Autophagy Markers p62, LC3II and Beclin1 Correlate with Clark Levels in Melanoma Tumors

Background The prognosis of melanoma depends on early diagnosis and timely treatment. Autophagy as a mechanism of degradation/recycling of cellular debris, has potential to be evaluated as prognostic biomarker in current research. Methods In this study, ATG5 and Beclin 1 gene expression in different Clark levels of melanoma were evaluated in a retrospective study of 10 years in the cancer institute of Tehran, Iran. The autophagy activity and the correlation with clinicopathological data were also investigated in a tissue microarray series of 52 melanomas after immunohistochemical staining for the autophagy-associated proteins p62, LC3II and Beclin1. The possibility of autophagy biomarkers were assessed by ROC curve analysis. Results The patterns of ATG5 and Beclin1 gene expression are different. While ATG5 was increased in the early stage and then decreased as the stage was progressed in comparison to tumor margin, the Beclin1 expression was decreased and not altered during tumor progression. However, Beclin1 expression at the protein level was increased with tumor progression. The expression of LC3II was also raised while the p62 levels were declined as the tumor progressed, suggesting an increased autophagy activity in melanoma patients. Melanoma ulceration was positively correlated with Beclin 1 and LC3II expression and inversely correlated with p62 (p<0.05). Autophagy markers expression did not signicantly correlate with melanoma mitotic rate and thickness.


Introduction
Melanoma is one of the most dangerous types of skin cancer caused by the abnormal proliferation of melanocytes (pigment cells) [1]. The incidence of melanoma is increasing globally each year, which raised the mortality rate [2]. To reduce melanoma-related mortality and well prognosis, effectual diagnostic biomarkers are always under investigation [3,4]. It has been demonstrated that the expression levels of many autophagy-related genes and proteins altered during process of melanoma malignancies, which might be considered as biomarkers for prediction its stages, invasiveness, patient's survival and treatment targets [5]. Macroautophagy (hereafter referred to autophagy), is a major lysosome-dependent process which is activated to digest or reuse cellular contents in a critical conditionssuch as nutrient de ciency and pathogenic infection to keepcell homeostasis [6]. Indeed, the cellcytoplasm are engulfed inside into a specialized vacuole, named autophagosome, which then fuses with lysosome to degrade the materials [7]. In this regulated process, the key proteins are the AuTophaGy-related (ATG). To date, at least 41 ATG genes have been identi ed [8]. Among autophagy proteins, ATG5 is essential for autophagosome formation. Knocking down of ATG5 leads in downregulation or autophagy inhibition, suggesting that ATG5 plays a central role in autophagy [9].
Beclin-1 is a mammalian ortholog of the yeast autophagy-related gene 6 (Atg6) is involved in both the autophagy (initial step of autophagosome formation) and cell death signaling pathway by interacting to either phosphoinositol-3 kinase (PI3k) class III or BCL-2 [10].
LC3 (microtubuleassociated protein 1A/1B-light chain 3) comprises of a solubleLC3 I and a lipidated form, called LC3 II. LC3 II is involved in autophagy and recruited into autophagosomes [11]. Different types of stressors promote the conjugation of LC3 I tophosphatidylethanolamine, to organize the autophagosome speci c LC3 II, which is considered the most reliablemarker of autophagy [12]. p62 or sequestosome-1/SQSTM1 is an adaptor protein transporting of ubiquitinated proteins during autophagy [13] and plays a substantial role in selective autophagic degradation of a number of cargoes.
Generally, p62 and its substrate are broken down within the autophagolysosome. This process is mediated by interaction with LC3, which is involved in phagophore/isolation membrane and autophagosome formation [14]. During tumorigenesis, impaired autophagy leads to p62/ubiquitinated protein aggregates [14] and visualized as autophagy marker [13].
The function of autophagy in cancer is quite complex and somewhat controversial. In the early stages of cancer onset, it is considered as tumor-suppressive to prevent the proliferation of precancerous cells.
However, in the later stages, autophagy supports tumor stability by increasing the resistance of cancer cells to chemotherapy and/or radiotherapy, consequently, inhibiting or activating autophagy in different stages of various cancers is a signi cant concern in treatment [6, 15].Since, previous studies have shown the diagnostic and therapeutic importance of the autophagy pathway in melanoma, the aim of this study was to investigate the important mediators involved in the autophagy pathway (ATG5, Beclin1 p62 and

RNA extraction
The RNA extraction from FFPE tissues was carried out manually according to our lab protocol. Brie y, the tumor portion of 5 shaved cuts (11µm thickness) from each selected block was depara nized with xylene. Thenthey were centrifuged at high speed and washed several times with 70% ethanol. The pellet was kept at room temperature until the ethanol was evaporated. Following by adding lysis buffer, chloroform, cold isopropanol and 70% ethanol, centrifuge was performed in each step. The supernatant was then drained and allowed to dry completely. Finally, extracted RNA was kept in DEPC waterand the concentration was assessed using a NanoDrop™ OneC Spectrophotometer (Thermo Fisher Scienti c Inc). Sample absorbance was measured at 260 nm and 280 nm and the ratio of optical density (OD)260/OD280 was used to test for protein or phenol contamination cDNA synthesis and qPCR 1000 ng / µl RNA was transferred to cDNA using the BIOFACT kit (South Korea, cat.no:BR631-096) according to the company instruction. ThemRNA expression levels were measured by qPCR using speci c pairs of primers in the following program: initial denaturation (10 min for 95°C), denaturation (10 s for 95°C),annealing (30 s for 60°C, 40 cycles), and elongation (10 s for 72°C).The oligonucleotide sequences of primers are as follow: ATG5 F5′-CACAAGCAACTCTGGATGGGATTG-3′ and ATG5 R5′- Tissue microarray: For characterization of immunohistochemical protein expression, three block arrays were constructed for three different autophagy biomarkers. Brie y, two 1.5-mm-diameter cylinders of tissue were taken from representative areas of each archival para n block and arrayed into a new recipient para n block. In addition, normal margins of tumors were placed as internal controls and to ensure the quality of staining slides.

Immunohistochemistry
Three-micrometer tissue sections from the TMA blocks were sectioned and applied to special immunohistochemistry coated slides. These slides were baked overnight in a 56°C oven. Then were depara nized by xylene, rehydrated through a graded ethanol series and washed with phosphatebuffered saline. Antigen retrieval was achieved by a 2-minute heat treatment in a pressure cooker, containing 1 liter of 10 mmol/L sodium citrate buffer that had been previously brought to the boil. Before staining, endogenous peroxidase activity was quenched with 1.5% hydrogen peroxide in methanol for 10 minutes. Immunohistochemical staining was performed on these sections using three different antibodies, (LC3β (H-50): sc-28266, p62/ SQSTM1 (D-3): sc-28359 and Beclin 1 (E-8) sc-48341, all were from Santa Cruz Biotechnology, Inc). Selected tissues were chosen to represent a range of anticipated staining intensities for both hematoxylin and eosin (H&E as imparted using a standard staining protocol [17].

Interpretation of tissue immunohistochemical reactions
The IHC sample scoring system is a method for converting qualitative data into measurable and comparable data. Scoring is based on the intensity of coloring and the number of positive tumor cells, which is shown in table 1. Among them 5 lymph node metastasis, 4 radial growth phase melanomas (RGPMs), 3 vertical growth phase melanomas (VGPMs), and 5 Melanoma Metastasis (MelMets)were recorded. The main clinicopathologic data are given in Table 2. Table 2 Frequency of anatomical stages and the demographic and pathological data in patients with melanoma.

Autophagy markersin different Clark level of malignant melanoma
To analyst the autophagy activity, two important genes in autopagosome preparation (ATG5 and Beclin 1) were checked. ATG5 expression was increased in early stage and thendecreased as the stage was progressed (Fig. 1a). Indeed, ATG5 expression inClark level IV and V samples (advanced stage) was signi cantly (p = 0.0004) lower thanin Clark I samples (early stage) and Clark II, III (intermediate stage) (P = 0.006 andP = 0.4, respectively) (Fig. 1b). In contrast to ATG5, expression of Beclin 1 was signi cantly decreasedin melanomas(fold change = 0.08) in compare with tumor margins (fold change = 1) with p = 0.00001. Indeed,the expression of Beclin1 was in lowest level and not altered during tumor progression (Fig. 1).
To test whether enhancement of ATG5 andBeclin1 expression actually results in autophagy induction in melanoma patients, we analyzed levels of LC3 and p62 as important protein markers to detect the extent of autophagy.The expression of LC3 levels were increased as the tumor progressed ( Fig. 2a and 2b). In talking speech, the mean percentage of expression were 86/66%, 72/72%, 80% and 100% in Clarck level I, II, III and IV, respectively (Fig. 3b). We observed markedly higher LC3 levels (mean expression 84.85%) in melanomas compared with tumor margin (mean expression 15.15%) (P = 0.001), suggesting an increased autophagy in melanoma cells (Figs. 2 and 3b).

Correlation of autophagy markers with clinopatholigic factors
Melanoma ulceration was positively correlated with Beclin 1 and LC3II expression and inversely correlated with p62 (p < 0.05). Autophagy markers expression in different Clark level were not signi cantly correlate with melanoma mitotic rate, melanoma thickness and speci c sites.
LC3 and Beclin1 protein expression wassigni cantly higher in late stage of melanoma whereas the p62 expression was decreased when tumor progressed (Figs. 2 and 3). Patients with high Beclin1 and LC3II contained a signi cantly low level of p62 compared with tumor marginindicating autophagy induction in melanoma patients (Fig. 2b) that meansBeclin1 and LC3II with p62 levels correlated inversely in these patients.

Biomarker possibility
The analysis of ATG5 and Beclin1 gene as biomarker with Roc curve (graph pad prism 8.0 software) showed that they could be valuable markerswith area under curve (AUC) 0.7 and 0.5 respectively, p < 0.05 (Fig. 4).

Discussion
Emerging evidences have shown the correlation of autophagy with tumor progression, so that impaired or inhibition of autophagy leads to tumorgenisity. Howeverdifferent results of autophagy activity during tumorigenesis have been recorded to date [18]. Therefore, it is expected that autophagy activity is low in the early stages while elevated with tumor progression. However, several experiment in melanoma or even other tumors reported controversial results. For instance, Beclin1 act as a haploinsu cient tumoursuppressor gene and can be either monoallelically deleted or display reduced expression in breast, ovarian, and prostatic cancer [19]. In contrast, other reports displayed that the overexpression of Beclin 1 correlated with tumorigenesis incolorectal and gastric cancer [20]. In cutaneous malignant melanomas, the results showed that Beclin 1 signi cantly decreased with tumor progression [21,22]. In current experiment,Beclin1 gene expression was reduced in melanoma patients in compare to tumor margins (P < 0.0001). However, the evaluation of Beclin1 level in all stages showed the decreased level in stage 1 and then went up as the tumor progressed, suggesting the autophagy activity during melanoma progression or due to autophagy induction following by chemotherapy. This is agreement with the study of advanced malignant melanoma (MM) which showed overexpression of Beclin-1 in advanced malignant melanoma [23].
LC3 expression as another autophagy marker has been reported to be either decreased in brain and ovary cancer [24,25] or increased in esophageal and gastrointestinal neoplasms [26]. LC3 expression was found associated with poor outcome in pancreatic cancer and with a better survival in glioblastoma patients with poor performance score [24,27]. In cutaneousmelanocytic lesions LC3 signi cantly decreased with tumor progression and the lowest expression of LC3 II protein was observed in melanoma metastases [18]. While in another experiment with TMA in melanoma and breast has con rmed overexpression of LC3 II and the positive correlation with Ki-67 [5]. In our study, high level of LC3 II protein was observed in melanoma tumors in compare to tumor margins, suggesting autophagy activity during melanoma tumorigenesis.Overexpression of p62 is also observed in a number of cancers, including melanoma, and is a marker of poor prognosis. Indeed, increasing levels of p62 expression in early stage of melanoma followed by subsequent decreasing it, suggesting the novel independent prognostic biomarkers for early stage melanomas [28]. In current study and in line with previous studies, during the progression of tumor, the autophagy marker of p62 was decreased. Analyzing autophagy in tissues with few markers is very di cult due to the paradoxical function of autophagy in tissues. Therefore, it is better that the expression levels of LC3II and p62 markers be done in parallel with the ATG study [27].There are several reports about the ATG5 gene and protein expression in different cancers and especially in melanoma patients. Liu H et al in 2013 reported ATG5, is often down-regulated inprimary melanomas compared to benign nevi. They also checked LC3II, Beclin1 and p62 in a few cases. The results have shown that Beclin1 was indistinguishable between melanomasand benign nevi, a reducedexpression of LC3 and increased expression of p62 was observed [29].In current experiment, the autophagy markers in different stages of melanoma was determined and revealed that although ATG5 was raised at early stage, the comparison of ATG5 expression in all melanoma tumors was decreased in compare with tumor margin (p < 0.05). Autophagy can protect tumor cells that are exposed to anticancer drugs, indicating the potential for inhibition of autophagy in cancer therapy [30].It has been shown that elevated autophagy in early stage is associated with invasiveness and drug resistance [31,32]. Moreover, autophagy was induced in melanoma cells under situations of metastasis [33] and/or hypoxia [34].All these data propose thatautophagy may be increased in melanoma cells as an adaptive stressresponse to chemotherapy and/or metabolic stress. In such studies, thequanti cation of additional ATG proteins should also be considered. So, in addition to ATG5, Beclin 1 expression at mRNA level was evaluated and revealed that the alteration of expression was notobserved during tumor progression. However,its expression was decreased in all stages in compare with tumor margin (p < 0.0001), suggesting that the expression of ATG proteins is, at least partially, differentially regulated. Previously published work, however, had reported a down-regulation of Beclin 1 in association with melanoma disease progression [18].
In contrast to the previous reports, our experiment in all stages was reported the autophagy activity was observed during tumor progression.Although in early stage, a reduction of autophagy may promote tumor growth, by preventing senescence, In later stages, basal autophagy activity resumes or is induced, leading to chemoresistance. Indeed, the regulation of autophagy in tumorigenesisdepends on the stage of the tumors [29].
Furthermore, Beclin 1 and LC3II expression level during tumor progression positively correlated with melanoma ulceration whereas p62 protein expression was inversely correlated (n = 22). And were not signi cantly correlate with melanoma mitotic rate, thickness and speci c siteswhereas previous study reported that in primary melanomas, melanoma thickness and ulceration were positively correlated with Beclin 1 expression and inversely correlatedwith cytoplasmic LC3 II protein expression.Melanoma mitotic rate wasinversely correlated with cytoplasmic LC3 protein expression [18], recommending more samples need to be analyzed.
Prognosis and survival for malignant melanoma is highly dependent on early diagnosis and treatment. Earlier experiments proposed p62, LC3 and Beclin 1 have been proposed as potential prognostic biomarkers for early stage of melanomas [35]. We also showed ATG5 and Beclin 1could be a reliable markers in early detection of melanoma.

Conclusion
Key autophagy regulatory proteins, including p62, LC3 and Beclin 1, have been proposed as potential prognostic biomarkers. p62 expression correlate inversely with Beclin1 and LC3II level, that means increasing of p62 followed by subsequent decreasing levels of Beclin 1 and LC3 expression in early stage, while as tumor progression, the pattern was changed and autophagy activity was increased, re ects the paradoxical role of autophagy in melanoma. Collectively, the de nition of ATG5 and Beclin 1 as prognostic biomarkers for early stage of melanomas provides novel and accurate means through which to identify tumors at risk of disease progression, facilitating earlier patient therapeutic intervention and strati cation tools for novel personalized therapeutic approaches to improve clinical outcome.  The level of Beclin1, LC3II and p62 expression based on Intensity score (0: negative, 1: weak, 2: Intermediate, 3: strong) were evaluated. All recordings calculated in percent compare to tumor margins.