 4-hy­droxy-3-methoxymandelic acid.

any metastasis, whereas a normal S-S-CD level is seen in both non-melanoma and melanoma patients with and without metastasis; c) that a high plasma 6H5MI2C level is seen in all melanoma patients with tumor thickness more than 3.0 mm regardless of the presence or absence of metastasis, whereas in thinn.er melanoma patients this is seen only in positive me­tastaSIS group; and d) that all melanoma patients with positive metastases showed a high plasma 6HSMI2C level (more than 1.75 ng/ml). We conclude that the measurement of plasma levels of melanin metabolites provides a method for detecting occult melanoma metastasis and estimating the prognosis of melanoma patients, plasma 6HSMI2C level being more sen­sitive and reliable than that of 5-S-CD, and its increased level being a high risk factor. Key words: melanoma/eumelanin/pheo­melanin/metabolites. ] Il1vest Dermatol1 02:501-505, 1994

methylated products, methoxytyrosine, homovanillic acid (HV A), and 4-hydroxy-3-methoxymandelic acid (VMA), was proposed for this purpo.se and tested to define the dopa turnover [18] . In spite of its potential usefulness for the diagnosis of advanced melanoma patients, the study indicated that other neural crest tumors (e.g., neuroblastoma) can also excrete DOPAC, HV A, VMA, and their conjugate in the urine [19].

Pavel et al introduced the measurement of indole metabolites in melanoma culture [20], and a number of studies have correlated the urin~ levels of indole derivatives, e.g., dihydroxyindole (DHI) and Its carboxylic form (DHICA), with melanoma progressions [21-24]. DHI and DHICA can be O-methylated in the liver as well as in melanoma cells. The O-methyl derivatives of DHI are then conjugated in the liver with glucuronic and sulfuric acid, and ex­creted in the urine as the isomeric glucuronite and sulphate conju­gates of 5-hydroxy-6-methoxyindole (5H6MI) and 6-hydroxy-5-methoxyindole (6H5MI), and a 6-0-sulphate of DHI. In contrast, DHICA and its O-methyl derivatives, 5-hydroxy-6-methoxyin­dole-2-carboxylic acid (5H6MI2C) and 6-hydroxy-5-methoxyin­dole-2-carboxylic acid (6H5MI2C), can be released in the free form (Fig 1). In a recent experimental murine melanoma study, the plasma level of 5H6MI2C plus 6H5MI2C correlated to the tumor size [25].

To develop a better method for the detection of occult melanoma and establishment of prognostic criteria, this study measures the plasma levels of indole derivatives, i.e., O-methyl derivatives of DHICA (5H6MI2C and 6H5MI2C) and a thiol derivative (5-S­CD), in melanoma and non-melanoma patients and compares their efficacy in estimating the prognosis of melanoma patients and in predicting the presence of occult melanoma metastases.

0022-202X/94/S07.00 Copyright © 1994 by The Society for Investigative Dermatology, Inc.

501

502 HAM ET AL

~COOB

..... 1 BO .... NB2

Tyrosine

l RO~COOB

,0 NB2 HO

) Tyrosinase

COMT

THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

BO~ CH30 ~N !J.-. COOH

H 5H6MI2C

CH30~ -... I I

HO N COOH H

6H5MI2C

Dra_4 ---Jl

Tau~

0~C008 O~

HO~

B~N':;""COOH H

.. ::WCOOH

DHICA oligomer. 5,6DHICA /. NB2 -+ __ o 0 ~ COOB ____

Dopaquinone Dopachrome "~ HO~ "I I BO N

H

HO~ BO~NJl..

B

EUMELANIN

5,6DHI DHI oligomer.

__ OB L k -... I S-{COOB • INTERMEDIATES PHAEOMELANIN

NB2 COOH NB2

5-S-CD

Figure L Metabolic pathway for the synthesis of melanin pigments, eumelanin, and pheomelanin. S-S-cysteinyldopa (S-S-CD) is a major precursor of pheomelanin synthesis whereas S,6-dihydroxyindole (5,6 DHI) and S,6-dihydroxyindole-2-carboxylic acid (5,6 DHI2C or DHICA) are eumelanin precur­sors. DHICA is further O-methylated to form metabolites of S-hydroxy-6-methoxyindole-2-carboxylic acid (SH6MI2C) and 6-hydroxy-S-methoxyindole-2-carboxylic acid (6HSMI2C).

MATERIALS AND METHODS

Subjects The following five groups cover a total of 88 subjects drawn from normal healthy volunteers and patients who visited the melanoma clinic at the Cross Cancer Institute at the University of Alberta: group A, 14 cO'!trol specimens from 14 normal healthy subjects without a personal or family history of cancer, including malignant melanoma; group B, 10 con­trol specimens from 10 patients with systemic development of non-mela­noma cancers; group C, 26 specimens from 21 patients who had cutaneous melanoma with no evidence of metastases (as judged by conventional rou­tine methods such as X-ray, ultrasound of abdomen, pelvis, CTscan of head, and bone scanning) at the time of plasma collection; group D, 24 specimens from 16 patients who had cutaneous melanoma with metastases in the past, but were disease free at the time of plasma collection; and group E, 14 specimens from 10 patients who had cutaneous melanoma with metastases at the time of plasma collection. Among 47 melanoma patients entered in this study, nine cases were of unknown primary melanomas, and measurement of tumor thickness was not availab le for 13 cases. Test plasma samples were obtained after the surgery, and patients were followed up every 3 months for the first 2 years and every 6 months thereafter up to 5 years. On each visit, blood samples were obtained and a chest X-ray was taken. Ultrasound studies were done on every second visit. If there was any suspicious metastatic lesion on chest X-ray and ultrasound examination, or from the finding of plasma melanin metabolites, further investigations, e.g., CT and bone scanning, were carried out. The staging of melanoma patients followed the pTNM classification of the American Joint Committee on Cancer [26]. Tumor thickness was measured by Breslow's classification [27].

Analysis of Plasma 5-S-CD Fifty milligrams of acid-washed alumina and 100 III of 2% sodium metabisulfite were applied in an Eppendorf tube (1.5 ml volume). In addition, 10 III of a-methyl-dopa (0.2I1g/ml) was included in hydrochloric acid (0.1 moljl) as an internal standard. The 1.0 ml of test plasma or 1.0 ml of 1 % ethylene diamine tetraacetic acid (EDT A) . 2 Na and 10111 of standard solution containing L-dopa and dopamine, each 0.1 I1g/ml, and S-S-CD, 0.2I1g/ml in hydrochloric acid (0.1 moljl) were then added to this tube. Catechols were absorbed onto alumina by adding 2.7 mol/I Tris-HCI buffer - 2% EDTA . 2 Na(pH 8.6) . After centrifugation, alumina was washed twice with 1 ml of water.

Catcchols were eluted with 100 111 of perchloric acid (0.4 mol/I) by mixing for 2 min. After centrifugation, 10 - 20 III of supernatant was injected into the high-performance liquid chromatograph (HPLC). Catechols con­tained in this sample were quantified from the relative ratio of peak height using dopa, dopamine, S-S-CD, and a-methyl-dopa as standards.

Analysis of Plasma 5H6MI2C and 6H5MI2C One milliliter of plasma was saturated with 100 mg of sodium ch loride and hydrochloric acid (1 moljl). The mixture was then extracted twice with 5 ml of ethyl acetate by shaking for 2 min and centrifugation. The pooled ethyl acetate extracts were evaporated at 40° C under vacuum and the residue was eluted with 200111 of mobile phase.

HPLC Chromatography A high-performance liquid chromatograph (Waters, model 600 E, Millford , MA) was employed with a Waters 460 electrochemical detector, as follows.

The AIJalysis oj 5-S-CD: The detector was set at 750 mY. Separation was achieved by a reverse-phase chromatography (11 BONDAPAK C18 lScm, Waters) at 25°C. The mobile phase contained 0.1 M phosphoric acid (1 2 g/I) , 0.1 M methanesulfonic acid (10 g/l), and 0.1 mM EDT A . 2 Na in water, pH being adjusted to 3.1. The flow rate was 1.0 ml/rnin.

Tile Analysis oj Iudoles: The detector was set at 600 mY using the same reverse-phase chromatography (11 BONDAPAK C18 25 cm, Waters). The mobile phase contained 0.1 M potassium phosphate buffer at pH 2.5 con­taining 1 mM EDT A . 2 Na and methanol (80: 20 in v /v). The column was maintained at 60°C. The flow rate was 1.0 ml/min.

RESULTS

Determination of 5H6MI2C and 6H5MI2C in Plasma Samples On HPLC chromatograms SH6MI2C and 6HSMI2C were eluted at 6.7 and 8.2 min, respectively. There were, however, unknown peaks that were eluted occasionally at the same position of SH6MI2C elution, causing the measurement of SH6MI2C to be unreliable. Further analysis was carried out only on 6HSMI2C.

Plasma Levels of 5-S-CD and 6H5MI2C in Normal Subjects and Non-melanoma Cancer Patients Plasma samples from 14 normal subjects (group A) were analyzed. The mean plasma con­centrations (±SD) of S-S-CD and 6HSMI2C were 1.71 ± 0.51 ng/ml and 0.51 ± 0.53 ng/ml, respectively. We defined the nor­mal plasma level of 5-S-CD and 6HSMI2C as less than 2 .22 ng/m1 and 1.04 ng/ml, respectively. Plasma levels of S-S-CD and 6HSMI2C in patients with non-melanoma cancer (group B) were 2.15 ± 0.66 ng/ml and 0.56 ± 0.36 ng/ml, respectively. These values were not significantly different from those of control subjects (group A) (Fig 2a,b).

VOL. 102, NO.4 APRIL 1994

a

0

b

Group A

Group A

B

B

•

•

c o E

•

•

• • •

y

C 0 E

Figure 2. Comparison of plasma level of 5-S-CD and 6H5MI2C in mela­noma patients with and without metastasis. a) 5-S-CD level: A) 1.71 ± 0.51 ng/ml in group A normal healthy volunteers; B) 2.15±0.66 ng/ml in group B patients with non-melanoma cancers; C) 3.37 ± 2.58 ng/ml in group C patients with no. metastasis in the past and the present; D) 2.65 ± 1.62 n~l.ml in group D patients With negatIve metastasIs at the present, but positive metastasis in the past; E) 5.62 ± 4.71 ng/ml in group E patients with metastasis by routine laboratory examinations (e.g., chest X-ray, ultrasound, CT scan, and bone scan). by 6H5MI2C: A) 0.51 ± 0.53 ng/ml in group A normal healthy volunteers; B) 0.56 ± 0.36 ng/ml in group B patients with non-melanoma cancers; C) 1.61 ± 2.57 ng/ml in group C patients with no metastasis in the past and the present; D) 1.81 ± 1.67 ng/ml in group D patients with metastasis in the past, but negative one at the time of blood collection; E) 4.09 ± 5.35 ng/ml in group E patients with metastases.

Tumor Thickness and plasma Level of 5-S-CD and 6H5MI2C in Melanoma Patients Group C, 0, and E mela­noma patients were analyzed. Table I summarizes the relationship between plasma levels of 5-S-CD and 6H5MI2C and the tumor thickness and positive versus negative metastases of melanoma. The definition of positive versus negative melanoma metastasis was based upon a minimum of 12-month follow-up of individual cases, with clinical observations and laboratory tests, including chest

PLASMA INDOLE AND MELANOMA METASTASIS 503

X-ray, ultrasound, CT scan, and bone scan. No significant differ­ence and correlation to the tumor thickness were noted in the indi­vidual four groups for measurement of plasma 5-S-CD level. In contrast, the plasma level of 6H5MI2C showed a significant differ­ence depending upon the tumor thickness. The melanoma groups 1 and 2 with tumor thickness less than 1.0 mm and 2.0 mm, respec­tively, did not show any increase of the plasma 6H5MI2C level. Importantly, however, there was a tendency towards a higher plasma 6H5MI2C level being in a positive metastasis group, even in those thin-melanoma groups. Group 3 and 4 participants with a tumor thickness of more than 2.0 mm and 3.0 mm, respectively, revealed significant and marked increases in the plasma levels of 6H5MI2C as compared to those of control subjects, and to groups 1 and 2 without metastases (Table I). Specifically, the latter group 4 (more than 3.0 mm) showed at least a threefold increase, as com­pared to normal control and non-melanoma cancer subjects (p < 0.001) in the plasma 6H5MI2C level and 74% of these patients had the melanoma metastases either at the time of sample collection (68%) or in the past (6%). Similarly, in group 3 (patients with tumor thickness of between 2.0 and 3.0 mm), 42% of the participants had melanoma metastases either at the time of plasma collection (25%) or in the past (17%). The findings indicate that the plasma 6H5MI2C level , but not 5-S-CD level, correlates well with the tumor thickness (in particular, > 2.0 mm), and a high risk of having metastatic melanoma .

plasma Level of 5-S-CD and 6H5MI2C, and Melanoma Stages Table II compares the plasma levels of 5-S-CD and 6H5MI2C with melanoma stages (I - IV development) . In stage I melanoma, none of the patients revealed an increased level of 6H5MI2C as compared to that of control subjects. Melanoma pa­tients in stages I - IV had a significant increase in the plasma level of 5-S-CD. However, there was no correlation between the increased 5-S-CD plasma level and melanoma stages, e.g., stage II group had a higher level than stage III. In contrast, a more significant difference was seen in the plasma 6H5MI2C level, in parallel to melanoma stages and metastases. There was no increase in plasma 6H5MI2C level in stage I patients, its mean being 0.38 ng/ml. In contrast, those patients in stages II, III, and IV who have either melanoma metastases or high tumor thickness showed a high plasma 6H5MI2c; level.. Importantly, 6H5MI2C levels of stage IV patients were 10 times higher than those of stage I patients.

plasma Levels of 5-S-CD and 6H5MI2C in Melanoma Pa­tients With and Without Metastasis The plasma levels of 5-S­CD and 6H5MI2C were compared with respect to the presence or absence of melanoma metastases (Fig 2a,b). As stated earlier, the mean plasma levels of 5-S-CD and 6H5MI2C in nOfmal control and non-melanoma patients were 1.71 and 2.15 ng/ml, and 0.51 and 0.56 ng/ml, respectively. A marked increase in the plasma level of both 5-S-CD and 6H5MI2C was seen in melanoma patients with metastases (p < 0.001), its mean 5-S-CD and 6H5MI2C levels being 5.62 and 4.09 ng/ml, respectively. Melanoma patients with metastases (group E) always showed a plasma 6H5MI2C level much higher than that of control subjects and non-melanoma cancer pa­tients, as well as those melanoma patients without metastases in the past or at the time of plasma collection, whereas the same group of patients often revealed a plasma 5-S-CD level lower than that of control. Melanoma patients in disease-free states at the time of sam­ple collection, i.e., either without metastases (group C) Of with positive metastases in the past (group D), indicated the mean value of the plasma 6H5MI2C at less than 1.8 ng/ml. When the group C patients (no metastases in the past) showed a high plasma 6H5MI2C level (> 2.0 ng/ml), they lISually revealed a high tumor thickness, more than 2.0 mm (2.0 < 3.0 mm = 40%; > 3.0 mm = 60%). In contrast, SlIch a clear correlation with the tumor thickness or the presence of melanoma metastasis was not seen in the group C pa­tients with high plasma 5-S-CD levels. One patient (a 70-year-old man; tumor thickness unknown) in group C was shown by routine chest X-ray examination to have a diffuse osteosclerotic lesion

504 HARA ET AL THE JOURNAL OF INVESTIGATIVE DERMATOLOGY

Table I. Plasma Level of 5-S-Cysteinyldopa (5-S-CD) and 6-Hydroxy-5-Methoxy-Indole 2-Carboxylic Acid (6H5MI2C) , Positive Versus Negative Metastases, and Tumor Thickness of Melanomas

5-S-CD 6H5MI2C Percent Distribution of Subjects

Negative Positive Negative Positive Number Metastasis Metastasis Metastasis Metastasis

Tumor of Negative Positive Level p Level p Level p Level p Thickness Subjects Metastasis' Metastasisb (ng/m1)' Value (ng/ml)' Value (ng/ml)' Value (ng/ml)' Value

< 1.0 11 64 36 2.24 ± 0.89 N,S'" 5.69 ± 2.53 P < 0.001' 0.41 ± 0.35 N,S'" 1.24 ± 1.81 N,S'" 1.0 < 2.0 9 67 33 2.35 ± 1.82 N,S 7.17 ± 4.21 p < 0.001 0.90 ± 1.66 N,S 0.82 ± 0.88 N,S 2.0 < 3.0 12 58 42 4.38 ± 1.91 N,S 1.78 ± 1.12 N,S 1.76 ± 2.07 P < 0.05 4.49 ± 1.65 P < 0.001 3.0 < 19 26 74 3.39 ± 1.57 P < 0.001 3.75 ± 3.44 N,S 2.38 ± 1.76 P < 0.001 2.17 ± 1.18 P < 0.001

• Percent distribution of group C melanoma patients who do not have any visible metastasis . • Percent distribution of group D and E melanoma patients. Group D patients bave a past history of melanoma metastasis, but no detectable metastasis at the time of plasma

collection as judged by routine conventional methods, e.g., chest X-ray, ultrasound of abdomen, pelvis, and CT scan of the head, and Gc-bone scintigraphy. Group E melanoma patients have metastasis at the time of plasma collection.

, ng/ml of mean ± SD. J Not significant in t test. . 'Comparison with normal control subjects, their mean ± SD being 1.17 ± 0.51 ng/ml and 0.51 ± 0.53 ng/ml in 5-S-CD and 6H5M12C, respectively.

throughout the bony thorax due to melanoma metastasis. The plasma 5-S-CD and 6H5MI2C levels were measured and all three measurements showed a normal value of 6H5MI2C (0.44-0.52 ng/ml) but an increased level of 5-S-CD (2.47 -4.44 ng/ml). The patient was subsequently found to have a bone metastasis of a prostate carcinoma, but none of melanoma metastasis, and died 2 years later of prostate carcinoma metastasis. In group D, with the exception of one case, the patients who had a plasma 6H5MI2C level higher than 2.0 ng/ml showed a tumor thickness of more than 3.0 mm, indicating that they are at high risk for metastases. Three of the group D patients showed lung metastases 3 months later, even though the tumor thickness of their primary melanoma was less than 0.75 mm and there was no visible melanoma metastasis on chest X-ray examination at the time of plasma collection. One patient in group E (a 47-year-old man; 1.35-mm tumor thickness) developed a tumor mass at the site of skin graft and lymph node swelling 2 years after melanoma excision. Three repeated biopsy reports of both subcutaneous tumors and lymph nodes indicated the simple fibrosis with negative immunostaining of all of the mela­noma markers tested. However, of the two, the plasma 5-S-CD and 6H5MI2C levels revealed a marked increase. Importantly, the rou­tine laboratory investigations, including chest x-ray, ultrasound of abdomen, CT scan of head, and bone scan, failed to demonstrate the presence of melanoma metastases. However, the plasma level of 6H5MI2C (21.74 ng/ml) was much higher than that of 5-S-CD (9.35 ng/ml), and the patient showed (3 months later) visible pul­monary and brain melanoma metastases on chest X-ray and CT scan; he died 6 months later. These findings may again indicate that the quantitation of plasma 6H5MI2C levels may be more reliable than that of plasma 5-S-CD levels. A patient with a high plasma 6H5MI2C level most likely has melanoma metastasis even though

it has not yet been identified by conventional, routine tests , e.g. , chest X-ray and ultrasound.

DISCUSSION

Among various melanin metabolites, 5-S-CD, a major pheomelanin precursor, is a marker that has been used commonly for analysis of melanoma metastasis. Our previous HPLC study, however, intro­duced the importance of measuring the urinary indole eumelanin metabolites, a combination of 6H5MI2C and 5H6MI2C, plus the 5-S-CD pheomelanin measurement rather than the latter alone [22]. The urinary and plasma levels of 5-S-CD in normal healthy subjects have been shown to possess large physiologic variations [8,17], often resulting in significant false results in the analysis of melanoma metastasis. 5-S-CD may not just be a precursor of ph eo­melanin. Along with others, we have previously reported the pres­ence of free and protein-bound 5-S-CD in non-melanocytic organs and tissues in mice, rats, and guinea pigs [28 ,29]. In addition, after UV exposure new melanin pigmentation occurs and there is in­creased 5-S-CD excretion [12-15]. 5-S-CD is excreted in patients suffering from both tyrosinase-positive and negative albinisms, the latter of which does not have any melanin synthesis [30].

In this study we found the normal range of the plasma 6H5M 12C concentration to be 0.07-2.15 ng/ml (0.32-9.90 nmol/I; mean value, 0.51 ng/ml, 2.35 nmol/I). This mean value was slightly lower than that (3 .6 nmol/I) reported recently in normal Oriental subjects [24]. Importantly, we found that the plasma 6H5M12C level correlates well with the tumor thickness, melanoma stage, and melanoma metastases. The increased plasma 6H5M12C concentra­tion was seen in melanoma patient groups with tumor thickness of more than 2.0 mm and 3.0 mm, and in stages II, III, and IV, indicat­ing that cases with a high plasma 6H5MI2C level are at high risk for

Table II. Plasma Levels of 5-S-CD and 6H5MI2C in Different Stages of Melanoma Patients

Percent Distribution of Subjects 5-S-CD 6H5MI2C

Melanoma Number of Metastasis Metastasis Metastasis Plasma p Plasma p Stage Subjects None' Pasrb Present< Level' (ng/ml) Value Level' (ng/ml) Value

I 8 100 0 0 2.18 ± 1.00 <0.021' 0.38 ± 0.34 NS'J II 18 100 0 0 3.50 ± 1.98 < 0.001 1.38 ± 1.55 <0.05 III 15 13 87 0 2.59 ± 1.73 < 0.02 1.97 ± 1.13 < 0.02 IV 30 0 47 53 4.54 ± 3.96 < 0.01 3.26 ± 3.97 < 0.01

• Group C melanoma patients with no metastasis. • Group 0 melanoma patients with metastasis in the past, but no metastasis detectable at the time of sample collection as judged by routine diagnostic protocols of chest X-ray,

ultrasound of abdomen, pelvis, and scalp, Ga-bone scintigraphy. , Group E melanoma patients with metastases at the time of sample collection. J ng/ml of mean ± SD. t N ot significant in t test. f Compared with normal healthy subjects, their means ± SD being 1.71 ± 0.52 ng/ml and 0.51 ± 0.53 ng/ml in 5-S-CD and 6H5MI2C, respectively.

VOL. 102, NO.4 APRIL 1994

the development of metastases, a visible metastasis, or a hidden non-visible melanoma metastasis that has not yet been identified by routine diagnostic methods.

In this study, 6HSMI2C was chosen over 5H6MI2C to measure the plasma indole metabolites as a melanoma-metastasis marker be­cause the HPLC peak of 5H6MI2C was masked by unknown peaks and the ratio of absorption peaks of 5H6MI2C and 6H5MI2C was approximately equal. A number of eu- and pheomelanin-related metabolites were recently measured in the urine of people with various capacities for melanin synthesis. 5H6MIC was found to be the best urinary marker of melanin pigmentation in the skin [31] . In that study, however, 6HSMI2C was not tested, though S-S-CD and other indole metabolites were compared. Another recent study, which measured the plasma and urinary levels of 5-S-CD and 6HSMI2C plus 5H6MI2C in B16-melanoma-bearing mice, indi­cated an excellent correlation between the plasma concentration of SH6MI2C plus 6H5MI2C and 5-S-CD, whereas the urinary excre­tion of 5H6MI2C plus 6H5MI2C and 5-S-CD showed no signifi­cant correlation [25]. The enzyme catechol O-methyltransferase has a broad spectrum of physiologic and pharmacologic roles in humans [32,33]. In melanoma cells, this O-methylation prevents the formation of highly reactive O-quinones and, hence, the production of indole metabolites. Therefore, 6H5MI2C and SH6MI2C could be a part of the mechanisms to protect the cells against their own toxic products [20]. A high plasma 6H5MI2C level in certain melanoma patients who did not have any melanoma metastasis may be attributable to this antioxidant process.

The diagnosis of occult melanoma metastasis is a serious medical problem. Spontaneous regression of primary cutaneous melanoma occurs fairly frequently and 3.6-15.0% of melanoma patients seek their medical advice for unknown primary lesions [34] . There is no tumor thickness of primary melanoma available. Even when a sec­ondary lesion is removed, it is not presently possible to determine whether these patients have another hidden secondary melanoma. Our measurement of the plasma eumelanin metabolite 6H5MI2C, combined with the 5-S-CD pheomelanin measurement would likely provide a sensitive method for this purpose and therefore should not lose its usefulness. This measurement would also help to judge the effect of chemotherapeutic and other treatment mo­dalities.

This study has bem supported by gra II tsfrom MRC-Ullillersity/llldustry Program (UI0049/11195) mId Ncr (1497) of Callada. Hiroyuki Hara is a visiting research fellolll from Ihe Departlllelll of Dermatology, NiI,on Utliversity School of Medicine, Tokyo,Japan . Nicholas Walsh is a Terry Fox Physiciall/Scielltistfellow.

Our appreciation to Drs. Stewart Adams and MillOko Jilllbow for the collectioll of patiwts' plasma alld for providing clinical data.

REFERENCES

1. Jimbow K, Salopek TG, Dixon WT, Searles GE, Yamada K: The epidermal melanin unit in the pathophysiology of malignant melanoma. Alii] Derlllato­patilo/13:179-188,1991

2. Ito S, Wakamatsu K, Horikoshi T: Serum 5-S-cysteinyldopa as a biochemical marker of melanoma: basic and clinical studies (abstr). Pigmelll Cell Res 6:314, 1993

3. Carstam R, Brinck C, Formstedt B, Rorsman H, Rosengren E: 5-S-cysteinyldopa in human urine. Acta Derlllatol Venereol (Stockil) 70:373-377, 1990

4. Takahashi H, Fitzpatrick TB: Quantitative determination of dopa; its application co measurlllcnt of dopa in urine and in the assay of tyrosinase in serum.] Invest Dem,ato/42:161-165, 1964

5 . Faraj BA, Lawson DH, Nixon OW, Murray DR, Camp WM, Ali FM: Melanoma detection by enzyme-radioimmunoassay of L-DOPA, dopam.ine, and 3-0-methyldopamine in urine. Clill Cirelli 27:108-112, 1981

6 . Hara H, Chino H, Kawanami T, Sameshima T, Morishima T: 5-S-cysteinyldopa in urine and tumor.] Dermato/19:806-808, 992

7. Peterson LL, Woodward WR, Fletcher WS, Palmquist M, Tucker MA, Ilias A: Plasma 5-S-cysteinyldopa differentiates patients with primary and metastatic melanoma from patients with dysplastic nevus syndrome and normal subjects. ] Am Acad Dermato/19:509-511, 1988

PLASMA INDOLE AND MELANOMA METASTASIS 505

8. Agrup G, Falck B, Kennedy B-M, Rorsman H, Rosengren A-M, Rosengren E: Dopa and 5-S-cysteinyldopa in the urine in healthy humans. Acta Dermatol Verrereol (Stock/,) 53:453-454, 1973

9. Agrup G , Hansson C, Juldorf F, Rorsman H, Rosengren A-M, Rosengren E: Excretion of 5-S-cysteinyldopa in humans with genetically dark skin. Acta Derlllatol Veuereol (Stock/,) 59:69-70, 1979

10. Wirestrand L-E, Hansson C, Rosengren E, Rorsman H: Melanocyte metabolites in the urine of people of different skin colour. Acta Dermatol Vwereol (Stockil) 65:345-347,1985

11. Rorsman H , Agrup G, Falck B, Rosengren A-M, Rosengren E: Exposure to sunight and urinary excretion of 5-S-cysteinyldopa. In: Riley V (cd.). Pigment Cell (2). Karger, Basel, 1976, pp 284-289

12. Hansson C, Rorsman H, Rosengren E, Tegner E: 5-S-cysteinyldopa and dopa in serum during treatment with 8-methoxypsoralen and UV A light. Acta Dermatol Venereol (Stock/,) 61 :251- 255, 1981

13. Hansson C, Wirestrand L-E, Aronsson A, Rorsman H, Rosengren E: Urinary excretlon of 6-hydroxy-5-methoxyindole-2-carboxylic acid and 5-S-cysteinyl­dopa during PtJVA treatment. Pirotoderlllatology 2:52 - 57, 1985

14. Ito S, K.,to T, JUJua K: Seasonal variation in urinary excretion of 5-S-cysteinyl­dopa in healthy Japanese. Acta Derlllatol Venereol (Stockh) 57:163-165 1987 '

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

Stiemer U, Rosdhall, Augustsson A, Kagedal B: Urinary Excretion of 5-S-cys­temyldopa m relation to skill type, UVB-indueed erythema, and melanocyte proliferation in human skin.] In,,est DennatoI91:506-510, 1988

Hamson C, Edholm LE, Agrup G, Rorsman H, Rosengren A-M, Rosengren E: The .quantltatlve determmatlO.n of 5-S-cysteinyldopa and dopa in normal serum and 10 serum from panents wlth malignant melanoma by means of high-pres­sure ILqUld chromatography. Clill Chelll Acta 88:419-427,1978

Rorsman H, Agrup G, Hansson C , Rosengren E: Biochemical recorders of malig­nant melanoma. In: Mackie RM (ed). Piglllellt Cell (b). Karger, Basel, 1983 pp 93-115 '

Goodall MC, Alton H: Metabolism of 3,4-dihydroxyphenyl-alanine (L-dopa) in human subjects. Biodrem Plrarmaco/21:2401-2408, 1972

Ito S, Kato T, Maruta K, Jimbow K, Fujita K: "Total" acidic metabolites of cate~holamines in urine as determined by hydriodic acid and hydrolysis and ILqUld chromatography: applLcanon to patients with neuroblastoma and mela­noma. Clill Chem 31:1185-1188,1985

Pavel S, Muskiet FAJ, de Ley L, The TH, van der Slik W: Identification of three indolic compounds in a pigmented-melanoma cell-culture supernatant by gas chromatography-mass spectrometry.] Callcer Res 105:275-279, 1983

Pavel S, Elzinga H, Muskiet FAJ, SmitJM, Mulder NM, Scharaffordt KH: Eume­lanin-related indolic compounds in the urine of treated melanoma patients.] Clitl Chelll Clin BiodreIll24:167-173, 1986

Yamada K, Walsh N, Han H, Jimbow K, Chen H, Ito S: Measurement of cl1mdanin precursor metabolites in the urine as 3 new marker for detection of occult melanoma metastases. Arch DermatoI128:491-494, 1992

Kagedal B, Lenner L, Astrand K, Hasson C: The stability of5-S-cysteinyldopa and 6-hydroxy-5-methoxymdole-2-carboxylic acid in human urine. Pigmetlt Cell Res (suppl) 2:304-307, 1992

Wakamatsu K,Ito S, Horikoshi T: Nonnal values of urinary excretion and serum concentratlon of 5-S-cystemyldopa and 6-hydroxy-5-methoxyindole-2-car­boxylLc aCId: blOchetnlcal markers of melanoma progression. Melanoma Res 1:141-147,1991

Wakamatsu K, Ito S, Fujita K: Production, circulation and excretion of melanin­related metabolites in B-16 melanoma-bearing mice. Acta Dermatol Verrereol (Stock/,) 70:367 - 372, 1990

Ameri~an Joint Committee on Cancer: Mam .. 1 Jar Stagitlg oj Catlw, 3rd ,d. J.B. LlPPlllcott, Phlladelphia, 1988, PI' 139- 144

Breslow A: Thickness, cross-sectional areas, and depth of invasion in the progno­Sts of cutaneous melanoma. A,,,, Slirg 172:902-908,1970

Ito S,Jlmbow K: Quantitative analysis of eumelanin and pheomelanin in hair and melanomas.] Invest Dermalol 80:268 - 272, 1983

Ito S, Jimbow K, Kato T, !Gyota M, Fujita K: Protein-bound dopa and 5-S-cys­temyldopa 10 non-melanogenic tissues. Acta Dermatol Vetlereol (Slockh) 63:463-467, 1983

Nimmo JE, HunterJAA, Percy-Robb IW, Jay B, Phillips CIL, Taylor WOG: Plasma 5-S-cyste1Oyldopa concentrations in oculocutaneous albinism. Acta Dermatol Venereol (Stock/,) 65: 169-171, 1985

WesterhofW, Pavel S, Kammeyer A, Beusenberg FD, Cormane R: Melanin-re­lated metabolites as markers of the skin pigmentary system.] [,west Dermatol 89:78-81,1987

Grossman MH, Creveling CR, Rybozynski R, Braverman M, Isersky C, Breake­field XO: Soluble and particulate forms of rat catechol-O-methyltransferase dLStlllguLShed by gel electrophoresis and immune fixation. ] Nel/rocilem 44:421-432,1985

Smit NPM, Pavel S, Kammeyer A, Westerhof W: Determination of cMechol O-methyl-transferase activity in relation to melanin metabolism using high­performance liquid chromatography with Iluoromerrie detection. Atlal Bio­chern 190:286-291, 1990

Kopf A W, Bart RS, Rodriquez-Sains RS, Ackerman AD: Occult primary malig­nant melanomas and spontaneous regression of malignant melanomas. In: Kopf A W etal (ed.). Maliglla"tMela"o",a . Masson Publishing USA, New York, 1979, pp 166 - 168