RESEARCH BMI1 and TWIST1 Downregulated mRNA Expression in Basal Cell Carcinoma

Background: BMI1, TWIST1 and SNAI2/SLUG have been implicated in aggressive behavior of squamous cell carcinoma (SCC) and melanoma and BMI1 expression could identify subtypes of Merkel cell carcinoma (MCC). However, BMI1, TWIST1 and SNAI2 expression levels in basal cell carcinomas (BCCs) have not been elucidated. We hypothesized BCC could be a good model system to decipher mechanisms which inhibit processes that drive tumor metastasis. The aim of this study was to examine the mRNA expression level of BMI1, TWIST1, and SNAI2 in BCCs. Materials and Methods: Thirty-five fresh non-metastatic BCC tissue samples and seven fresh normal skin tissue samples were evaluated by real-time RT-PCR. Results: BMI1 and TWIST1 demonstrated marked down-regulation (p<0.00l, p=0.00l respectively), but SNAI2 showed no significant change (p=0.12). Conclusions: Previous literature has clearly demonstrated a positive association between BMI1 and TWIST1 expression and metastatic BCC, aggressive SCC and melanoma. Here, we demonstrated a negative association between BMI1 and TWIST1 mRNA expression level and BCC.


Introduction
Basal Cell Carcinoma is the most common type of skin cancer and probably one of the most common cancers in the world. BCC is well known for its slow growing feature within the epidermis of skin (Moad et al., 2012). BCC is locally invasive and rarely metastatic with a high rate of recurrence. Why does BCC rarely metastasize in comparison with other skin cancers such as Melanoma and Squamous Cell Carcinoma (SCC)? The Hedgehog (Hh) pathway has been proven to be involved in pathogenesis of BCC, but this pathway cannot solely explain metastatic inefficiency and restriction to local invasion (Schwartz and Pirrotta, 2007). Many studies have been conducted on molecular causes of cancer metastasis and have revealed certain transcription factors which act in Epithelial Mesenchymal Transition (EMT) pathway such as TWIST1, SNAIL and SNAI2 (Zhu et al., 2013). In EMT, epithelial cells undergo transition into mesenchymal phenotype and are believed to play a critical role in aggressiveness and metastasis (Zhu et al., 2013). Most EMT related molecules show up-regulation in metastatic cancer and have been investigated especially in SCC (Mikheeva et al., 2010;Guo et al., 2012;Reinisch et al., 2007;Mihic et al., 2007;Gao et al., 2013;Wang et al., 2013)). Since BCC rarely metastasizes, it can be a proper model to study the inhibitory mechanisms of metastasis. Therefore, we decided to test the expression of transcription factors which have been demonstrated to have a major role in aggressiveness of other skin cancers.
SNAI2 and TWIST1 cooperatively with BMI1 are demonstrated to be the important transcription factors in EMT and most of these transcription factors regulate each other. BMI1 (B lymphoma Mo-MLV insertion region 1 homolog) is a component of the polycomb repressive complex 1 which epigenetically silences genes (Vander et al., 1994). BMI1 was first identified as an oncogene in mouse lymphomas Casas et al., 2011). It has a role in proliferation and self-renewal (Vander et al., 1994). BMI1 expression has been demonstrated to be regulated by TWIST1 in Squamous Cell Carcinoma (SCC) cell lines (Guo et al., 2011). It was shown that BMI1 and TWIST1 cooperatively silence the E-cadherin locus which is a marker of epithelial cells (Guo et al., 2011). TWIST1 also regulates SNAI2/SLUG (Yang et al., 2004). Interestingly, it has been proven that SNAI2 is essential for TWIST1 function (Yang et al., 2004). The BMI1-TWIST1-SNAI2 axis has a role in EMT (Mihic et al., 2007;Reinisch et al., 2007;Mikheeva et al., 2010;Guo et al., 2012). The significance of this axis in Basal Cell Carcinoma (BCC) was not previously explored though there is one study, to the best of our knowledge, documenting elevated BMI1 expression via immunostaining in sixteen BCC cases (Bachmann et al., 2008). In two separate studies, expression of BMI1 in Melanoma was reported with conflicting results, but ample evidence provides strong support for BMI1 involvement in endowment of aggressive behavior to cancerous cells. Recently, one study demonstrated use of BMI1 expression in identification of Merkel Cell Carcinoma (MCC) subtypes, a rare and highly aggressive skin cancer (Kouzmina et al., 2012). However, mRNA expression level of BMI1, TWIST1, and SNAI2 was not previously quantitated in a large group of BCC cases in comparison with normal skin.
Since BCC rarely metastasizes and is one of the least dangerous cancers in the world, we hypothesized BCC would be a good model system to elucidate mechanisms which inhibit processes that drive tumor metastasis. The aim of this pilot study was to examine the mRNA expression level of BMI1, TWIST1 and SNAI2/SLUG in basal cell carcinoma to generate a more complete picture of these genes in a non-metastatic and non-aggressive cancer.

Patients
We collected thirty-five fresh tissue samples from individuals diagnosed with basal cell carcinoma in 2010 at the Tumor Clinic and Reconstructive Surgery Center of the Razi Dermatology Hospital, Tehran, Iran. The procedures followed were in accordance with the ethical standards of Tehran University of Medical Sciences and the Helsinki Declaration of 1975, as revised in 1983. Our collection of BCC cases included twelve cases of nodular, four cases of superficial, five cases of infiltrative, nine cases of micronodular, three cases of meta-typical and two cases of the basosquamous type tumor. The thirty-five BCC cases contained twenty-eight males and seven females with an average age of 65±2. Seven fresh normal skin tissues were obtained from cosmetic surgeries such as face lifts. All cases were reviewed by a dermatopathologist to confirm the diagnosis and the subclassification.

Real-time RT-PCR
RNA was extracted from tissue samples using the Tripure isolation reagent (Roche, Mannheim, Germany) according to the manufacturer's instructions. cDNA synthesis was performed using the PrimeScript RT reagent (Takara Bio Inc, Shiga, Japan) according to the manufacturer's protocols. For expression analysis, primers were designed using the NCBI. Primer-design tool (Table   1). Quantitative Real-time PCR was carried out using the SYBR Premix Ex Taq TM (Takara Bio Inc, Shiga, Japan) on rotor gene 6,000 corbette detection system. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene expression was used as a housekeeping gene for normalization of data. All experiments were performed in duplicates.

Statistical analysis
Statistical analysis was performed using the SPSS statistical package (SPSS, Chicago, IL, U.S.A.). The statistical importance of differences was assessed using the χ 2 test. p values less than 0.05 were considered as statistically significant. The specificity and sensitivity of data was determined by the Receiver Operating Characteristic (ROC) curve.

Results
Mean Relative Expression (MRE) of all samples is reported in Figure 1. BCC samples were divided into two groups based on invasive and aggressive behavior as evaluated by a dermatopathologist; highly invasive subtypes were included in the high risk group (micronodular, infiltrative and meta-typical) and the remaining subtypes were included in the low risk group. Although BMI1 and TWIST1 mRNA expression decreased in BCC samples compared to controls (p<0.00l and p=0.00l, respectively), there seems to be no significant BMI1, TWIST1 and SNAI2 MRE difference between high risk and low risk BCC groups. SNAI2 expression did not demonstrate a significant difference between BCC and controls. Expression pattern of BMI1 and TWIST1 seems to be corroborative and in line with previous studies demonstrating TWIST1 regulation of BMI1.
The ROC curve was used to assess the sensitivity and   (Table 2). In addition, TWIST1 with 94% sensitivity, has a lower specificity of 80% in comparison with BMI1.

Discussion
Primary tumors are responsible for 10% of deaths but metastases are responsible for 90% of cancer deaths. BMI1, as a stemness marker, and TWIST1, as a mesenchymal marker, have been implicated in aggressive behavior of cancers. Stemness markers and mesenchymal markers are involved in Epithelia-Mesenchymal Transition which occurs in cancer metastasis. Strong evidence has established a positive association between BMI1 and TWIST1 expression and metastatic cancers such as metastatic skin cancers. On the flip-side of the coin is Basal Cell Carcinoma (BCC), well known as one of the least aggressive cancers in the world. BCC is also a rare example of cancers which rarely metastasize. Thus, BCC provides a unique natural human model to monopolize in trying to understand regulatory steps involved in cancer metastasis prevention. We hypothesized BCC could be a good model system to elucidate mechanisms which inhibit processes that drive tumor metastasis.
The results confirm the study hypothesis. A stemness marker, BMI1, and a mesenchymal marker, TWIST1, clearly shown to be up-regulated in aggressive cancers are shown to be down-regulated in Basal Cell Carcinoma, our proposed model for studying natural intrinsic inhibitory metastatic steps utilized in non-aggressive cancers.
The initial utility of these results in clinical settings could be that down-regulation of BMI1 and TWIST1 mRNA could forecast metastatic inefficiency status of a patient and could be used for follow-up purposes. Further biochemical studies on signaling pathways involving and connected to BMI1 and TWIST1 could provide more details to the molecular mechanism of cancer metastasis responsible for 90% of cancer deaths.
TWIST1 which is a major mesenchymal mediator, with BMI1, a stemness molecule, cooperatively have an important role in EMT none of our BCC tumor samples were metastatic, but previous studies have demonstrated BMI1, TWIST1 and SNAI2 up-regulation in aggressive SCC (Patel et al., 2007;, SNAI2 upregulation in melanoma patient tumor samples (Shirley et al., 2012), and TWIST1 up-regulation in one case of metastatic BCC (Majima et al., 2012). In the present study, BMI1 and TWIST1 demonstrate marked down-regulation whereas SNAI2 expression shows no significant change in BCC tumors compared to controls. The limited data generated in this small group of BCC cases demonstrates BMI1 and TWIST1 dichotomy of action in locally invasive BCC in contrast to metastatic BCC, aggressive SCC, and melanoma. Local invasion and metastatic inefficiency of BCC shows no correlation with SNAI2.
BMI1 is a potential mediator of BCC carcinogenesis downstream the Hedgehog signaling pathway demonstrating down-regulation in the present study of non-metastatic BCC cases (Vander et al., 1994). It seems that BMI1 up-regulation has a role in aggressive behavior of metastatic cancerous cells. In the present study, with a small sample of non-metastatic, slow growing BCC cases, BMI1 and TWIST1 cooperatively demonstrate down-regulation. This may provide a clue for molecular differentiation between metastatic and non-metastatic cancerous cells, as perfectly illustrated for MCC subtyping via BMI1 expression (Kouzmina et al., 2012). However, more studies need to clarify these findings.
Our results suggest BMI1 as a potential biomarker for distinguishing controls from BCC tumor samples. The BMI1-TWIST1-SNAI2 axis we initially proposed was confirmed in this pilot study with limited number of cases, but further studies need to be conducted on greater number of BCC cases including metastatic BCC alongside SCC and melanoma cases. It may be a possibility that suppressed or elevated BMI1 and TWIST1 expression can distinguish locally invasive behavior from metastatic behavior. In conclusion, we have shown decreased BMI1 and TWIST1 mRNA expression level in non-metastatic BCC tissue samples in comparison with normal skin tissue samples. SNAI2 mRNA expression did not show significant change in BCC compared with controls.