INTRODUCTION

The epithelioid and spindle cell nevus was first described in 1948 by Spitz and named juvenile melanoma (1). It is a benign melanocytic lesion found predominantly in children and adolescents and that can be identified reliably in many cases by previously well-described major and minor histopathological criteria (2, 3, 4, 5). However, there are subsets of Spitz nevi and of malignant melanoma that closely resemble each other and represent diagnostic challenges. Additionally, some lesions are only partially excised (shave biopsies), thus often preventing the assessment of the lateral edges and deep components. The importance of distinguishing between these two entities lies in their different prognosis and treatment. Multiple studies using several approaches (immunohistochemistry, flow cytometry and molecular analysis) have been performed to identify potential differences between Spitz nevi and melanomas, but the results have been controversial or difficult to apply to routine material (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38). S100 protein is not a single protein; instead there are a group of S100 proteins with diverse functions. The S100 protein family, a subclass of low molecular weight Ca2+-binding proteins, regulate a variety of cellular processes via interaction with different target proteins in a calcium-dependent manner. They are of interest because of their involvement in neoplastic processes and their genes are clustered in chromosome 1q21. Chromosome 1 is frequently altered in several tumors, including melanomas (28, 39, 40). The most widely known commercially available antibody “anti-S100” mainly reacts with cells containing S100B polypeptide chains. This antibody was used as an early marker for melanocytic neoplasms and remains sensitive but not specific (41, 42, 43). Recently, monoclonal antibodies have become available that mainly react with S100aa and have identified several variants of the S100a chains (28). The expression of different S100 proteins (A2, A6, and A8/A9 or A12) was examined in Spitz nevi, melanomas, and melanocytic nevi cases to test the hypothesis that their expression differs among these entities and may contribute to the distinction between these entities.

MATERIALS AND METHODS

The cases were obtained from the files at the Dermatopathology Division of the New York Presbyterian Hospital-Cornell University Weill Medical College, between 1997 and 2002, based on the availability of adequate tissue material for study. Forty-two cases of Spitz nevi, 105 cases of melanomas, and 73 cases of different types of melanocytic nevi were selected. The diagnosis was made on multiple H&E routine-stained sections before this study. After review of newly cut routine microscopic sections, additional unstained sections were cut, deparaffinized, pretreated with pepsin (0.25% w/v, pH 2.0) for 5 minutes at 45° C, and then reacted with the monoclonal antibodies: anti-S100A2, anti-S100A6, MAC387, and HMB45 (Table 1). A streptavidin-alkaline phosphatase detection system with a red chromogen (Vector Lab) was used. The reaction time for the red chromogen was 20 minutes. Two investigators verified the results of the immunostains without references to the original diagnosis, although in some cases the diagnosis was obvious. The staining was evaluated semiquantitatively and read as positive (strong or weak) and negative. In the cases regarded as strongly positive, staining was observed in >50% of the lesional cells whereas the weakly, positive cells were between 1 and 50% positive. The negative cases showed <1% of positive lesional cells. Appropriate positive and negative controls were also included.

TABLE 1 Monoclonal Antibodies Used in this Study
Full size table

The χ2 test was used for comparison of groups. A P value of <.05 was considered significant.

RESULTS

All Spitz nevi cases (42/42) strongly expressed S100A6 protein in a diffuse pattern, regardless of the subtype (compound, dermal, or junctional; Figs. 1, 2). The melanocytic cells in all Spitz nevi cases did not react with anti-S100A2 or MAC387 antibodies. However, the keratinocytes in the epidermis overlying Spitz nevi reacted with anti-S100A2 antibody in some cells in the basal layer and with the MAC387 antibody in some cells in the upper third of the epidermis. The HMB45 antibody was weakly positive on the junctional component, whereas the dermal component usually was negative with only scattered positive cells.

FIGURE 1
figure 1

Spitz nevus. This photograph shows a strong and uniform reactivity with the monoclonal antibody against S100A6 protein in the nevus cells.

Full size image
FIGURE 2
figure 2

Spitz nevus. As in Figure 1, this photograph shows a strong and uniform reactivity with the monoclonal antibody against S100A6 protein in the nevus cells.

Full size image

S100A6 protein was expressed in 35 of 105 melanomas (33%), 30 of them with weak positivity (86%). The distribution of positive cases in the melanoma subtypes was as follows: 10 of 12 nodular (83%), 3 of 7 metastatic (43%), 16 of 54 superficial spreading (30%), 2 of 7 lentigo maligna (28.5%), 3 of 21 in situ (14%), and 1 acral (Figs. 3, 4). Only 3 nodular melanomas and 1 superficial spreading and 1 acral melanoma were strongly positive. The expression in the positive melanoma cases was patchy throughout the tumor and negative or minimal in the junctional component. Three desmoplastic melanomas were negative. In all melanoma cases, the melanocytes did not stain with anti-S100A2 or MAC387 antibodies. However, S100A2 protein was expressed in the keratinocytes located in the basal or parabasal layer of the epidermis adjacent to melanoma cells and MAC387 antibody was often observed in the full thickness of the overlying epidermis. HMB45 was positive in some of the cells in the melanoma cases, except for the dermal component of the desmoplastic ones.

FIGURE 3
figure 3

Melanoma in situ. The melanoma cells in the epidermis are negative for anti-S100A6 antibody. Some dermal dendrocytes and endothelial cells are positive, which represent internal positive controls for the adequacy of the reaction.

Full size image
FIGURE 4
figure 4

Melanoma, superficial spreading type. The melanoma cells in the epidermis and dermis are negative for anti-S100A6 antibody.

Full size image

S100A6 protein was expressed in 41 of 73 melanocytic nevi (56%), 40 of them with weak positivity (98%). The distribution of positive cases in different melanocytic nevi was as follows: 22 of 25 nevus with architectural disorder (88%), 6 of 12 combined nevi (50%), 3 of 6 blue nevi (50%), 5 of 13 compound nevi (38%), 4 of 12 intradermal nevi (33%), and 1 of 3 acral nevi (33%; Fig. 5). Only 1 nevus with architectural disorder was strongly positive. The immunoreactivity in the positive cases was restricted to the deep dermal component. Two pigmented spindle cell nevi were negative. Neither anti-S100A2 nor MAC387 antibodies reacted with the nevus cells. However, anti-S100A2 and MAC387 antibodies were observed in a few cells in the basal and upper layers of the epidermis, respectively. HMB45 was positive in the junctional component of the nevi and in scattered nevus cells in the dermis. Blue nevi were strongly positive throughout.

FIGURE 5
figure 5

Melanocytic nevus with architectural disorder and atypia of melanocytes. The nevus cells are negative for anti-S100A6 antibody. Some dermal dendrocytes and endothelial cells are positive, which represent internal positive controls for the adequacy of the reaction.

Full size image

The differences observed in the number of positive cells for anti-S100A6 antibody in Spitz nevi, melanomas, and melanocytic nevi cases were statistically significant, either comparing the three entities as whole groups (P < .001) or comparing more similar histologic subgroups as junctional Spitz nevi–melanoma in situ (P < .001) and Spitz nevi–nodular melanoma (P < .001). S100A6 protein results are summarized in Tables 2, 3, 4, 5, 6 and 7.

TABLE 2 Expression of S100A6 Protein in Spitz Nevi Cases
Full size table
TABLE 3 Expression of S100A6 Protein in Melanoma Cases
Full size table
TABLE 4 Expression of S100A6 Protein in Melanocytic Nevi Cases
Full size table
TABLE 5 Comparative S100A6 Expression
Full size table
TABLE 6 Comparative S100A6 Expression
Full size table
TABLE 7 Comparative S100A6 Expression
Full size table

DISCUSSION

Spitz nevus was first described by Spitz in a paper entitled “Melanomas of childhood” (1). Dr. Spitz (1) thought that these lesions were melanomas with an unusually good prognosis in children. Later, these “melanomas of childhood” or “benign juvenile melanomas” were renamed “spindle and epithelioid nevi” or “Spitz nevi” (reviewed by Elder and Murphy, 44). Pathologists' task of using routine light microscopy to differentiate between these Spitz nevi and melanomas has been difficult, sometimes leading to overdiagnosis of melanoma when the lesion is a Spitz nevus (leading to excessive surgery) or to misdiagnosis as Spitz nevus when the lesion is in fact an invasive melanoma, with catastrophic consequences for the usually young patient. One of the cases described by Spitz (1) in her original series did in fact metastasize, illustrating the difficulty that even experts have in separating these lesions (1). The differentiation between Spitz nevus and melanoma remains problematic also as indicated by the number of slides sent for diagnostic consultation. The occasional failure of routine histopathology to separate adequately the cases of Spitz nevi from melanomas has been widely acknowledged and has led a group of authors to recommend complete excisions of Spitz nevi in adults just because of this uncertainty factor in diagnosis (45). Melanomas are increasing in the general population and often are being treated at less advanced stages than previously. In contrast to large melanomas, small ones present even greater difficulty in distinction from Spitz nevi (46, 47). These problems have led to a number of attempts to define useful adjunctive techniques to help in the distinction between Spitz nevus and melanoma. These techniques include the use of immunohistochemistry, flow cytometry, and nuclear morphometry, among others, and more recently, genetic analysis (6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38). Some of these techniques have been very cumbersome and are not readily adaptable to use for clinical diagnosis, which needs to be more rapid.

With the aim of contributing to the distinction between these two entities, the expression of the S100 protein subtypes (S100A2, A6, A8/A9/A12) was studied in Spitz nevi and melanomas as well as in melanocytic nevi. S100A6 (Calcyclin, CACY, 2A9, PRA, CaBP 5B10) can be induced by growth factors and has possible roles in the regulation of cell growth and division. It has been found in abundance in certain cancer cell lines and tumors, but is restricted to a specific subclass of cells in normal tissues (48, 49, 50, 51, 52, 53, 54, 55, 56). In the current study, S100A6 protein was strongly expressed in all Spitz nevi cases, suggesting that calcium-dependent signal transduction pathways may be abnormally regulated in these nevi. It was also present in the deep component of the compound, predominantly intradermal nevi and of nevi with architectural disorder, but with a weak and patchy staining. Although neuroid changes in ordinary melanocytic nevi expressed S100A6 protein (55), this reaction does not create problems in the differential diagnosis with Spitz nevus because these neuroid features are almost never described in them. Six of 12 combined nevi expressed S100A6 protein, mainly in a Spitz nevus component of the combined nevi. Interestingly, three of six blue nevi were weakly positive. Thirty-three percent of melanomas were anti-S100A6 positive. The positivity was mainly observed in the dermal component. The junctional component was negative or with scant positive cells. Low percentage positivity in the in situ, lentigo maligna, and superficial spreading subtypes contrasts with high percentage positivity in the nodular and metastatic subtypes of melanomas. Five of the nodular melanomas had spitzoid features. Nonetheless, the immunoreactivity was not as strong and uniform as in Spitz nevi. It has been previously reported that Type C nevus cells express S100A6 protein, supporting their terminal maturation along a pathway of schwannian differentiation (55). Sixty-six melanomas were negative for anti-S100A6. However, very undifferentiated or aggressive melanomas could express S100A6 protein again in progressive deregulation of gene expression similar to CEA expression in colon cancer.

Most of the melanoma cases of our study were primary melanomas and were anti-S100A6 negative. This negativity could result from a variety of changes, such as a mutation in a gene controlling S100A6 expression, a mutation deleting the immunogenic epitope detected by the monoclonal antibody, the expression of different S100A6 isoforms produced, or protein–protein binding blocking of the immunoreactive site, among other possibilities. Further molecular analyses of the S100A6 protein and gene should be performed in Spitz nevi, melanomas, and melanocytic nevi cases to study the underlying genetic bases for the differences observed. MAC387 antibody, long recognized as a nonspecific macrophage marker (57), has also been reported to react with keratinocytes related to cutaneous malignancies and melanocytic nevi (58, 59, 60, 61). We also observed MAC387 positivity in the keratinocytes overlying melanomas, melanocytic nevi, and Spitz nevi. The reaction was slightly stronger in melanomas and junctional Spitz nevi, comparing to other subtypes of Spitz nevi and melanocytic nevi, where it was weaker and uneven. Therefore, the use of MAC387 in this setting is non-contributory. Nevus and melanoma cells themselves were always negative with MAC387. S100A2 gene is down-regulated in some tumors and activated upon the interaction with wild-type p53; thus, it has been suggested that the S100A2 gene could act as a tumor suppressor-related gene (51). On the other hand, high expression of S100A2 protein has been reported in skin appendage tumors and in basal and squamous cell carcinomas (62). In our study, S100A2 protein expression was observed in the basal keratinocytes immediately adjacent to the nevus cell nests and to melanoma cells, without any significant difference between the cases. Nevus and melanoma cells did not express S100A2 protein.

In summary, we propose anti-S100A6 protein as a potential immunohistochemical marker that could be used diagnostically to distinguish Spitz nevus from melanoma, not only in classic cases but also in cases that are difficult to distinguish by routine microscopy. Moreover, the S100A6 expression in Spitz nevi suggests that calcium-dependent signal transduction pathways play an important role in regulating cellular processes in these nevi.