Shoshuke Ito

Fujita Health University School of Health Sciences
Toyoake, Aichi 470-11, Japan

The incidence of malignant melanoma is increasing worldwide and Japan is no exception. Three hundreds Japanese died of melanoma in 1985 and the number had doubled in the past 20 years (1). The diagnosis and treatment at the early stage are important for patients with malignant melanoma. Furthermore, the progression of melanoma is critical for patients at the advanced stages. Thus, tumour markers that sensitively detect metastasis have long been awaited.
Two types of melanin pigments, the black eumelanin and the reddish-brown pheomelanin, are produced not only in melanocytes but also in melanoma cells (2). They are formed from 5,6-dihydroxyindoles or cysteinyldopas (CD) through the tyrosinase oxidation of tyrosine in the absence or presence of cysteine, respectively. Major portions of these melanin precursors may be oxidized to yield the melanin pigments. However, minor portions may be leaked into the blood stream, partly O-methylated in the liver, and excreted into the urine. Thus, it would be possible to estimate the progression of melanoma by measuring the concentrations of these melanin-related metabolites in blood or urine (3). In fact, the urinary excretion of the major isomer of cysteinyldopas. 5-S-CD, has been used as a biochemical marker of melanoma in some countries, especially in Sweden (4). On the other hand, clinical significance of the indolic metabolites such as 5(6)-hydroxy-6(5)-methoxyindole-2-carboxylic acid (5H6MI2C and 6H5MI2C) has also been suggested (5).
Several years ago, we started comprehensive studies to evaluate the clinical significance of these melanin-related metabolites as markers of melanoma progression (for our latest publication, see ref. 6). In this "discussion", the recent progress in this field and my view regarding the clinical significance of 5-S-CD and 6H5MI2C in melanoma patients will be summarized briefly.

Brief history of 5-S-CD as a marker of melanoma : the high level of urinary excretion of 5-S-CD in patients with melanoma was first reported in 1973 and since then many groups have studied its clinical significance. The Swedish group has published a paper summarizing the results on 571 melanoma patients (4). Of 161 patients with metastases 60% showed values exceeding the pathological level of 400 µg/day (1.3 µmol/day) while less than 2% of 410 patients without metastases exceeded this level. In recent years 5-S-CD in plasma (or serum) has drawn more attention. Significantly higher levels of plasma 5-S-CD were found not only in patients with extraregional metastases (stage IV) but also in those with regional lymph node metastases (stage III) (7).

5-S-CD genesis in melanocvtes and other normal cells : cysteinyldopas are precursors of pheomelanin. However, the presence of cysteinyldopas does not necessarily indicate pheomelanogenesis, as will be discussed below. During the sunny season urinary excretion of 5-S-CD is significantly increased. Thus, the mean value of excretion in the Swedish people in summer was threefold greater than that in winter (8). This has been considered as a major limitation of 5-S-CD in following up melanoma patients. We have also observed a seasonal variation in the excretion of 5-S-CD in Japanese (9). It should be stressed, however, that the degree of variation in Japanese was not so pronounced as in people living in Sweden. The finding that albino mice excreted as much 5-S-CD as black mice led us to speculate the extramelanocyte origin of cysteinyldopa genesis (10). We have found substantial amounts of dopa and 5-S-CD in the hydrolysates of non-melanogenic tissues such as liver and kidney from mice and have thus indicated that 5-S-CD in normal subjects might be derived mostly from the protein-bound 5-S-CD that is produced by non-specific oxidation taking place on the tyrosine residue in proteins (10). In accordance with this, plasma 5-S-CD concentrations in both tyrosinase-positive and negative albino patients were found to be almost identical to the control values (11). It has also been shown that urinary excretion of 5-S-CD is independent of skin color (12). Furthermore, no correlation was found between the basal 5-S-CD excretion and skin type or number of melanocytes, and subjets with skin type II had a significantly higher 5-S-CD excretion than those with skin type III-IV after UV-B exposure (13). These results led the authors to suggest that the increase in 5-S-CD excretion during UV irradiation is due to UV damage of melanocytes rather than to the stimulation of melanogenesis (13). It now appears that Japanese are less vulnerable to sun exposure than fair-skinned Caucasian and have thus the serum (plasma) 5-S-CD levels less variable throughout a year. Our ongoing research is in support of this view.

5-S-CD genesis in melanoma cells : next problem to be considered is the production of high levels of 5-S-CD in eumelanic melanoma. B16 melanomas contain almost pure eumelanin, yet they produce as much 5-S-CD as the eumelanin-related metabolite 6H5MI2C (14). Two explanations seem to be possible. One hypothesis is that the tyrosinase leaked from melanosomes catalyzes dopaquinone production in the cytoplasm leading exclusively to cysteinyldopa genesis. In fact, the soluble fractions from B16 and Harding-Passey melanomas contained greater concentrations of 5-S-CD than the melanosome fractions (15). The other hypothesis is that the defect of membrane in abnormal and incomplete (aberrant) melanosomes found in melanomas facilitates leakage of dopaquinone into the cytoplasm (16). Whatever the mechanism of cysteinyldopa genesis, these results suggested that although 5-S-CD has no value as an indicator of melanogenesis in normal melanocytes, it is a good indicator of the tyrosinase activity in melanoma cells and thus of the progression of melanoma (Table).

6H5MI2C as a marker of melanogenesis : 5,6-dihydroxyindole (5,6DHI) and its carboxy derivative (5,6DH12C) are intermediates of eumelanogenesis. However, these dihydroxyindoles cannot be used as biochemical markers, because they are extremely labile to oxidation and most of the indoles are O-methylated either in melanoma or liver (5). Thus, the O-methyl derivatives of 5,6DHI and 5,6DHI2C may instead serve as markers reflecting the degree of eumelanogenesis or the progression of melanoma.
In contrast to 5-S-CD, the eumelanin-related metabolite 6H5MI2C (or 5H6M12C) has been shown to reflect well the skin type and hair color (17). There have been no reports studying the seasonal variation in the levels of these indoles. However, it has been shown that after PUVA treatment the urinary excretion of both 5-S-CD and 6H5MI2C elevated several-fold and paralleled with each other (18). These results suggest that 6H5M12C can be used as a marker of melanogenesis.

Comparison of 5-S-CD and 6H5M12C as melanoma markers : we estimated normal
values from 33 Japanese. The mean values of urinary excretion of 5-S-CD and 6H5MI2C were found to be both ca 0.4 µmol/day and those of serum concentrations ca 4 nmollL (6). Levels of these markers in urine were much more variable than those in serum. Considering the wide variation of these values, we have adopted 1.5 µmol/day and 10 µmol/L as the upper limits of normal values for both markers. The urinary excretion of both 5-S-CD and 6H5MIC2 decreased in elderly subjects, while their serum concentration showed no age-dependent differences. This might be ascribed to the decrease of renal clearance associated with aging.
In collaboration with Dr. T. Horikoshi, Dept of Dermatology, Sapporo Medical College, we have been carrying out since 1989 a comparative study to determine which of the 4 markers (Table) reflects most sensitively the progression of melanoma (19). Among more than 30 patients with primary or metastatic melanoma, eigth had advanced to the stage IV and eventually died. In those 8 patients serum 5-S-CD values had been elevated to the pathological level before or when the metastases were detected clinically. In 7 of the 8 cases the elevation of 5-S-CD was more pronounced in serum than in urine. 5-S-CD and 6H5MIC2 levels in the remaining patients without metastasis were within normal ranges. Serum 6H5MI2C exceeded the normal range only at the end stage in 2 patients. Urinary excretion of 6H5MI2C did not reflect the progression of disease.

HPLC measurements of 5-S-CD and 6H5M1C2 : the concentration of -S-CD in urine or serum has usually been determined by HPLC with electrochemical detection. With the original method using alumina to extract 5-S-CD from serum (20), it was not possible to remove 3,4-dihydroxypenylacetic acid (DOPAC) which is eluted out much slower (52 min) than 5-S-CD (12 min). We have recently developed a new method to overcome this difficulty by washing alumna with an acidic buffer prior to the elution of 5-S-CD with perchloric acid (to be published). It has now become possible to analyze more than 20 serum or urine samples within 24 hrs using HPLC equipped with an automated sampler. This can be compared to the on-line automated HPLC procedure for the determination of 5-S-CD in urine (21).

Conclusion and perspective
It now became apparent that 5-S-CD in serum reflects the progression of melanoma more sensitively than in urine. It is possible to expect that serum 5-S-CD gives more direct informations on melanogenic activity of melanoma. Urinary 5-S-CD is, on the contrary, more influenced by metabolic activities, such as oxidation, 0-methylation, and cojugation, that may be taking place in tissues other than melanoma. Furthermore, it should be stressed that serum sample is much easier to collect than 24-hr urine sample. This is particularly important for following up outpatients with suspected melanoma metastases. We have just started a comprehensive and collaborative project involving dermatology departments of more than 60 universities and cancer institutes in Japan to evaluate in melanoma patients the clinical significance of serum 5-S-CD as a marker of 1) the progression of disease (detection of metastasis); 2) the efficacy of therapy; 3) the prognosis.

References

01. Kukita A, Ishikawa K: J Invest Dermatol 92:210s-213s, 1989
02. Prota G: J Invest Dermatol 75:122-127, 1980
03. Rorsman H, Agrup G, Hansson C, Rosengren C: In Pigment Cell, Vol. 6, Basel, Karger, pp. 93-115, 1983
04. Agrup G, Agrup P, Andersson T, et al: Acta Dermatol Venereol 59:381-388, 1979
05. Pavel S, Erzinga H, Musket FAJ, et al: J Clin Chem Clin Biochem 24:167-173, 1986
06. Wakamatsu K, Ito 8, Horikoshi T: Melanoma Res 1:141-147, 1991
07. Petersson LL, Woodward WR, Fletcher WS et al: J Am Acad Dermatol 19:509-515, 1988
08. Rorsman H, Agrup G, Falck B, et al: In Pigment Cell, Vol. 2, Basel, Karger, pp. 284- 289, 1976
09. Ito S, Kato T, Fujita K: Acta Dermatol Venereol 67:163-165, 1987
10. Ito 8, Jimbow K, Kato T, et al: Acta Dermatol Venereol 63:463-467, 1983
11. Nimmo JE, Hunter JAA, Percy-Robb IW, et al: Acta Dermatol Venereol 65:169-171, 1985
12. Wirestrand LE, Hansson C, Rosengren C, Rorsman H: Acta Dermatol Venereol 65:345-348, 1985
13. Stierner U, Rosdahl I, Augustsson A, Kagedal B: 3 Invest Dermatol 96:506-510, 1988
14. Wakamatsu K, Ito 8, Fujita K: Acta Dermatol Venereol 70:367-372, 1990
15. Ito S, Homma Y Kiyota M, et al: J Invest Dermatol 80:207-209, 1983
16. Borovansky J, Mirejovsky P, Riley PA: Neoplasma 38:393-340, 1991
17. Westerhof W, Pavel 8, Kammeyer A, et al: J Invest Dermatol 89:78-82, 1987
18. Hansson C: Acta Dermatol Venereol Suppl 138:1-60, 1988
19. Ito S, Wakamatsu K, Horikoshi T, et al: 3rd ESPCR Meeting, Abstract p. 30, 1991
20. Ito 8, Kato T, Maruta K, et al: J Chromatogr 311:154-159, 1984
21. Kagedal B, Kallberg M, Arstrand K, Hansson C: J Chromatogr 473:359-370, 1989

Table : Comparison of 5-S-CD and 6H5MI2C in serum and urine

^a : in Japanese
^b : no pretreatment required
^c: for repeated injections