Comparison of the effects of C-KIT, MMP-2, Ki-67, Bcl-2 and metallothionein expression on prognostic factors in skin and non-skin malignant melanomas


 Objectives: We aimed the relationship between with selected prognostic factors of MM, 1) C-KIT (CD117), which related to MM pathogenesis, 2) Ki-67, which is the determinant of cell kinetics, 3) MMP-2, which degradation of the basement membrane and extracellular matrix required for cell migration, invasion and metastasis in melanogenesis, 4) bcl-2, which reduces the apoptotic response to cytotoxic chemotherapy and 5) MT’s whose overexpression shows resistance to anticancer drugs and radiotherapy and thus poor prognosis.Material and Methods: In this study, 30 skin and 15 non-skin MMM cases diagnosed between January 01, 2000 and June 01, 2011 in SCU-TF Medical Pathology Department were examined. C-Kit, MMP-2, Ki-67, bcl-2 and MT markers were applied to these cases.Results: When comparing the skin and non-skin groups, the difference between Ki-67 and MMP-2 expressions was found to be statistically significant, but no significant difference was found in staining with c-Kit, MT and bcl-2. It was thought that high MMP-2, C-KIT, bcl-2 expressions and high Ki-67 index as well as loss of MT expression might be associated with poor prognosis. Despite the high level of KIT immunostaining in melanomas, this parameter does not appear to be a good predictor of the presence of molecular mutations. Mutations that activate KIT should be considered a rare event in this tumor.


Introduction
MM is one of the most important malignant tumors of the skin, with a high mortality rate and high invasion capacity. It can also arise from mucosal surfaces. These surfaces include oral and anogenital mucosa, sinonasal mucosa, esophagus, intestines, meninges, all layers of the eye and bladder [1][2][3] .
Major prognostic factors of malignant melanoma are Breslow thickness, tumor size (largest tumor size and others), ulcer, histological type, Clark level in the skin, tumor in ltrating lymphocytes, lymphovascular invasion, perineural invasion, and regression. The rst three of these (Breslow thickness, largest tumor size and ulcer) together with lymph node and distant organ metastases determine the pathological stage according to the TNM classi cation.
In the late 1980s, Holzmann B et al. proposed a model of melanoma in which benign melanocytes gradually evolved into melanoma cells with metastatic capability, in which every step was de ned by the acquisition or loss of certain cellular markers that were easily detected by immunohistochemical analysis 4,5 . Of these, c-Kit (CD117) is a growth factor for melanocyte migration and proliferation. It shows different staining properties in various benign and malignant melanocytic lesions. While different rates of c-Kit expression were observed in some studies, the development of melanoma was found to be associated with loss of c-Kit expression in some studies 6,7 .
Degradation of basement membranes and extracellular matrix is an essential step for melanoma cell migration, invasion, and metastasis formation in melanogenesis. Matrix metalloproteinases and their tissue inhibitors play a crucial role in these complex multistep processes. Melanoma cells may express a number of matrix metalloproteinase family members (MMP-1, MMP-2, MMP-9, MMP-13, and MT1-MMP) as well as their tissue inhibitors (TIMP-1, TIMP-2, and TIMP-3) 8 . The combination of determining MMP-2, Ki67, and p53 immunoreactive proteins could be bene cial in the selection of high-risk melanoma patients for future adjuvant trials.Increased expression and functionally active of MMP-2 are associated with the progression of melanoma 9 .
Ki-67, a determinant of proliferation kinetics, is not expressed in resting G0 cells of the cell cycle, but is expressed in G1, S, G2 and M phases. Therefore, the Ki-67 level is expected to be high in many aggressive MM cases [10][11][12] .
Programmed cell death (apoptosis) has been implicated in tumor development and may affect the metastatic potential of tumor cells. The role of bcl-2, a proto-oncogene that inhibits apoptosis, has been studied in several malignancies, including cutaneous melanoma (CM). It is a family of pro-and antiapoptotic proteins, regulates mitochondrial membrane permeability and proapoptotic mitochondrial pathway 13 . Dysregulation of Bcl-2 proteins appears of critical importance for melanoma cell survival and drug resistance, and among the fast increasing number of new targeted therapies, those, which affect Bcl-2 proteins, may especially apply for melanoma 14,15 .
Metallothioneins (MT) are intracellular, low molecular weight, cysteine rich proteins. Its overexpression, which can be seen in many tumors, indicates resistance to anticancer drugs and radiotherapy and poor prognosis. A similar situation is also present in melanoma and non-melanoma skin tumors 16-18 . In our study, the relationship of these markers with selected prognostic factors (histological type, Breslow thickness, mitotic index, ulceration and localization) in cutaneous and mucosal malignant melanomas was investigated.

Materials And Methods
Ethics statement. Ethical approval for the study was obtained from Sivas Cumhuriyet University Scienti c Research Assessment Board dated 04.12.2012 and numbered 2012-12/13. Institutional Review Boards of all participating institutions approved the study protocols and written informed consent was obtained from all study participants. Research methods of procedures were carried out in accordance with relevant guidelines and regulations. 30 CMM and 15 MMM cases diagnosed in SCU-TF-PD between 01 January 2000 and 01 June 2011 were included in the study. C-Kit, MMP-2, Ki-67, bcl-2 and MT markers were applied to these cases. The relationship of these markers with selected prognostic factors (histological type, Breslow thickness, mitotic index, ulceration and localization) was investigated.

Evaluation
Histological evaluation was performed by re-examining all H&E stained preparations of the cases, NICON, OPTIPHOT, JAPAN under the light microscope.
Information on the localization of the tumor was obtained from the SCU-TF-PD report archive.
Histological type was speci ed as NM, SSM, ALM, LMM and other types.
The mitotic index was counted in the "hot area" where the mitotic activity was the highest, that is, in 4x400 area (in mm 2) in our microscope and grouped as 0, (0.1-6) and 6.
It was stated that the ulcer was present or absent.
External control preparations were prepared for each IHC staining.
MMP-2 expression was in the form of cytoplasmic staining and it was evaluated as negative if there is less than 5% staining, weak positive if there is 5-20%, moderate positive if there is staining in 21-50% malignant cells and strong positive if there is> 50% staining 8 .
Ki-67 expression was counted in the area of highest staining, given as a percentage. If staining <20%, it was considered as low expression, if ≥ 20% as overexpression 9 .
Bcl-2 expression was also in the form of cytoplasmic staining, and staining less than 5% in tumor cells was accepted as 0 (negative), between 5-50% staining 1 (weak positive), and more than 50% staining 2 (strongly positive) 12 .
MT expression is in the form of both cytoplasmic and nuclear staining. <10% staining in tumoral cells was considered negative, if there was more than 10% staining, it was accepted as overexpression. To prevent false positive results, 10% was taken as the cut-off value [14][15][16] .

Results
The ages of the individuals in the CMM group varied between (3-89), with a mean (67.68 ± 16.17), and the mean age (49-88) of the individuals in the MMM group (69.84 ± 12.44). The difference between the two groups in terms of age was found to be insigni cant (x²: 0.41; p: 0.681; p> 0.05).
Again, 17 (57%) of the individuals in the CMM group are male, 13 (43%) are female; 7 (47%) of the individuals in the MMM group were male and 8 (53%) were female, and the difference between groups in terms of gender was found to be insigni cant (x²: 0.43; p: 0.508; p> 0.05).
Histological type: In CMMs, 22 cases (73.3%) were NM, 4 cases (13.3%) were SSM, 3 cases (10.0%) were LMM and 1 case (3.3%) was ALM. In MMM, 14 cases (93.3%) were in the NM type and 1 case (6.7%) was in the SSM type. When these groups were compared in terms of histological type, the difference between them was found to be insigni cant (p> 0.05).
Ulceration: While no ulcer was found in 13 cases (43.3%) in CMMs, ulcers were observed in 17 cases (56.7%). While there were no ulcers in 5 cases (33.3%) in MMMs, ulcers were observed in 10 cases (66.7%). When CMM and MMM were compared in terms of ulcer, the difference between the groups was found to be statistically insigni cant (p> 0.05).
Furthermore, when CMM and MMMs were compared in terms of Ki-67 expression according to their histological types, the difference between CMM cases was statistically signi cant (p <0.05), and the difference between MMMs was found insigni cant (p> 0.05). In the CMM group, in NM cases, 15 cases (.68.2%) with a rate of <20% were detected. In non-skin MM, in NM cases, 10 cases (71.4%) with ≥20% staining were observed (see in Table. 3).
When CMM and MMMs were compared in terms of MMP-2 expression according to their histological types, the difference was found to be statistically insigni cant (p 0.05). However, in 10 (45.5%) of the NM's in the CMM group, MMP-2 expression was detected at a rate of 50%. MMP-2 expression was 50% in 8 (57.1%) of the NM in the MMM group (see in Table 4).
When CMM and MMMs according to histological types were compared in terms of bcl-2 expression, the difference was found to be statistically insigni cant (p 0.05). However, bcl-2 expression was 50% in 10 (45.5%) of the NMs in the CMM group, while bcl-2 expression was 50% in 8 (57.1%) of the NMs in the MMM group (see in Table 5).
When CMM and MMMs according to histological types were compared in terms of MT expression, the difference was found to be statistically insigni cant (p 0.05). However, while there were 17 cases (77.3%) with MT expression <10% in NM in the CMM group, 8 cases (57.1%) with MT expression <10% in NM in the MMM group were detected (see in Table 6).
When CMM and MMMs were compared in terms of c-Kit expression according to histological types, the difference was found to be statistically insigni cant (p 0.05). On the other hand, 8 cases (36.4%) were detected with <10% staining in NM in the CMM group, and 10-50% was observed in 5 cases (35.7%) of the NM type in MMM (see in Table 7).

Discussion And Conclusion
The frequency of MM among all skin cancers is 3-5%. It is responsible for 75% of deaths due to skin cancers and 1-2% of deaths due to all cancers [1][2][3]20 . It is stated that the incidence of whites in industrial countries has increased signi cantly in the last two decades, and nowadays its incidence has stabilized and even decreased. Survival rates increase in patients with melanoma, as patients present earlier and with thinner melanoma.
Five-year survival rates have been reported at 85% or more in women and 75% in men. Again, in another cohort analysis of 4791 patients diagnosed with primary CMM between 1976 and 2001 in southern Germany, primary CMM diagnosis during 1990-2001 was associated with a more favorable 10-year survival (88.6% versus 80.0%, P <.0001) compared with 1976-1989. Median tumor thickness at primary diagnosis was signi cantly lower in the second period (0.75 mm vs 1.07 mm, P <.0001). That is, tumor thickness has been found to be a dominant prognosis determinant 23 .
In the study, the number of patients with ulcers (66%) and the number of cases with Breslow thickness 4 mm (53%) were higher in MMMs compared to CMMs. This situation may be associated with the worse prognosis in MMMs compared to skin melanomas.
In MMMs, there are many prognostic factors other than tumor thickness that have different importance. These are histological type, ulceration, mitotic index, presence of satellites, angiolymphatic invasion, advanced stage, occult metastases, local recurrence, Clark invasion depth, location, accompanying nevus, lymphocytic in ltration, regression, nuclear volume, sex, vitiligo, age and can be counted as pregnancy [1][2][3]24 .
Also, risk factors for MM development include hair and skin color, the presence of multiple freckles, PUVA therapy, the presence of multiple large atypical nevi, genetic factors (CDK2A and CDK4 mutations), xeroderma pigmentosum, immunosuppression, chemicals, oil, and exposure to printer products. involves trauma (for subungual melanoma) and burns counted 1-3 .
According to AJCC staging, three important histopathological criteria are Breslow thickness, ulcer and mitotic index (25). Age, gender, and anatomic location of the primary tumor that are not included in the AJCC system also affect the survival of primary MM cases (26,27). When these factors are combined with histopathological features, they show a better prognosis than AJCC system 24 .
Gonzalez et al. showed that melanoma expresses c-kit, a gastrointestinal stromal tumor marker, but has not been extensively evaluated for protein kinase Cθ (PKCθ) or DOG1, and these stains have not been associated with prognostic factors. They immunostained 62 primary cutaneous and 15 metastatic melanomas for polyclonal c-kit (pc-kit), monoclonal c-kit (mc-kit), PKCθ, and DOG1, and correlated the results with prognostic parameters and survival. 34 (55%) of cutaneous melanomas were stained for pckit, 30 (48%) for mc-kit, 11 (18%) for PKCθ, and 2 (3%) for DOG1 29 . anorectal melanoma cases had a c-kit mutation, whereas 16 of 40 cases had CD117 expression, and 3 of these 16 cases also found c-kit mutations 31 . In the study of Alessandirini et al., C-KIT stains were documented in 48% of melanomas, 50% of PAMs and 24% of nevi. The mean score of KIT staining in the melanomas/PAMs group was signi cantly different from nevi (p=0.0076). No statistically signi cant differences were detected between either c-kit immunostaining score or pattern and each of the other clinico-pathologic parameters considered. No KIT gene mutations were detected in melanomas and nevi.
A silent mutation/polymorphism in KIT exon 13 was found in one PAM 32 .

Liu et al. investigated c-Kit and Sox10 expressions in 28 patients with sinonasal mucosal melanoma, and
determined c-Kit expression in 24 patients, therefore they emphasized that c-Kit expression may be useful in the regulation of treatment. However, they also stated that Sox10 is a sensitive determinant in SSMs 33 . In the study, c-Kit expression was detected in all conjunctival MM's. In addition, C-KIT expression was found in all but 3 of 11 nonconjunctival cases in the MMM group.
Weinlich et al. investigated the relationship between MT overexpression and prognostic factors in their second study, which included 1270 cases in 2006. While MT overexpression was determined in 310 cases (24.4%), MT was found negative in 960 cases (75.6%). The female / male ratio (51% / 49%) of the patients were close to each other, and the mean age at which lesions were excised was 54 years (7-95).
During their follow-up, it was found that the disease progressed in 167 cases and death due to metastatic disease occurred in 110 cases. MT overexpression was detected in 117 of 167 cases and 80 of 110 cases in these groups. The proportion of MT positive melanomas increases with greater Clark invasion depth and Breslow thickness 17 . Accordingly, they also investigated whether MT overexpression was an independent prognostic factor or only a parameter correlated with increasing degree of invasion or tumor thickness. For this purpose, they found that MT overexpression was signi cant in a statistical model for disease progression (p <0.001). In conclusion, MT overexpression is a strong, highly signi cant factor independent of tumor thickness, and determines and measures the risk of disease progression.  18 . In the study, the number of cases with MT expression <10% was found to be higher in the skin and extradermal group than the cases with MT overexpression. However, most of the MT negative cases were histologically of NM type and Breslow thickness was higher. In addition, in the MMM group, in patients with low MT expression, ulcers were observed at a higher rate. These ndings oft he study can be interpreted as indirectly, that the loss of MT expression indicates a serious prognosis, albeit a little.
Väisänen et al. in their CMM series of 157 cases, investigated the effect of MMP-2 expressions on Ki67 and p53 expression as well as survival, overexpression of the MMP-2 protein and high metastatic potential melanomas characterized by Ki67 and low MMP-2. They founded overexpression of the matrix metalloproteinase 2 protein in conjunction with overexpression of Ki67 characterized melanomas with high metastatic potential and was associated with declined survival with a 10-year disease-speci c survival of 33% compared with 85% in the cases with low matrix metalloproteinase-2 and low Ki-67 levels (p = .002). Similarly, in cases with overexpression of matrix metalloproteinase-2 and a positive immunoreaction for p53, the 10-year disease-speci c survival was only 42% compared with 80% in patients with matrix metalloproteinase-2 less than 20% and a negative immunostaining for p53 (p < .001). The presence of all 3 adverse prognostic factors was prognostically more signi cant than any marker alone with a 10-year survival of only 28%. This combination of determining matrix metalloproteinase 2, Ki67, and p53 immunoreactive proteins could be bene cial in the selection of highrisk melanoma patients for future adjuvant trials (9). In our series, 7 cases (24.1%) with <5% MMP-2 expression in their CMMs, 6 cases (20.7%) with 5-20%, 4 cases (13.8%) with 21-50% and 12 cases with 50% (% 41.4) was determined. There were no cases with MMP-2 expression rate <5% and 5-20% in MMMs. 7 cases (46.7%) with 21-50% staining and 8 cases (53.3%) with 50% staining were identi ed. When compared in terms of MMP-2 expression, the difference between two groups was found to be statistically signi cant (p <0.05).
As a result, when the CMM and MMM groups were compared, the difference between them was statistically signi cant in staining with Ki-67 and MMP-2, while staining with c-Kit, MT and bcl-2 was statistically insigni cant. It was thought that high MMP-2, c-Kit, bcl-2 expressions and high Ki-67 index as well as MT expression loss were thought to be associated with poor prognosis when evaluated with associated histological prognostic factors. However, in order to precisely determine the prognosis of MMs and to determine more effective treatments, larger case series should be followed with molecular pathology studies.