SIRT1 and Ki67 immunohistochemical expression in progression of cutaneous malignant melanoma

The current study evaluated the proliferation index and immunohistochemical expression of SIRT1 on normal skin and cutaneous melanocytic nevi (CMN) and cutaneous malignant melanoma (CMM). Formalinfixed paraffin-embedded tissue samples from 43 CMN, 22 CMM and 10 normal skin were obtained and clinical data were abstracted from the electronic medicalrecord.SIRT1 and Ki67 proteins expressions were evaluated regarding to clinic pathological behavior in CMM. The level of significance was set at α = 5% (p < 0.05) .Our findings showed that SIRT1 positivity was significantly higher in benign melanocytic nevi than that in cutaneous malignant melanoma (p=0.002). As expected, the proliferation index was significantly higher in samples of cutaneous malignant melanoma as compared to the normal skin and melanocytic nevi (p=0.001). However, the expression of Ki67 protein was not also significantly related to the expression of SIRT1 (p > 0.05). In conclusion, low expression of SIRT1 and high proliferation index may play an important role in progression of cutaneous melanoma. Keywords— melanocytic nevi; skin lesions; proliferation; sirtuins.


INTRODUCTION
Skin cancer is the third most co mmon hu man malignancy and its global incidence is rising at an alarming rate, with basal cell carcino ma, squamous cell carcino ma and melanoma being the most common forms [1].There are an estimated 2-3 million cases of skin cancer across the world each year, and although cutaneous malignant melano ma (CMM) only accounts for about 200,000 of these (World Health Organization), it is the most dangerous form, accounting for most skin cancer deaths [2]. CMM diagnosed in early stage can be cured by surgical resection, and about 80% of cases are dealt with in this way [3]. However, metastatic malignant melano ma is largely refractory to existing therapies and has a very poor prognosis, with a median survival rate of 6 months [4].
Malignant melanoma is a tu mor that originates fro m the melanocytes and manifests main ly on the skin [5]. UVR exposure is a major risk factor, especially in light skin populations [6]. Lightly p ig mented skin and a large nu mber of melanocytic nevi are associated with increased risk of developing malignant melano ma [6,7] Cutaneous melanocytic nevi (CMN) are benign proliferations of melanocytes that are postulated to result fro m sun-induced mutations [8],typically BRAF [9]and genes associated melanocytic system [10]. In the majority of such neoplasms, subsequent melanocyte senescence is induced by tumor suppressor proteins such as p16 and the nevus therefore ceases to grow and becomes stable or even involutes [11]. The accumu lation of alterat ions in key genes controlling processes such as proliferation, apoptosis, senescence, and response to DNA damage can favor the format ion morphologically atypical melanocytes predisposing the risk of developing melanoma [12,13]. Cutaneous melanocytic nevi can progress to the intraepidermal lesion that can involve some local micro invasion of the dermis. The next phase, the cells can progress invading the dermis, a mo re dangerous stage in which the cells have metastatic potential, with nodules or nests of cells. Not all melanomas pass through each of these individual phases but develop directly fro m isolated melanocytes or nevi, and both can progress directly to metastatic malignant melanoma [14]. Silent mating-type informat ion regulation 2 homologue 1 (SIRT1) is a protomember of the sirtuin family (SIRT1-7) that is involved in a variety of biological processes, including genetic control of aging, regulating transcription, apoptosis, stress resistance and energy efficiency during low-calorie conditions [15,16]. To date, the role of SIRT1 remains controversial as previous data suggest that SIRT1 can act as an oncogene or a tumor suppressor, likely depending on cell type, its distribution and biological targets [17][18][19].
Recent studies demonstrated that SIRT1 levels are reduced in some types of cancers, and that SIRT1 deficiency results in genetic instability and tumorigenesis [20,21]. SIRT1-deficiency resulted in an increased tumor format ion in p53-null mice [22]. SIRT1 inhibits proliferation of pancreatic cancer cells expressing pancreatic adenocarcinoma up-regulated factor [23]. On the other hand, SIRT1 has been considered as a tumor promoter because of its increased expression in some types of cancers and its role in inactivating proteins that are involved in tu mor suppression and DNA damage repair [24]. The role and functional significance of SIRT1 in cancer development and progression is currently an intense area of research investigation. SIRT1 has been shown to be upregulated in several cancers such as prostate cancer, cutaneous T-cell ly mphoma, colo rectal cancer and pancreatic [16,[25][26][27][28].SIRT1 is also overexpressed in nonmelano ma skin cancers, including squamous and basal cell carcino mas, actin ic keratosis, and especially in Bowen's disease [25,29]. Ho wever, several studies have shown that both the overexp ression and low exp ression of SIRT1 has been linked to poor disease prognosis and survival depending on the type of cancer [30][31][32][33][34]. In spite of the controversial role of SIRT1 in tumorigenesis [35,36], it is evident that SIRT1 is significantly involved in the process of tumorigenesis, however, its exp ression status in melano ma is poorly defined and are required further investigation. In this research, we observed the expressions of SIRT1 and index proliferation in skin lesions, and investigated the association between the expressions and clinicopathological characteristics.

Patients
This retrospective and cross -sectional study analyzed samples of hu man tissues in 17 normal skins (control), 40 benign cutaneous melanocytic nevi (CMN) and 22 cutaneous malignant melanoma (CMM) with confirmed histopathological diagnosis. Clinic data were obtained from medical records of patients attended at public health centers for Oncology treatment at Montes Claros city, Minas Gerais state, Brazil. The normal hu man skin samples were obtained fro m patients who experienced aesthetic or corrective surgical procedures in women's breasts.
Ethical approval for this study was obtained from a relevant local ethic co mmittee (Co mmittee on ethic in research -FaculdadesIntegradasPitágoras: protocol no: 714.865/2014).

Immunohistochemical reactions
The 5-μm tissue sections were deparaffinized, hydrated and the antigen retrieved. The tissue sections were incubated with 3% (v:v) hydrogen peroxide for 30 min at roo m temperature to quench the endogenous peroxidase.
After blocking in normal goat serum, the tissue sections were incubated with the primary antibodies anti-Ki67 (Mouse monoclonal, Dako, 1:200 dilution, clone MIB-1, Glostrup, Denmark) and anti-SIRT1 (Clone sc-15404, Santa Cruz Biotecnology, Dallas -TX) overn ight at 4°C. The slides were then washed in PBS and incubated with LSA BTM-Kit Plus Pero xidase® for 1h. Tissues were stained with a chro mogen amino-etilcarbazol, counterstained with Mayer's hemato xy lin, cover slipped, and visualized under an optical microscope Oly mpus® BH2 microscope (model: CX31; RTSF, Miami, USA).Positive and negative controls were applied according to the manufacturer's instructions (DakoCyto mation, Glostrup, Den mark).All control, CMN, and CMM samples were examined by two independent investigators who were blind to the clinical data.

Counting of immunostained cells with Ki67and SIRT1
The immunohistochemical staining for Ki67 and SIRT1 were measured manually using custom software ImageJ®, version 1:44, for Windows® to assist in performing the cells counts. Ten images were acquired per case at a total magnification of ×400 using an optical inverted Oly mpus® FSX100 microscope (model: CX31; RTSF, M iami, USA). Selected fields were those with highest density of Ki67 or SIRT1 positive cells. Ki67 labeling index was performed as fo llo ws: % marking = (positive nuclei/[positive nuclei + negative nuclei]) [41,42].
The immune reactiv ity of SIRT1 was evaluated in the normal epithelial /nevi/neoplastic cells considering the cytoplasmatic and/or nuclear staining, or even absent. It was estimated the proportion of cells labeled with both cytoplasmic and nuclear expression in each one of the photomicrographs. Next, the average of the ratios was calculated for each case, considering individual nuclear and cytoplasmic expression for statistical analysis. It aimed to further determine the best cutoff point to define the expression of the protein as positive or negative in samples to the lesions types and location of this p rotein, using the receiver operating characteristic curve (ROC curve). In cytoplasmic exp ression, it was found that any ratio of h igher than 36.1% represented a good cut-off for positive cytoplasmatic staining (p <0.021). In the evaluation of nuclear exp ressions identified that positivity would be better represented in values above 1.8150% markup (p = 0.198). In the final evaluation o f the positive immunohistochemical exp ression of SIRT1, cases were further ranked as cytoplasmic or nuclear staining to study statistical inferences.
Statistical analysis Statistical analyses were performed using SPSS® 18.0 (SPSS Inc., Illinois, USA) and Graph Pad Prism®5.0 Softwares. Results were exp ressed as mean ± SE or as percentages. P values ≤0.05 were considered statistically significant.
Co mparisons of immunohistochemical expressions of studied proteins between the lesions were evaluated using Mann-Whitney and Fisher's exact test. Kruskal-Wallis was used to evaluate the differences between SIRT1 and Ki67 exp ressions and clin icopathological characteristics of melano ma.The analysis immunohistochemical of the expression of SIRT1 and variables clin icopathological lesion and comparing of the expression of SIRT1 between the types of study samples was performed by Ch i-square test and Fisher's exact test with applicat ion of ROC (" Receiveroperator curves") curve. The curve ROC was used to assess the sensitivity and specificity as the cutoff point for analysis of SIRT1 expression.

SIRT1immunohistochemical expression is reduced in human cutaneous malignant melanoma
SIRT1 immunohistochemical staining was localized in tu mor and normal cells (ly mphocytes and fibroblasts) with varying intensities.The immunohistochemical staining of SIRT1 in CMM, CM N and normal skin is shown in Figure 1.
Based on the ROC curve, it was simulated a cutoff point to distinguish samples with low and high staining of SIRT1, according to the d iagnosis of the sample. Applying the values on a ROC curve, the area under the curve [43,44]was 0.684 (95% CI) with best estimates occurring in the amount of 36.1, wh ich was a sensitivity of 59.1% and specificity of 23.5% (p= 0.021).
Tumors with scores above the 36.1 cut-off values were considered positive for the cytoplasmic expression of SIRT1 protein. According to ROC curve analysis, expression percentage for nuclear SIRT1 above the critical value 1.8150% was defined as positivity.
Applying the values on a ROC curve, the area under th e Normal skin samples (control) showed weak or negative cytoplasmic staining to SIRT1. In normal skin, cytoplasmic SIRT1 staining was weakly and d iffusely expressed in suprabasal epidermal keratinocytes, with only faint and focal staining in the granular layer and stratum corneum. The SIRT1 protein was detected in the normal epithelial t issues in 35.29% of normal skin (Tab le 1).
The SIRT1 protein was present in both cytoplasmic and nuclear co mpart ments of the cutaneous melanocytic nevi. Therefore, SIRT1-positive cases were classified into two categories (nuclear or cytoplasmic SIRT1). SIRT1 was positive in 76.74% of benign melanocytic nevi cases. A mong the SIRT1-positive cases, 63.63% were cytoplasmic positive and 36.36% were nuclear positive. In part icular, the junctional co mponent of benign melanocytic nevi was positive in the most cases. SIRT1 positivity is observed in the majority melanocytes, especially those arranged in bridging nests at the dermoepidermal junction and the intrader mal component.
The SIRT1 protein was present in cytoplasmic compart ment of the malignant melanocytic cells. SIRT1 was positive in twelve of 23 MM C (52, 17%), and all SIRT1-positive MM C cases showed cytoplasmic positivity. SIRT1 positivity was significantly h igher in benign CMN than that in MMC (p = 0.002). The invasive component of melano ma shows a weak and diffuse SIRT1 cytoplasmic staining. Most MMC displayed high rate of melanocytes expressing SIRT1 in intradermal component. SIRT1 exp ression was not significantly related to any of the clinicopathological parameters ( Table 2).
Ki67 immunostaining was nuclear and nucleolar (Fig.1). The average nu mber of Ki67-positive cells was significantly h igher in samp les of cutaneous malignant melano ma as compared to the normal skin and melanocytic nevi (p <0.001) (Table1). The normal skin and benign CMN displayed positive Ki67 immunostaining in basal keratinocytes whereas benign CMN was absent in the majority of nevi cells. The expression of Ki67 it found in intradermal co mponent of MMC.
The associations between the Ki67 expression and clinicopathological factors did not have statistical significance (Tab le 2). Furthermore, the expression of Ki67 was not also significantly related to the expression of SIRT1 (p>0.05).

IV. DISCUSSION
This study investigated the immunohistochemical exp ression of Ki67 and SIRT1 in normal skin, CM N and CMM samp les. In this study, we noted a significant decrease of staining of SIRT1 fro m CMN to CMM samples. Also, normal epithelial cells showed weak o r negative staining to SIRT1 wh ile CM N samples exhib ited a h igher SIRT1 exp ression as compared to the CMM samp les. Co mparatively, the SIRT1 expression was gradually decreased during carcinogenesis and tumour progression of colorectal adenocarcinoma [33]. Th is suggests that loss of SIRT1 expression in tu moral lesions may be associated with a more aggressive phenotype. However, the ro le of SIRT1 in human malignant tumors is controversial. So me previous studies have reported that SIRT1 overexpression was associated with shorter overall survival o r poor prognostic indicators in breast and gastric carcino ma [35,45,46].
The exp ression of SIRT1 is relat ively higher in hepatocellular carcino ma, breast cancer, and thyroid cancer but lower in colon and lung cancer [35,[47][48][49][50]co mpared with their corresponding normal tissues. In cancer, SIRT1 has been reported as either an oncogenic or a tumor suppressive role, depending on the type of cancer and the context of the analysis [17,21,51]. These results suggest that SIRT1 may acts differently depending on the specific organ or type of tumor involved.
We demonstrated that SIRT1 is predominantly localized in the cytoplasm of CMM. SIRT1 cytoplasmic localization is not common ly identified in cancercells and it is unclear if SIRT1 localizat ion has any changes during carcinogenesis. Similar results have also identified aberrant cytoplasmic localizat ion in human cancer cells [52][53][54].This finding may suggest a new mechanis m fo r SIRT1 function as a cancer-specific survival factor by targeting cytoplasmic proteins.
In contrast to its well-described role in the nucleus, the deacetylation function of cytoplasmic proteins caused by SIRT1 provides important insights into the function of cytoplasmic SIRT1. Zhang, 2007 showed that SIRT1 enhanced IGF-1 signaling by deacetylating the ISH-2 cytoplasmic protein [55]. SIRT1 also deacetylates cytoplasmic cortactin and promotes cancer cell [35,52] In addit ion, SIRT1 was found to promote the activation of cytoplasmic kinases, including AMPK,.Ras-MAPK, Erk and S6K1 [52].
In the case of certain cancers, includ ing prostate cancer, lung cancer, breast cancer, and melano ma, SIRT1 has been shown to localize to the cytoplasm, while being located predominantly in the nucleus in the corresponding normal t issues [52,53]. Th is change in localization could theoretically min imize the deacetylation of TP53 in the to play a more significant role, especially those that are localized to the cytoplasm [17]. Cellu lar localization o f SIRT1 also has been shown to differ among different tissue types in mice [56], wh ich could exp lain why SIRT1 sometimes exhibits tumor suppressor properties in certain types of cancer but not in others.
In current study, there was no association significant between SIRT1 and Ki67 immunostaining, however, fu rther functional study will be needed to investigate the relationship between SIRT1 and cellular proliferation. Melanoma is known to exh ibit aberrant expression of proliferation markers, and these abnormalities are considered important steps in the genesis and progression of melano ma [57].An increasing literature describes the role o f pro liferat ion markers in the evaluation of melanocytic tumors [58]. Ki67 staining has been shown positive in mu ltip le lesions, 5% of positivity on melanocytic cells in most benign nevi, although there have been reports of up to 15% positivity in Spit z and dysplastic nevi [59][60][61][62]. Conversely, Ki67 staining is reported as positive in 13-30% of the cells in a malignant melano ma, although individual cases can show almost 100% nuclear positivity [62,63]. In our study, we found a lower average than 5% staining in benign melanocytic nevi wh ile in melanoma was greater than 15%. Therefore, Ki67 index was reported to be h igher in malignant melano mas than in benign nevi. Correspondingly, no associations between Ki67 and measures of tumor size (thickness and diameter) and invasion (Clark's level) were found. Other studies on cutaneous melanoma have suggested that increased Ki67 expression might be associated with tumor thickness and tumor cell proliferation [62].
Our experiments have some limitations, such as small samp le size. The tumor heterogeneity and staining scoring ethod also may to interfere the results. In summary, we need further study on the roles of SIRT1 and Ki67 on clinical and pathological behavior of melanoma.

V.
CONCLUSION Low expression of SIRT1 and high proliferation index may p lay an important role in progression of cutaneous melanoma.