Skin cancer is the most frequent malignancy in the world (1). It constitutes a public health problem and causes high costs for the health system (2,3). In Colombia, there was an increase in new cases of basal cell carcinoma (BCC) from four new diagnoses per 1,000 individuals in 2003 to 11 per 1,000 individuals in 2005, and squamous cell carcinoma (SCC) increased from one to three new diagnoses per 1,000 individuals in the same period (4). The risk factors related to developing these two types of cancer in our population are having skin phototypes 1 to 3, the presence of actinic keratosis, a family history of skin cancer, living in rural areas and working outdoors (5). However, these factors do not fully explain why some people develop skin cancer and others do not. Therefore, understanding the genetic factors involved in tumor development is of great importance for elucidating skin cancer biology. Non-melanoma skin cancer (NMSC) is divided into SCC, BCC, and other less common cancers (6). BCC is the most frequent skin cancer (80%), (7) and it is caused by intermittent exposure to ultraviolet radiation (UVR), especially during youth (8). SCC is less frequent but is a more aggressive tumor, with a greater probability of metastasis and eventual death (9). SCC is more common among people who work outdoors and are exposed to UVR (8).

UVR induces direct and indirect DNA damage, and these DNA lesions must be repaired to prevent the onset of carcinogenic events (10-12). A key molecule during tumor development is the p53 protein, encoded by the TP53 gene. In response to cellular stress, p53 modulates the transcription of genes involved in cell cycle arrest, apoptosis, DNA stability and the inhibition of angiogenesis (13-15). Therefore, the alteration of these pathways due to mutation or loss of TP53 predisposes individuals to tumor development (16). The TP53 mutations associated with cancer are very diverse and are located throughout the entire sequence of the gene (17-19). However, the most important mutations are located in exons 5 to 8, which encode the DNA-binding domain, and these mutations impair the DNA binding ability (17,20). In NMSC, TP53 mutations have been described in up to 50% of cases (21,22).

High-resolution melting (HRM) scans for mutations by using polymerase chain reaction (PCR) assays. It has been used as an alternative strategy for the screening of variants in DNA due to differences in DNA melting temperatures and curve profiles from samples controls (23). This method is fast and much cheaper than other mutation genotyping technologies and does not require post-PCR handling (24). The aim of this study was to describe mutations in exons 5-8 of the TP53 gene in a sample of Colombian patients with non-melanoma skin cancer using PCR-High resolution melting as prescreening method.

Patients and samples. A descriptive cross-sectional study was carried out. We included 116 subjects older than 18 years and diagnosed with non-melanoma skin cancer (58 with BCC and 58 with SCC) between 2012 and 2014. Samples of tumor material from each patient were embedded in paraffin (TTEP) and fixed in formalin. The study was conducted in accordance with ethical standards for human research. The patients were asked to participate voluntarily and signed their consent. The study and the informed consent were approved by the Research Ethics Committee of the Hospital Universitario-Centro Dermatológico Federico Lleras Acosta, E.S.E (Bogota, Colombia).

Cell culture. The cell lines used as controls were HT-29 (Human Colorectal Adenocarcinoma; c.818G>A; exon 8), NCI-H727 (Human lung carcinoma, non-small cell; c.492_493ins9 exon 5), DU145 (c.820G>T; exon 8), HEK-293 (Human kidney; wild type), AU565 (Breast carcinoma; c.524G>A), NCI-H82 (Small cell lung cancer; c.375G>T), NCI-H520 (Lung squamous cell carcinoma; c.438G>A) and PC3 (Prostate adenocarcinoma; c.413del1). Cell lines were kindly donated from Fabio Aristizabal (Universidad Nacional de Colombia, Bogotá, Colombia).

All cell lines were cultured in Corning Glass Culture Flasks (Sigma-Aldrich, St. Louis, MO, USA). All cell lines except HEK-293 were cultured in RPMI 1640 culture medium (Gibco, Life Technologies Corporation, Grand Island, NY, USA), supplemented with 10% fetal serum bovine (v/v). Cultures were maintained at 37˚C, in a humid atmosphere with 5% CO₂. HEK293 cells were cultured using DMEM supplemented with 5% fetal serum bovine (v/v) and streptomycin -penicillin (Gibco, Life Technologies Corporation, Grand Island, NY, USA). Genomic DNA was extracted from confluent cultures (70-80%) and used as control in the PCR-High Resolution Melting.

DNA extraction. From 7 μm-thick histological sections of TTEP stained with haematoxylin and eosin, the tumor tissue was microdissected using light microscopy to ensure 80% morphological tumor characteristics when possible. The DNA was extracted using the commercial QIAamp DNA FFPE Tissue Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. The DNA integrity was verified by the absorbance ratio at 260/280 nm wavelength measured by NanoDrop 2000c (Thermo Scientific, Waltham, MA, USA), and amplification of the GAPDH gene and electrophoresis on 1% agarose gels. Also, DNA from control cell lines were extracted using the same method.

Screening of mutations by PCR-HRM. High resolution melting (HRM) analysis was used to screen mutations in exons 5-8 of the TP53 gene. PCR and HRM were performed in a CFX96 system thermocycler (Bio-Rad Laboratories, Hercules, CA, USA). For HRM of exons 7 and 8, a single segment was amplified per exon using the primers published by Krypuy et al. (23). For exon 5, three segments were amplified using the primers described by Mitchell et al. (24). For exon 6, primers were designed in Primer3 (25) (Table I). All analyses were performed in duplicate in 96-well plates with 2 ng of DNA from each sample for a final volume of 20 μl. The Precision Melt Supermix (Bio-Rad) kit was used for HRM. Amplification conditions included an initial denaturation of 5 min at 95˚C, followed by 45 cycles of 95˚ for 10 s, annealing according to Table I and by 10 s at 72˚C. The samples were kept at 60˚C for 30 s, and a denaturation curve was then generated by increasing the temperature by 0.1˚C/s. Electrophoresis was performed on 2% agarose gel stained with SYBR-Safe (Thermo Fisher Scientific) to verify the size and identify the amplified products.

DNA sequencing analysis of TP53 exons 5-8. The sequence of exons 5-8 was determined by direct sequencing from DNA extracted from the cell lines mentioned in Table II. According to the presence mutations or WT sequence in exons 5-8 in each cell line, these were used to optimize the HRM-PCR conditions of each exon as indicated in Table I. The primers used were reported by Krypuy et al. (23). The primers sequences and the conditions of PCR are shown in Table I. Samples with differences in the denaturation curves adjusted to temperature and normalized to DNA control of each cell line during the HRM protocol were subjected to amplification by conventional PCR using the primers as shown in Table II. Subsequently, the PCR products were purified using the QIAquick PCR Purification Kit (Qiagen) according to the manufacturer’s instructions, and bidirectional direct sequencing was carried out with the Sanger method using BigDye Terminator V1.1 Cycle Sequencing Kit and an ABI PRISM 3130xl Genetic Analyzer (Applied Biosystems, Waltham, MA, USA). Once the sequence was obtained, it was analyzed using the novoSNP program version 3.0 (Department of Molecular Genetics VIB and University of Antwerp) against the TP53 reference sequence extracted from the ENSEMBL database (ENST00000269305). Additionally, variants were analyzed using the Excel MUT-TP53 2.0 open-access data sheet. The variants were verified in the International Agency for Research on Cancer (IARC) mutation database for TP53 (http://p53.fr) (IARC TP53 Database).

TP53 loss of heterozygosity testing. The DNA from anti-coagulated blood was extracted to identify TP53 loss of heterozygosity (LOH). It was done using QIAamp DNA blood kit (Qiagen) according to the manufacturer’s instructions. Peripheral blood DNA samples of patients that showed variants in TP53 in tumor tissue were amplified and sequenced.

The average age was 61 years for BCC patients and 68 years for SCC patients. 46% (27/58) of patients diagnosed with BCC and 39% (23/58) with SCC were women. 35% of BCC patients (20/57) and 57% of SCC (33/58) patients reported working outdoors. 47% (27/57) and 60% (34/57) of the patients with BCC and SCC, respectively, reported having a family member with a history of cancer. Family history of gastric cancer was the most frequent antecedent, in 25% of patients with BCC (7/57) and 23% (8/34) with SCC. Twenty-three percent (8/34) of patients diagnosed with SCC reported a family history of melanoma.

None of the patients studied was immunosuppressed. Likewise, none of the patients presented with metastasis. Only one patient with SCC had a history of radiotherapy prior to the tumor diagnosis. Concerning BCCs, 98.2% (56/57) were primary; 64% of these cases were nodular and 31% were mixed. 43% of the patients with SCC histopathology had tumors with moderate differentiation, while 32% of SCC patients had well-differentiated tumors.

To detect samples with probable mutations in exons 5-8 of TP53, 115 samples from patients diagnosed with BCC and SCC were analyzed using HRM (Figure 1). The variants present in the cell lines used as controls for the PCR-HRM analysis are found in Table II. The subsequent direct sequencing of samples with differences in the HRM analysis identified 10 pathogenic variants, 4 variants of uncertain significance (VUS) and one benign. Of the variants detected, 80% were substitutions and 20% were deletions (Figure 2 and Table III). 16% of patients diagnosed with BCC and 26% of patients with SCC had variants in the TP53 gene sequence. The most recurrent mutation in these patients was the missense mutation p.R158P (c.472C>G) resulting in an amino acid substitution from Arginine to Glycine. This mutation was found in 12% of patients with BCC and 10% with SCC. Of the samples with TP53 mutations, the most frequent histological subtype was nodular in BCC and poorly differentiated in SCC (Table IV). Any of the variants found in the tumor tissue were found in the matched peripheral blood DNA of each patient.

Non-melanoma skin cancer (NMSC; BCC and SCC) has several risk factors, among which are prolonged UVR exposure, skin type, gender, and ability to repair DNA (5,22,26). A key molecule during carcinogenesis is the p53 protein, encoded by the TP53 gene (13,14).

UVR produces a skin response because it induces the formation of cyclobutane pyrimidine dimers (CPDs) and pyrimidine-pyrimidone photoproducts (6-4 PPs) (9,12). Despite the clear relationship between TP53 mutations and the risk of developing skin cancer (18,19,27,28), the mutation rates reported in the literature vary according to the population analyzed (29). In this study, tumor tissue was obtained from biopsies of patients with BCC and SCC diagnoses, and samples were screened for mutations in exons 5 through 8 of TP53 with HRM analysis and subsequent direct sequencing of the DNA fragments that differed from control cell lines in their denaturation curves.

Among the 115 samples analyzed, 16% of the BCCs and 26% of the SCCs had mutations in exons 5 to 8. This frequency is similar to that reported for populations in China (20% BCC) (30) and Korea (33%) (31), while it is lower when compared to populations such as England and Sweden (58%) (32) and the United States (between 45 and 58%) (33,34).

Although Colombia is a tropical country, it has an estimated incidence of skin cancer that is lower than other locations with fewer geographic risk factors. It is possible that skin cancer rates in Colombia are not as high as those reported in Australia or some European countries because of phototype and individual genetic susceptibility. It may be that a combination of certain phototypes with continuous but controlled exposure (not intense enough to produce sunburn) generates skin changes, such as increased melanin production and thickening of the corneum stratum, that protect the skin from the deleterious effects of UVR (5,32). In addition, the genetic background of Colombian populations is shaped by the mixture of native, European and African migration patterns. This heterogeneity may have an impact on the genetic susceptibility to the development of squamous cell carcinoma or basal cell carcinoma (35).

In this work, we found that 80% of the mutations are substitutions, which is in agreement with the data reported in the IARC database. The most studied genetic alterations are in the region encoding the DNA-binding domain, which includes exons 5 to 8. The majority of the mutations in this domain (87.9%) are substitutions that generate missense variants (18,36). These mutations generate proteins that form a p53 tetramer that is less competent to specifically bind DNA, an effect that has been correlated with loss of gene function (18).

On the other hand, frameshift mutations represent 11% of the mutations reported in all types of cancer (37). In this study, deletions were detected in three patients; one of the deletion mutations eliminated 10 nucleotides, resulting in a frameshift mutation and the generation of a premature stop codon.

The most frequent mutations in non-melanoma skin cancer are C:G,T:A, and CC:TT substitution in pyrimidine dimers; this mutation type represents 80% of all TP53 mutations (31,38,39). Twenty-five percent of these substitutions are found in CpG dinucleotides, which have a 10-fold greater mutation frequency than other dinucleotides. The cytosines in CpG dinucleotides are methylated in normal tissue and are frequently deaminated relative to non-methylated cytosines. CpG substitutions generate poor base pairing and represent at least a third of the known TP53 substitutions (37). In this work, the C>T and G>A substitutions represented 47% of the point mutations, indicating that the TP53 mutations in these individuals may be related to sun exposure.

The majority of TP53 mutations reported here are in the DNA-binding domain (DBD). This result is likely because the DBD region contains 22 CpG dinucleotides and three codons (175, 248, and 273) that are considered to be mutation hot spots (40). Likewise, the effect of each genetic variant on protein function depends on its site within the DBD and the type of amino acid substituted. Changes in Loop1 (L1) (residues F113- T123), Loop2 (L2) (residues K164-C176), and Loop3 (L3) (M237-P250) as well as in the amino acids involved in the coordination of zinc binding (R175, C176, H179, C238 and C242) affect the antiproliferative function of p53 and the thermodynamic stability of the DBD (41). In the samples analyzed in this study, a group of mutations located between L1 and L2 (139, 142, 151, 153, 155, 156, 158, and 159) was found, indicating a pattern of variants different from that reported in the literature; however, since the substitutions introduce amino acids with different polarity, these variants impact the protein function.

Three mutations in the DBD region were detected that generate stop codons and frameshifts (p.K139Xfs, p.P153Xfs, and p.P155Xfs); these variants generate premature proteins lacking the DNA-binding segment, and they likely play an important role in carcinogenesis. The p.V173M and p.H179Y mutations detected are located within Loop2 of the DBD; specifically, they are in residue 179, which binds to zinc and coordinates the binding to DNA. These mutations are considered disruptive mutations and impact protein function (41).

The mutation hot spot located at codon 273 was found in two patients diagnosed with SCC. The change from arginine to histidine at position 273 (p.R273H) causes reduced selectivity of DNA-binding compared to the normal protein (40). This variant is associated with carcinogenesis because it participates in invasion, increased migration, increased cell proliferation, and drug resistance, thus preventing the death of the tumor cell (42).

Pfeifer et al. (2005) showed that mutation hot spots in skin cancer are found in codons 152, 158, 196, 213, 245, and 282, which contain methylated cytosines and are altered by UVR (44). In contrast, Giglia-Mari and Sarasin report that codon 177 is the most frequently mutated in BCC and codon 278 in SCC (22). In the samples analyzed in this work, we found that codon 158 (p.R158G; c.472C>G) had the highest mutation rate. This codon is between L1 and L2 of the DBD. The prediction analysis on the impact of protein function shows a score of 98.26, indicating loss of function (43,44).

The discovery of the recurrent p.R158G mutation is important due to its high frequency in this population; 12% (7/57) of the BCC patients and 9% (5/58) of the SCC patients had this mutation. This mutation is reported only 21 times in the p53 database and decreases protein activity by approximately 90% (IARC) (25), which could suggest a preponderant role in the development of NMSC in the population studied. Future studies with a larger sample size are required to verify whether this is a high-frequency (hot spot) variant in our population.

In 80% of patients with identified TP53 mutations, only one mutation was detected, while the remaining 20% (5/24) had more than one mutation found. The presence of multiple mutations has been reported in human skin cancer, human head and neck cancer, and mouse skin cancer induced by UVR (42,45-48). This finding is explained by continuous exposure to UVR generating multiple mutational events. On the other hand, the cells that make up the tumor come from multiple clones each with different mutations (49).

This study shows that the frequency of TP53 mutations in tumor tissue from a group of Colombian patients diagnosed with BCC and SCC was similar to that reported for other geographic areas. The sample analyzed has a different pattern of mutations than that reported by other authors for this type of cancer; our discovery of the recurrent p.R158G mutation is most relevant, as it suggests that this mutation could have an important role in the development of SCC and BCC. This study screened for mutations in exons 5 to 8 and did not exclude the presence of mutations in the other exons of the gene.

The informed consent, clinical data, and molecular biology results of this study are restricted by the ethics committee of the Hospital Universitario-Centro Dermatológico Federico Lleras Acosta E.S.E. and are available from the file with the code 4000-16.6W. These files can be requested in the following e-mail: Comitedeeticaeninvestigacion@dermatologia.gov.com .

The Authors declare no conflicts of interest.

Conceptualization: Luz D Gutiérrez-Castañeda; methodology: Maria Irene Cerezo Cortés and Luz D Gutiérrez-Castañeda; software: Maria Irene Cerezo Cortés and Luz D Gutiérrez-Castañeda; validation: Luz D Gutiérrez-Castañeda; formal analysis: Maria Irene Cerezo Cortés and Luz D Gutiérrez-Castañeda; investigation, Maria Irene Cerezo Cortés and Luz D Gutiérrez-Castañeda; resources: Luz D Gutiérrez-Castañeda; data curation: Maria Irene Cerezo Cortés and John Nova; writing – original draft preparation: Maria Irene Cerezo Cortés and Luz D Gutiérrez-Castañeda; writing – review and editing: Maria Irene Cerezo Cortés, Luz D Gutiérrez-Castañeda and John Nova; supervision: Luz D Gutiérrez-Castañeda; project administration: Luz D Gutiérrez-Castañeda; funding acquisition: Luz D Gutiérrez-Castañeda. All Authors have read and agreed to the published version of the manuscript.

This work was funded by Hospital Universitario-Centro Dermatológico Federico Lleras Acosta, E.S.E., D C, Bogotá, Colombia (project Grant 4000-16.6W). We acknowledge the Hospital Universitario Centro Dermatológico Federico Lleras Acosta, E.S.E. D.C, Colombia for supporting this study. We thank the patients who gave their informed consent for participating in this study. We also thank the staff of the diagnostic support department for the collaboration in obtaining patients.