Abstract
Background/Aim: Nodular and superficial are the most common subtypes of basal cell carcinoma (BCC). Signaling pathways such as Hedgehog (HH) and Wingless (WNT) signaling are associated with BCC phenotypic variation. The aim of the study was to evaluate of the expression profiles of 84 genes related to the WNT and HH signaling pathways in patients with nodular and superficial BCC. Materials and Methods: A total of 58 BCCs and 13 samples of normal skin were evaluated by quantitative real-time polymerase chain reaction (qPCR) to detect the gene-expression profile. Results: qPCR array showed segregation in BCC subtypes compared to healthy skin. PRKX, WNT3 and WNT16 were significantly (p<0.05) altered: PRKX was up-regulated, and WNT3 and WNT16 were down-regulated in nodular BCC. Conclusion: PRKX, WNT3 and WNT16 genes, belonging to the WNT signaling pathway, are involved in the tumorigenic process of nodular BCC.
Basal cell carcinoma (BCC) is a common skin tumor in humans (1). It presents slow growth and if left untreated can be locally destructive (2-4). It does not have a high mortality rate, although it is responsible for substantial morbidity, imposing a growing burden on healthcare services (4, 5). Significant biological BCC subtypes are nodular, superficial, infiltrative, sclerodermiform, micronodular and mixed BCC forms (4, 6, 7). Histological BCC subtypes may display differences in their biological behavior, such as tumor growth pattern, potential for recurrence and metastasis, histological pattern and genetic factors. In addition, it is important to consider extrinsic factors, such as site of origin, therapeutic choice and immunological state of the person with the tumor (3). Nodular BCC is the most common biopsied subtype; it usually manifests as a single lesion and mostly affects head and neck areas (8). Histologically, the tumor is a well-defined structure with precise contours; it presents basaloid cells of nodular mass separated from the dermis by a typical artefact of separation (8-10). Superficial BCC can be seen as multiple lesions. It is the second most common biopsied subtype, affecting mostly the trunk and shoulder areas. It is seen as a tumoral focus, which goes from the epidermis to papillary dermis, showing peripheral refraction around the tumor (8, 9).
The wingless (WNT), hedgehog (HH), epidermal growth factor (EGF), fibroblast growth factor (FGF), insulin-like growth factor (IGF) and transforming growth factor beta (TGFβ) signaling pathways are related to development mechanism of skin stem cells (11). WNT and HH have gained notorious attention in skin cancer, especially in BCC. The HH signaling pathway is known to be activated in most BCCs and mutations in HH genes, especially Patched 1 (PTCH1) and Smoothened, frizzled class receptor (SMO), are pivotal to the development of BCC (3, 12, 13). Target genes in which expression is up-regulated directly by HH signaling are PTCH1, GLI family zinc finger 1 (GLI1), GLI2, protein kinase cAMP-activated catalytic subunit alpha (PRKACA) and hedgehog-interacting protein (HHIP) (2, 4, 14). AKT serine/threonine kinase 1 (AKT1), IGF binding protein 4 (IGFBP4), IGFBP2, secreted frizzled-related protein 2 (SFRP2), bone morphogenetic protein receptor type 2 (BMPR2), leucine-rich repeat containing G protein-coupled receptor 5 (GPR49) and platelet-derived growth factor receptor-alpha (PDGFRα) have been reported to be up-regulated in BCC (4, 15, 16). There is evidence that WNT activation is a downstream consequence of HH signaling in tumors (17, 18). Activation of the WNT pathway, typically caused by gene mutations, leads to increased transcription of genes related to growth, proliferation, differentiation, apoptosis, genetic stability, migration and angiogenesis (17,19). It has also been reported that genes such as cyclin D2 (CCND2), calcium/calmodulin-dependent protein kinase II gamma (CAMK2G), casein kinase 2 alpha 2 (CSNK2A2), and frizzled receptors such as FZD7, FZD8 and FZD2 were deregulated in BCC (4). Larsimont et al. used a mouse model of BCC to show early expression of SRY sex-determining region Y-box 9 (SOX9) during tumorigenesis in a WNT-β-catenin-dependent manner. Deletion of SOX9 and constitutive activation of HH signaling prevents BCC development (20).
In the present study, the objectives were to focus on differential expression of genes belonging to the WNT and HH signaling pathways in patients with superficial and nodular BCC. Using quantitative real-time polymerase chain reaction (qPCR) array techniques and validation assay by qPCR, we identified alterations in the gene expression patterns of human nodular BCC.
Materials and Methods
Human tissue specimens. The first part of the present study involved the analysis of six BCC samples (three nodular and three superficial) and three healthy human epithelial tissue biopsies (normal skin), totaling nine distinct samples (Table I). Tissue biopsies were obtained from the Service of Dermatology at the Asa Norte Regional Hospital between 2011 and 2014 in Brazil. Normal tissue biopsies were collected from patients who underwent surgery for general dermatological purposes but not for neoplasm. The samples were collected and immediately transferred to RNA later® (Invitrogen, Life Technologies) and stored at −80°C until required for analysis. BCC tumor types were confirmed by histopathological tests. The second part of the present study involved 48 samples of nodular BCC and 10 samples of healthy human skin, obtained from the same source. All pathologies (or lack thereof) were confirmed by histological examination. Samples with ambiguous histopathology were excluded. The study received approval from the Ethical Committee of the Catholic University of Brasilia (no. CEP/UCB 112/2009). The patients gave their written consent for use of their biopsy samples for this study.
RNA extraction and analysis and cDNA synthesis. Tissue disruption was carried out using a TissueLyser II (Qiagen, Hilden, Germany) with 5 mm steel beads, RNase-Free, 20 Hz, 3×3 min. Total RNA was isolated using the RNeasy Mini kit (Qiagen) according to the manufacturer's instructions. In order to evaluate the RNA integrity, 1 μl of each sample was analyzed by gel electrophoresis. The RNA amount was quantified by Qubit® spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). In addition, we used a Bioanalyzer 2100 (Agilent Technologies, Waldbronn, Germany) to ensure the quality of the samples. Genomic DNA contamination was ruled out using a gDNA random primer (Qiagen) for qPCR. 1 μg of total RNA from each sample was used to synthetize cDNA using the RT2 First Strand kit (Qiagen) or the High Capacity cDNA Reverse Transcription Kit (Thermo Fisher Scientific) according to the manufacturer's instructions. The RT2 RNA QC PCR Array (Qiagen) was used for quality control.
qPCR array. In order to analyze the gene-expression profile, we used the Human Hedgehog Signaling RT2 Profiler™ PCR Array version 3.0 (Qiagen) according to the manufacturer's instructions. This array contains 84 genes from the HH and WNT signaling pathways including genes involved in cell differentiation and multicellular organism development. Five housekeeping genes [beta-2-microglobulin (B2M), hypoxanthine phosphoribosyltransferase 1 (HPRT1), ribosomal protein L13a (RPL13A), glyceraldehyde-3-phosphate dehydrogenase (GAPDH), actin beta (ACTB)] and controls were also included on each array for genomic DNA contamination detection, RNA quality, and general PCR performance. Genorm software (available at http://medgen.ugent.be/~jvdesomp/genorm/), integrated by qbasePLUS software version 2.1 (Biogazelle NV, Zwijnaarde, Belgium) was used to identify the best housekeeping genes to analyze the gene expression profiles. qPCR array was carried out using Step One Plus (Thermo Fisher Scientific).
Validation assay by qPCR. The differential gene expression pattern was validated using qPCR with TaqMan® probes for 10 selected genes (casein kinase 1 delta (CSNK1D), PRKACA, protein kinase cAMP-activated catalytic subunit beta (PRKACB), protein kinase X-linked (PRKX), WNT inhibitory factor 1 (WIF1), WNT2B, WNT3, WNT10A, WNT10B and WNT16) based on the expression pattern and the hierarchical cluster, p-value and fold-change values. ACTB and GAPDH genes were used as housekeeping genes for reference. We compared 48 samples of human nodular BCCs with 10 samples of healthy human epithelial tissue biopsies obtained in the same manner as mentioned above. Superficial BCC was not analyzed in this phase due to low sample availability compared to nodular BCC cases. qPCR assays followed the manufacturer's instructions in a 10 μl reaction using Step One Plus (Thermo Fisher Scientific) equipment. The gene expression assays were approximately 100% efficient, with a slope value of - 3.32 and an r-value >0.99.
Data processing and cluster analysis. Gene-expression profiles were analyzed using Web-based PCR Array Data Analysis Software (Qiagen; http://pcrdataanalysis.sabiosciences.com/pcr/arrayanalysis.php). Student's t-test was used to compare average gene expressions between groups of samples. After normalization of the fold-change values, performed on Gene Cluster 3.0 software (21), unsupervised hierarchical clustering analysis was performed by the centroid linkage method. The k-means method was applied to organize genes and arrays. Similarity metric was based on Euclidean distance. Genes were then normalized and clustered using Cluster software (21) and visualized using Gene Tree View software (22). qPCR data analysis used qBasePlus software version 2.1 (Biogazelle NV, Zwijnaarde, Belgium) to acquire the relative expression of the target genes in the 48 nodular BCC samples compared to controls.
Results
Several lines of evidence implicate cross-talk between the WNT and HH pathways in tumorigenesis and suggest their involvement in progression to aggressive, invasive, and metastatic disease stages (3, 16, 23, 24). Scant data have been published regarding gene involvement in either nodular or superficial BCC tumors and no extensive analysis of gene expression modulation exists (3, 4, 13, 25).
In the current study, six BCC specimens were analyzed by qPCR array to investigate gene expression profiles. Nodular and superficial BCC samples were obtained upon excision of primary skin tumors. The results obtained demonstrated modulation of several genes belonging to the WNT and HH pathways (Table II). From the initial gene-expression analysis, we observed 30 genes with differential expression in three comparisons: nodular BCC compared to normal skin, superficial BCC compared to normal skin and nodular BCC compared to superficial BCC. In nodular BCC, 22 differentially regulated genes (p<0.05; fold change cut-off >2.5). Only patched domain containing 4 (C6ORF138) and WNT5A were up-regulated whereas the other 20 genes were down-regulated in nodular BCC compared with normal skin. Compared with normal skin 17 genes were differentially regulated in superficial BCC. Nine genes were down-regulated in superficial compared with nodular BCC. In all groups, several genes belonging to the WNT family were down-regulated in both nodular and superficial BCC samples compared to normal skin (Table II).
Hierarchical clustering of HH and WNT pathway genes in BCC. In order to select the genes for validation in a higher number of nodular BCC samples (n=48), we performed hierarchical clustering analysis of the dataset from the previous analyses. Figure 1 shows the clustering dendrogram identifying the main outcomes. It lists 63 genes with correlation higher than 0.90. Normal skin specimens formed a distinct cluster relative to the six BCC samples. Nodular BCC samples presented a different gene-regulation profile in relation to normal skin tissue, with relative similarities to superficial BCC. Nodular BCC presented up-regulation of CSNK1D, RAS-associated protein RAB23, PRKACB, PRKX, WNT5A, bone morphogenetic protein 6 (BMP6), F-box and WD repeat domain containing 11 (FBXW11), intraflagellar transport 52 (IFT52), and siah E3 ubiquitin protein ligase 1 (SIAH1); and down-regulation of PRKACA, WIF1, WNT2B, WNT3, WNT10A, WNT10B, WNT16, FK506 binding protein 8 (FKBP8), NPC intracellular cholesterol transporter 1 (NPC1), and mitogen-activated protein kinase 1 (MAPK1).
Since the WNT pathway is poorly studied in skin cancer, we decided to focus on WNT pathway genes that were highlighted in the clustering analysis. In order to reinforce the differential expression pattern in a population of nodular BCC, we performed an extended qPCR assay using tumor samples from 48 patients for analysis of 10 selected genes. The secondary qPCR analysis focused on the following genes: CSNK1D, PRKACA, PRKACB, PRKX, WIF1, WNT2B, WNT3, WNT10A, WNT10B and WNT16. WNT5A was not selected since it has been extensively studied and our data from the first experimental phase corroborated its up-regulation in nodular BCC, as described previously (26, 27). The second experimental phase confirmed PRKX as being up-regulated (p=0.01016), and WNT3 and WNT16 as being down-regulated (p=0.000081 and p=0.0117, respectively) in nodular BCC. The expression profiles of the remaining seven genes did not significantly differ.
Discussion
BCC is a cutaneous malignant tumor that is locally invasive and has the highest incidence in Caucasian individuals, but rarely causes metastasis or death (5, 8, 9). Active HH signaling is implicated in BCC malignancy (3, 24). WNT family genes are implicated in many cell processes such as cell proliferation, cell fate, polarity, differentiation and migration (18, 19). To date, more than 19 human WNT genes have bene identified and many of them have different isoforms (28). Hierarchical clustering analysis of our dataset showed nodular BCC specimens clustered together and had a different gene modulation in relation to normal skin, relatively similar to superficial BCC, in agreement with Yu et al. (4).
Our study focused on nodular BCC and confirmed up-regulation of PRKX, whereas WNT3 and WNT16 were down-regulated. Human PRKX is a type-I cAMP-dependent protein kinase and is associated with the WNT pathway (29). There is evidence that PRKX stimulates endothelial cell proliferation, migration, and vascular-like structure formation in multiple developing tissue processes related to modulation of angiogenesis (29, 30). Studies report that PRKX binds to peptidylprolyl cis/trans isomerase, NIMA-interacting 1 (PIN1), membrane-associated guanylate kinase, WW and PDZ domain containing 1 (MAGI1) and BCL2-associated athanogene 3 (BAG3), which have been implicated in a variety of cellular functions, including proliferation, apoptosis, development, differentiation and tumorigenesis (29, 31, 32). We report for the first time that the PRKX gene is up-regulated in nodular BCC. Since nodular BCC has a more aggressive phenotype than superficial BCC, it is reasonable to expect that the up-regulation observed for PRKX could be related to neoplastic transformation due to stimulation of angiogenesis. Newell et al. showed that BCC exhibited a 5-fold increase in angiogenesis compared to normal skin (33).
WNT3 belongs to the WNT1 class of ligands, which stimulate the canonical WNT/β-catenin pathway by binding to a Frizzled receptor (34). There is evidence that WNT3 also acts through a non-canonical WNT signaling pathway that involves phosphorylation of mitogen-activated protein kinases (MAPKs), the Ras-dependent extracellular-signal-regulated kinases 1/2/5 (ERK1/2/5) (35). WNT3 overexpression is implicated in tumor proliferation, inhibition of tumor apoptosis, poor prognosis and promotion of malignant transformation in cancer such as human hepatocellular carcinoma, non-small cell lung cancer, and squamous cell carcinoma of the oral cavity (34-38). It is expressed in the NT2, MCF-7 and MKN45 cell lines (39). In addition, Kimura et al. presented the hypothesis that WNT3 signaling is required for increased mantle cell lymphoma lymphomagenesis demonstrated in hematological malignancies (40). The literature suggests that WNT3 overexpression is associated with the development of aggressive malignancy during tumor progression. We studied nodular BCC, which is not a very aggressive tumor compared to those mentioned above. Our results suggested that nodular BCC exhibits a different gene expression profile from very aggressive human tumors, which would be compatible with down-regulation of WNT3 in nodular BCC.
The other differentially regulated gene was WNT16, which is involved in oncogenesis and in several developmental processes, including regulation of cell fate and patterning during embryogenesis (41, 42). WNT16 expression is found in spleen, appendix, lymph nodes, and bone marrow of patients with pre-B acute lymphoblastoid leukemia with a t(1;19) chromosomal translocation, and is implicated in lineage specification in B-cell development (43). Fear et al. identified alternative WNT16 isoforms A and B (41). The isoforms showed differential expression in adult human tissues. Teh et al. reported WNT16B but not WNT16A isoform as being up-regulated in BCC compared with normal skin (42). Nonetheless, the up-regulation seen by Teh et al. was analyzed using only three human BCC samples. In the current study, we analyzed 48 samples of nodular BCC, although the data do not differentiate between WNT16 isoforms.
In conclusion, functional studies will be needed in order to confirm the role of PRKX, WNT3 and WNT16 in nodular BCC. The gene-expression profiles obtained here give us a starting point for understanding the molecular basis of nodular BCC. Hierarchical clustering showed distinct profiles for the samples, suggesting a strong influence of the WNT pathway in nodular BCC. Taken together, our findings suggest that WNT3, WNT16 and PRKX are involved in the modulation of WNT signaling, possibly leading to an imbalance in growth, proliferation and differentiation. It will be important to investigate whether skin cancer subtypes exhibit distinct gene-expression profiles associated with WNT signaling leading to aggressive skin tumor, as compared to superficial BCC. It is tempting to speculate that WNT signaling may have an important role in differentiation of epithelial tissues. The modulation of genes belonging to the WNT pathway may contribute to the dysregulation of nodular BCC differentiation as observed in skin cancer subtypes.
Acknowledgements
The Authors are grateful to patients and their families for authorizing the molecular assays of their tumor biopsies and normal skin biopsies in Asa Norte Regional Hospital (HRAN). We thank the dermatological staff from the Service of Dermatology from HRAN for all their help. Rosângela Vieira de Andrade Ph.D. had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
- Received July 20, 2016.
- Revision received August 15, 2016.
- Accepted August 18, 2016.
- Copyright© 2016 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved