Molecular Medicine Reports Molecular Medicine Reports 1791-2997 1791-3004 D.A. Spandidos 10.3892/mmr.2022.12771 MMR-26-02-12771 Articles A novel 7-hypoxia-related long non-coding RNA signature associated with prognosis and proliferation in melanoma LuoYi 1 LiTinghao 2 ZhaoHengguang 3 ChenAijun 1 Department of Dermatology, The First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, P.R. China Department of Urology, The First Affiliated Hospital of Chongqing Medical University, Chongqing 400016, P.R. China Department of Dermatology, The Second Affiliated Hospital of Chongqing Medical University, Chongqing 400010, P.R. China Correspondence to: Dr Aijun Chen, Department of Dermatology, The First Affiliated Hospital of Chongqing Medical University, 1 Youyi Road, Yuzhong, Chongqing 400016, P.R. China, E-mail: chenaijunhx@126.com 08 2022 14 06 2022 26 2 255 16022022 12052022 Copyright: © Luo et al. 2022 This is an open access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-commercial and no modifications or adaptations are made.

Hypoxia-related long non-coding RNAs (lncRNAs) are important indicators of the poor prognosis of cancers. The present study aimed to explore the potential relationship between melanoma and hypoxia-related lncRNAs. The transcriptome and clinical data of patients with melanoma were downloaded from The Cancer Genome Atlas database. The prognostic hypoxia-related lncRNAs were screened out using Pearson's correlation test and univariate Cox analysis. As a result, a hypoxia-related-lncRNA signature based on the expression of 7 lncRNAs was constructed, with one unfavourable [MIR205 host gene (MIR205HG)] and six favourable (T cell receptor β variable 11-2, HLA-DQB1 antisense RNA 1, AL365361.1, AC004847.1, ubiquitin specific peptidase 30 antisense RNA 1 and AC022706.1) lncRNAs as prognostic factors for melanoma. Patients with melanoma were divided into high- and low-risk groups based on the risk score obtained. Survival analyses were performed to assess the prognostic value of the present risk model. Potential tumour-associated biological pathways associated with the present signature were explored using gene set enrichment analysis. The CIBERSORT algorithm demonstrated the important role of the hypoxia-related lncRNAs in regulating tumour-infiltrating immune cells. Clinical samples collected from our center partly confirmed our findings. Cell Counting Kit-8 and flow cytometry assays indicated the suppression of proliferation of melanoma cells following inhibition of MIR205HG expression. Indicators of the canonical Wnt/β-catenin signalling pathway were detected by western blotting. The present study demonstrated that MIR205HG could promote melanoma cell proliferation partly via the canonical Wnt/β-catenin signalling pathway. These findings indicated a 7-hypoxia-related-lncRNA signature that can serve as a novel predictor of melanoma prognosis.

 hypoxia melanoma long non-coding RNA MIR205 host gene proliferation Funding: No funding was received. Introduction

Melanoma is the 19th most common cancer diagnosis worldwide, representing <5% of skin cancers and accounting for 80% of skin cancer-related deaths (1). Compared with other cancer types, the global incidence of cutaneous melanoma has been increasing annually at a more rapid rate (2). Every year, ~160,000 new cases of invasive melanoma and 48,000 deaths due to the disease are recorded worldwide (3). The oncogenesis and developmental mechanisms of melanoma remain unclear and require further exploration. Furthermore, it is of vital importance to explore efficient biomarkers for early and accurate diagnosis.

Tumorigenesis and malignancy development are complex processes affected by various factors. Due to genetic changes and epigenetic defects, the tumour microenvironment, which consists of cellular and non-cellular components, also serves an important regulatory role in cancer progression (4). Tumour hypoxia is a condition wherein solid tumours have lesser oxygenation than that observed in normal tissues (5), which results in the obstruction of therapy and the progression of malignancy (6). Various studies have demonstrated the crucial role of hypoxia in the malignant behaviour of melanoma cells, including abnormal proliferation, distant metastasis and immune evasion (7–9). As a type of RNA that does not encode proteins and has a length >200 nucleotides, long non-coding RNAs (lncRNAs) have been explored in the occurrence and development of numerous tumours (10,11). lncRNAs participate in the modification of proteins that affect the survival and proliferation of cancer cells and thereby serve a role in cancer cell survival and development, even in hypoxic circumstances (12). It has been reported that hypoxia yield proliferation-associated lncRNA acts as a competing endogenous RNA (ceRNA) that binds to microRNA (miRNA/miR)-431-5p under hypoxic conditions to upregulate CDK14 expression, resulting in gastric cancer cell proliferation (13). In colorectal cancer cells, lncRNA COL4A2 antisense RNA 1 facilitates growth and glycolysis under hypoxia via the miR-20b-5p/hypoxia-inducible factor 1α subunit axis (14). Several studies have also explored the relationship between malignant melanoma progression and hypoxia-related lncRNAs, including long intergenic non-protein coding RNA 518 and non-coding RNA activated by DNA damage (15,16). Therefore, lncRNAs may serve a crucial role in hypoxia-related malignancy development and poor melanoma prognosis.

lncRNA-based signatures have recently drawn attention due to their high predictive accuracy (17,18). However, hypoxia-related lncRNAs as biomarkers for prognostic prediction in melanoma are yet to be investigated. The present study aimed to build a hypoxia-related-lncRNA signature and nomogram that can improve the overall survival (OS) predictability in patients with melanoma using bioinformatics. The present study identified a 7-hypoxia-related-lncRNA signature through a comprehensive analysis of The Cancer Genome Atlas (TCGA) database and two hypoxia-related gene sets. The present signature was verified and analysed by combining the clinical features of melanoma, and validated using clinical samples and in vitro experiments.

 Materials and methods <sec> <title>Data collection

The transcriptome expression profiles corrected by FPKM (fragments per kilobase of exon per million mapped fragments) and clinical information were downloaded from TCGA database (https://portal.gdc.cancer.gov). The details of the clinical characteristics, including OS, survival time, age, sex, stage, tumour size, distant metastasis and lymph node metastasis, of the included cohort of 471 patients are listed in Table SI. Among these 494 patients, 23 patients with incomplete data or vague living status were excluded from the present study. The hypoxia-related gene expression profiles KRIEG and WINTER were downloaded from the Gene Set Enrichment Analysis (GSEA) database (http://www.gsea-msigdb.org) and are listed in Table SII. The data used in the present study were downloaded on May 21, 2021.

 Establishment of a hypoxia-related-lncRNA signature as a prognostic model

Pearson's correlation analysis was used to identify hypoxia-related lncRNAs. The correlation between lncRNA expression and hypoxia-related gene expression was calculated. The selection criteria included values with |R2|>0.7 and P<0.001. Univariate Cox regression analysis was used to screen for prognosis-associated hypoxia-related lncRNAs. A hazard ratio (HR) <1 represented favourable OS outcomes, and a HR >1 represented poor OS outcomes. lncRNAs with P<0.05 were selected as hypoxia-related lncRNAs and used to construct the present prognostic model.

 Evaluation of the hypoxia-related-lncRNA prognostic signature

Multivariate Cox regression analysis was used to create a signature, and the risk score formula was as follows: (RiskScore) = ∑i=1n(Expi ⁎Coei). In the aforementioned formula, ‘n’ indicates the number of prognostic genes, ‘Expi’ indicates the level of lncRNA I expression, and ‘Coei’ indicates the regression coefficient of the corresponding lncRNA obtained using the multivariate Cox regression model. Patients with melanoma were divided into high- and low-risk subgroups according to the median risk score. A survival curve was used to compare the OS of patients between the high- and low-risk groups. The diagnostic efficacy and clinicopathological characteristics of the 7-hypoxia-related-lncRNA signature were evaluated using receiver operating characteristic (ROC) curves. The risk score efficiency of the present signature to independently predict the survival rate was assessed using univariate and multivariate Cox regression analyses.

 Estimation of clinical independence and construction of the nomogram

The ‘rms’ R package was used to integrate the clinical features with the risk scores obtained. The present 7-hypoxia-related-lncRNA signature was used to construct a nomogram for further clinical prediction. Variates of clinical features were categorized into dichotomous variables and defined by Cox regression analysis. Samples with incomplete clinical information were excluded to make the results consistent. A horizontal straight line was drawn to ascertain the points for each variable by variables feature. Subsequently, the total points of each patient were calculated by adding the points for all variables together, followed by normalization to a distribution of 0 to 100. The performance of the present nomogram was evaluated using calibration plots and time-dependent ROC curves.

 Principal component analysis (PCA)

PCA was conducted using the ‘scatterplot3D’ R package to illustrate the level of expression of melanoma samples in the low- and high-risk groups. Patients were divided into different subgroups using different gene sets, including the present 7-hypoxia-related-lncRNA gene set, hypoxia-related-lncRNA gene set, hypoxia-related gene set and whole gene sets.

 Construction of the co-expression network

The lncRNA-miRNA-mRNA co-expression network was constructed using Cytoscape (version 3.8.2) to elucidate the association of the hypoxia-related lncRNAs with their target miRNAs and downstream mRNAs. The lncRNA-miRNA connection was predicted using data downloaded from miRcode (http://www.mircode.org). TargetScan (version 8.0) (https://www.targetscan.org), miRDB (http://www.mirdb.org) and miRTarBase databases (http://mirtarbase.mbc.nctu.edu.tw/php/index.php) were used to screen for miRNAs and their corresponding target mRNAs. The selected mRNAs were compared with hypoxia-related mRNAs to obtain reliable hypoxia-related lncRNAs.

 GSEA

GSEA was employed to identify enriched gene sets in the high-risk groups. The Gene Ontology-Biological Process (GO-BP) ontology gene sets (c5.go.bp.v7.4.symbols.gmt) were downloaded from the Molecular Signatures Database (http://www.gsea-msigdb.org). Gene set permutations were performed 1,000 times for each analysis.

 Immune cell infiltration analysis

The CIBERSORT algorithm of the R software was used to determine the profile of tumour-infiltrating immune cells (TIICs; including 22 immune cells) in all tumour samples. The relationship between TIICs and risk scores was analysed using the limma package. Comparisons of the TIIC levels between the high- and low-risk groups were estimated using the unpaired Student's t test.

 Cell culture

A375 human melanoma cells were purchased from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences. Cells were incubated in DMEM (Gibco; Thermo Fisher Scientific, Inc.) supplemented with 10% FBS (Gibco; Thermo Fisher Scientific, Inc.). The cells were seeded in cell culture plates and cultivated in a humidified incubator at 37°C with 5% CO2.

 Human tissue specimens

The human melanoma specimens were provided by the Department of Dermatology of the Chongqing Medical University (Chongqing, China) from October 2021 to March 2022. The present study was performed using 20 melanoma samples and their paired adjacent tissues, which were collected from 10 males and 10 females, with age range from 49 to 58 years old. Inclusion criteria were as follows: People who were diagnosed with melanoma by pathological examination at first time with medical history which was detailed enough to meet the statistical needs of this trial. Exclusion criteria were as follows: People who were diagnosed as melanoma or treated previously; people with other special conditions (such as severe chronic secondary fungi or bacterial infection); and people with metastatic melanoma. The expression levels of three lncRNAs were detected using reverse transcription-quantitative PCR (RT-qPCR). The present study was approved (approval no. 2021-487) by the Institute Research Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (Chongqing, China). All individuals provided written informed consent for the use of their samples for clinical research.

 Small interfering RNA (siRNA/si) interference assay

All siRNAs targeting human MIR205 host gene (MIR205HG, Gene ID: 642587; si-MIR205HG-1 and si-MIR205HG-2) and scrambled negative control (si-NC) were designed and synthesized by Shanghai GenePharma Co., Ltd. and are listed in Table I. The siRNAs were transfected into A375 cells using Lipofectamine® 2000 (Invitrogen; Thermo Fisher Scientific, Inc.) according to the manufacturer's protocol. In brief, A375 cells were seeded into 6-well plates at a density of 1.2×105 cells per well and routinely cultured until the confluence reached 70%. Then the medium was changed to 2 ml DMEM without serum, supplemented with 5 µl Lipofectamine® 2000 and 5 µl (20 µM) siRNA to transfect for 6 h at 37°C with 5% CO2. Subsequent experimentation was performed 48 h after transfection. The efficiency of MIR205HG knockdown was verified using RT-qPCR.

 RT-qPCR

Total RNA was extracted from the two groups using TRIzol® reagent (Takara Bio, Inc.) according to the manufacturer's protocol. First-strand cDNA synthesis was performed using the PrimeScript RT reagent kit (Takara Bio, Inc.). The following thermocycling conditions were used for RT: 25°C for 5 min, 42°C for 40 min and 85°C for 2 min. Subsequently, qPCR was performed using an ABI7500 real-time PCR system using SYBR green (Takara Bio, Inc.). The qPCR thermocycling conditions were as follows: 94°C for 2 min, followed by 40 cycles of 94°C for 30 sec and 55°C for 45 sec. All RT-qPCR primers were purchased from Shanghai GenePharma Co., Ltd. and the sequences are listed in Table I. Each sample was detected at least three times independently. The 2−ΔΔCq method was utilized to quantify the expression levels (19).

 Cell proliferation assay

The growth rate of A375 cells was assessed using a Cell Counting Kit-8 assay (CCK-8; Boster Biological Technology). A375 cells transfected with different siRNAs were seeded into 96-well plates (5,000 cells per well). After incubation for 48 h at 37°C with 5% CO2, 10 µl CCK-8 reagent was added to each well. The absorbance of the cell medium at 450 nm was recorded to assess the cell viability in each group.

 Cell cycle analysis

A375 cells cultured in six-well plates were transfected with different siRNAs. After transfection for 48 h, cells were collected and fixed with 70% cooling ethanol at 4°C overnight. Following staining with propidium iodide solution containing RNase A, cells were washed with PBS three times. After an incubation of 30 min at room temperature, the cell cycle distribution was analysed using a FACSCalibur flow cytometer (BD Accuri C6 Plus). CytExpert software (version 2.4.0.28; Beckman Coulter, Inc.) was employed for analysis.

 Western blotting

Total protein was extracted using RIPA lysis buffer (Beyotime Institute of Biotechnology) containing 1% PMSF and 1 mmol/l β-glycerophosphate sodium salt hydrate, which were both purchased from Selleck Chemicals. The protein concentration was calculated using a BCA protein quantitative kit (Beyotime Institute of Biotechnology). A 10% SDS-polyacrylamide gel (Beyotime Institute of Biotechnology) was used to separate the same amount of protein sample (20 µg) and protein was transferred to nitrocellulose membranes (MilliporeSigma). Membranes were blocked with 5% skim milk in Tris-buffered saline containing 1% Tween 20 for 2 h at room temperature. Membranes were incubated overnight at 4°C with the following primary antibodies: Anti-phospho-GSK3-β (1:1,000; cat. no. 5558; Cell Signaling Technology, Inc.), anti-GAPDH (1:10,000; cat. no. 10494-1-AP; ProteinTech Group, Inc.), anti-c-Myc (1:1,000; cat. no. A5011; Bimake), anti-β-catenin (1:1,000; cat. no. 8480; Cell Signaling Technology, Inc.), anti-phospho-β-catenin (1:1,000; cat. no. 4176; Cell Signaling Technology, Inc.), anti-GSK3-β (1:1,000; cat. no. 121456; Cell Signaling Technology, Inc.). Following primary incubation, membranes were incubated with HRP-labeled Goat Anti-Rabbit IgG (H+L) (1:1,000; cat. no. A0208; Beyotime Institute of Biotechnology) at room temperature for 2 h. Subsequently, enhanced chemiluminescence reagents (MilliporeSigma) were employed to assess protein expression and protein expression was normalized to that of GAPDH. The blots were quantified using Evolution (version 18.0.8.0, Viber Lourmat).

 Statistical analysis

The R software (version 3.5.1) with corresponding packages and GraphPad Prism 7 (GraphPad Software, Inc.) were used for statistical analyses. Unless specified otherwise above, P<0.05 was considered to indicate a statistically significant difference. The PERL programming language (version 5.30.2; http://www.perl.org) was employed to process data. The Kaplan-Meier method and log rank test were performed to compare the OS between the high- and low-risk groups. The results of analyses were presented as the mean ± standard deviation. The Student's t-test was utilized to analyze the differences between two groups according to the specific. Dunnett's test was the post hoc test used following one-way ANOVA to compare three or more groups.

 Results <sec> <title>Construction of the hypoxia-related-lncRNA signature

RNA expression data from 471 patients with melanoma and 1 normal adjacent tissue were extracted, and two hypoxia-related gene sets (KRIEG and WINTER) were selected to identify hypoxia-related oncogenes. A total of 882 hypoxia-related mRNAs were acquired (Fig. 1A). Using Pearson's correlation analysis between the lncRNA samples from 471 patients with melanoma and hypoxia-related genes, 72 hypoxia-related lncRNAs were identified. The selection criteria were |R2|>0.7 and P<0.001. Univariate Cox regression analysis was then performed to analyse the 72 lncRNAs. The survival data revealed that 23 hypoxia-related-lncRNAs were independent prognostic factors in patients with melanoma. To eliminate the mutual influence among lncRNAs, multivariate Cox regression analysis was used (Fig. 1B). Finally, a hypoxia-related-lncRNA signature that included 7 lncRNAs was constructed, with one unfavourable (MIR205HG) and six favourable [T cell receptor β variable 11-2 (TRBV11-2), HLA-DQB1 antisense RNA 1 (HLA-DQB1-AS1), AL365361.1, AC004847.1, ubiquitin specific peptidase 30 (USP30) antisense RNA 1 (USP30-AS1) and AC022706.1] lncRNAs as prognostic factors for melanoma (Fig. 1C; Table II).

Using the present 7-hypoxia-related-lncRNA signature, patients with melanoma were classified into two groups based on their risk scores (Fig. 1D). The OS time and survival rate were longer and higher, respectively, in the low-risk group compared with the high-risk group (survival rate, 62.33% in the low-risk group vs. 44.84% in the high-risk group; Fig. 1E), which was also demonstrated in survival analysis (P<0.001; Fig. 1F). The diagnostic efficacy of the signature was estimated using time-dependent ROC curves, wherein the area under the curve (AUC) was >0.7 for both 1 and 5 years (Fig. 1G). To validate the reliability and stability of the present risk model, the training cohort was divided into two sub-validation cohorts (n=314 in sub-validation cohort 1 and n=132 in sub-validation cohort 2) at random. Kaplan-Meier survival analysis revealed a worse prognosis in high-risk group patients in both cohorts (P<0.001 and P<0.001; Fig. 1H and J). Time-dependent ROC curves showed promising accuracy of the present risk model, particularly at 5 years in both sub-validation cohort 1 (AUC=0.706) and sub-validation cohort 2 (AUC=0.732) (Fig. 1I and K). These results suggested that the present signature could predict the survival period of patients with melanoma.

 Association of the hypoxia-related-lncRNA signature with the clinicopathological features of melanoma

The association between the risk scores and clinicopathological features was assessed. The results suggested that the 7-hypoxia-related-lncRNA signature was closely related to melanoma progression. Specifically, patients with American Joint Committee on Cancer (AJCC) stage III–IV, T stage 3–4 and N stage ≥1 had higher risk scores than patients with AJCC stage I–II (P=0.0246), T stage 0–2 (P<0.001) and N stage <1 (P=0.0104) (Fig. 2C, E and F). However, differences in age, sex and M stage were not observed (P>0.05; Fig. 2A, B and D). The results indicated that the present 7-hypoxia-related-lncRNA signature was closely related to the progression of melanoma.

 Independent risk characteristics for the hypoxia-related-lncRNA signature

The prognostic ability of the 7-hypoxia-related-lncRNA signature in patients with melanoma was assessed via univariate and multivariate Cox regression analyses. The results demonstrated that the risk score of the present signature [HR, 1.694; 95% confidence interval (95% CI), 1.396-2.056; P<0.001], age (HR, 1.015; 95% CI, 1.004-1.026; P=0.008), T stage (HR, 1.283; 95% CI, 1.081-1.522; P=0.004) and N stage (HR, 1.587; 95% CI, 1.245-2.024; P<0.001) were significantly related to the OS of patients with melanoma as independent prognostic indicators (Fig. 3A and B). The time-dependent ROC curve also revealed a 0.789 AUC value of the risk score, indicating appropriate sensitivity and specificity of the present 7-hypoxia-related-lncRNA signature in predicting the survival of patients with melanoma compared with other clinicopathological parameters (Fig. 3C).

 Stratification analysis of clinical features using the hypoxia-related-lncRNA signature

A stratified analysis of patients with melanoma based on clinical features, including sex, age, AJCC stage and TNM stages, was performed. The Kaplan-Meier survival curve analysis illustrated that patients with melanoma had higher hypoxia-related-lncRNA signature-predicted risk scores and a shorter OS period compared with the low-risk groups for most clinical and clinicopathological features (Fig. S1A-K); however, the patients at the M1 stage (P=0.959) were an exception, probably due to the small sample size (Fig. S1L). These results indicated that the present 7-hypoxia-related-lncRNA signature could accurately predict the survival period of patients with melanoma.

 PCA using the hypoxia-related-lncRNA signatures

PCA was employed to detect the different distribution patterns between the low- and high-risk groups of patients with melanoma. The 7-hypoxia-related-lncRNA gene set (Fig. 4A) was used to create a low- and high-risk group. A significant separation of the risk score based on the hypoxia-related lncRNA gene set (Fig. 4B), hypoxia-related gene set (Fig. 4C) and all gene sets (Fig. 4D) was not observed, indicating that, compared with other distribution patterns, the present 7-hypoxia-related-lncRNA signature could accurately divide patients with melanoma into risk groups.

 Establishment and validation of the nomogram

To diagnose or predict the onset or progression of melanoma, a nomogram was established based on the risk score and other clinical features, including age, AJCC stage, T stage and N stage (Fig. 5A). The calibration plots satisfactorily predicted the 3-year OS rate compared with the ideal model (Fig. 5B) but unsatisfactorily predicted the 5-year OS rate (Fig. 5C). Time-dependent ROC curves revealed that the AUC values of the nomogram at 3 and 5 years were 0.709 and 0.715, respectively (Fig. 5D).

 Construction of the ceRNA network and GSEA

To explore the mechanism of the present 7-hypoxia-related-lncRNA signature, the 7 hypoxia-related lncRNAs and their corresponding miRNAs were first predicted using miRcode. The results indicated that 92 miRNAs could interact with lncRNA MIR205HG, lncRNA TRBV11-2 and lncRNA HLA-DQB1-AS1 (Table SIII). Additionally, the corresponding miRNA target proteins were predicted using three databases (TargetScan, miRDB and mieTarBase). Via comparison with the obtained hypoxia-related mRNAs, it was determined that the present signature targeted 17 miRNAs and 54 mRNAs, and a co-expression RNA network was constructed (Fig. 6A). To explore the relationship between the expression of the 7-hypoxia-related-lncRNA signature and melanoma classification, including biological processes involved in progression, GSEA was employed to identify key Gene Ontology terms and pathway enrichment analysis of high-risk score patients with melanoma in the GO-BP gene sets. Genes associated with a high-risk score were mainly enriched in the activation of immune response [normalized enrichment score (NES)=1.75; P=0.012], regulation of immune system process (NES=1.63; P=0.002), adaptive immune response (NES=2.05; P<0.001), development of immune system (NES=1.82; P<0.001), positive regulation of immune response (NES=1.94; P<0.001), positive regulation of cell death (NES=1.77; P=0.002), negative regulation of cell cycle (NES=1.57; P=0.028), positive regulation of Wnt signalling pathway (NES=1.85; P<0.001) and regulation of Wnt signalling pathway (NES=1.75; P=0.004) (Fig. 6B-K).

 Regulation of TIICs by the 7-hypoxia-related lncRNAs

Based on the GSEA results in Fig. 6C-G, the present study explored the association between immune cell infiltration and the risk score. The CIBERSORT algorithm was used to validate the association of the hypoxia-related lncRNAs with TIICs in patients with melanoma. The proportion of TIICs in each patient with melanoma was investigated using the CIBERSORT algorithm (Fig. 7A). A comparison of the TIIC levels between the high- and low-risk groups showed an increased level of M0 macrophages (P<0.001) and resting mast cells (P=0.046) in the high-risk group, and a decreased number of plasma cells (P=0.002), CD8+ T cells (P<0.001), resting memory CD4+ T cells (P=0.036), activated memory CD4+ T cells (P<0.001), follicular helper T cells (P<0.001), resting natural killer (NK) cells (P=0.007) and M1 macrophages (P<0.001) in the high-risk group (Fig. 7B). These results revealed the important role of the hypoxia-related lncRNAs in regulating TIICs.

 MIR205HG silencing inhibits melanoma cell proliferation via the canonical Wnt/β-catenin signalling pathway

The expression levels of human leukocyte antigen (HLA), MIR205HG-1 and TRBV11 in melanoma and paired adjacent tissues of 20 clinical samples were detected using RT-qPCR. The expression levels of HLA and TRBV11 were elevated, whereas the expression levels of MIR205HG-1 were decreased in the normal tissues (Fig. 8A-C). These results suggested the specificity of siRNA to MIR205HG. To investigate the role of MIR205HG in melanoma cells, A375 cells were transfected with two sequences of siRNAs that target MIR205HG and negative control (NC) siRNA. As shown in Fig. 8D, the knockout efficiency was detected using RT-qPCR. Then si-MIR205HG-1 was selected for further experimentation since it significantly inhibited the expression of MIR205HG in A375 cells compared with the NC group. The results of cell cycle distribution analysis using flow cytometry indicated that the percentage of cells in the S-phase was greater in the siMIR205HG-1 group than in the NC group in A375 cells (Fig. 8E). Additionally, detection of cell viability using a CCK-8 assay also demonstrated the proliferation of A375 cells, which was inhibited to a greater extent compared with the NC group cells after si-MIR205HG-1 transfection (Fig. 8F). Based on the aforementioned GSEA results, the potential relationship between the MIR205HG and the Wnt/signaling pathway was investigated. Subsequently, the expression levels of relative proteins were detected by western blotting (Fig. 8G). The results revealed that the expression levels of p-β-catenin and p-GSK3-β in A375 cells were increased after siMIR205HG-1 transfection, while the expression of c-Myc and β-catenin was decreased. The ratio of phosphorylated protein to total protein (p/t, phosphorylated/total) was calculated. The results demonstrated that phosphorylation modification on GSK3-β was enhanced (p/t, NC. vs. si-MIR205HG-1, 0.8782 vs. 1.219, P=0.014) and the same trend was also observed in β-catenin (p/t, NC. vs. si-MIR205HG-1, 0.4397 vs. 1.262, P=0.002). Therefore, these results indicated that silencing MIR205HG could partly inhibit the proliferation abilities of melanoma cells through the Wnt/β-catenin signalling pathway.

 Discussion

The incidence of melanoma has increased rapidly over the last 50 years worldwide (20). Surgical excision and molecular targeted drugs, which improve the OS and progression-free survival of patients, are the main treatment options for cutaneous melanoma (21). Insufficient understanding of melanoma molecular markers poses a hindrance to melanoma diagnosis and treatment. Therefore, identifying sensitive and specific biomarkers and understanding the mechanisms underlying melanoma development are required to improve the survival rate of patients with melanoma.

Functioning differently in comparison with protein coding sequences, lncRNAs serve an important role in regulating gene expression, which affects chromatin modification, as well as RNA splicing and protein activity. Emerging evidence suggests that lncRNAs could have an effect on the occurrence, development, prognosis and chemotherapy resistance of tumour cells. Recent studies have reported that lncRNAs, as novel biomarkers, could be used to predict cancer prognosis (22,23). lncRNAs have been reported to be involved in melanoma development by regulating the hypoxia pathway (23,24). Due to consistent overexpression of HOX transcript antisense RNA (HOTAIR), the incidence of lymph node metastasis has been increased compared with primary lesions. However, HOTAIR knockdown suppresses the motility and invasion of melanoma cells in vitro (25). Aberrant metastasis associated lung adenocarcinoma transcript 1 (MALAT1) upregulation promotes melanoma metastasis by stimulating cell migration and de-suppressing miR-140-mediated snail family transcriptional repressor 2 and ADAM metallopeptidase domain 10 inhibition, suggesting that MALAT1 serves as an oncogenic lncRNA in the development of melanoma (26). BRAF-activated non-protein coding RNA knockout markedly reduces the proliferation of melanoma cells via the MAPK signalling pathway (27), which may act as a prognostic predictor and possible drug target. FGD5 antisense RNA 1 could promote the prognosis in patients with melanoma (28). Furthermore, several lncRNAs, such as antisense RNA in the INK4 locus, SPRY-IT1, Llme23, urothelial cancer associated 1, SRA-like non-coding RNA and survival associated mitochondrial melanoma specific oncogenic non-coding RNA, have been demonstrated to be differentially expressed in patients with melanoma, and to serve a role as potent melanoma progression and metastasis regulators (29). Although recent studies have illustrated the mechanism of melanoma development, the involvement of regulatory lncRNAs in melanoma prognosis under hypoxic conditions remains unclear.

LncRNAs can modify the regulatory effect of miRNAs on mRNAs (30). Additionally, miRNAs can regulate translation and stability of mRNAs by controlling cellular processes, such as differentiation, apoptosis and migration. In cancer, miRNAs serve an important role in controlling metastasis, invasion and proliferation (31). Zhang et al (32) revealed that miR431 can silence DAB2 interacting protein, which is a Ras GTPase activating protein tumor suppressor, to activate Ras/Erk and promote metastasis of pancreatic neuroendocrine tumors. Feng et al (33) revealed that the upregulation of miR-409-3p and Kruppel like factor 17 could inhibit the invasion and migration of gastric cancer cells. In the present study, three lncRNAs from the 7-hypoxia-related-lncRNA signature were further investigated, and these may serve essential roles in the prognosis of patients with melanoma by interacting with the 17 miRNAs and 54 mRNAs predicted. But the potential role of these crucial miRNAs that effect on the prognosis of melanoma remains unknown. Previous studies proved that miRNAs could regulate the biological process of cancer cells through Ras GTPase activity, pErk or EMT process (34,35). In order to render the present study more complete, close attention shall be addressed on how our 7 hypoxia-related-lncRNA affects Ras GTPase activity, as well as the relationship between our signature and pErk and EMT markers.

The present study was based on data downloaded from TCGA database, which contains complete clinical and survival data of patients with melanoma. In the present study, a 7-hypoxia-related-lncRNA signature was constructed, and this included one unfavourable and six favourable lncRNAs as prognostic factors for melanoma. Certain of the lncRNAs have been previously reported to influence tumour prognosis (36). However, there remains a lack of focus on hypoxia-related lncRNAs in melanoma. It has been reported that abnormal HLA expression may influence cytotoxic T lymphocyte and NK cell responses in uveal melanoma (37). According to a study by Serana et al (38), TRBV contributes to the promotion of the antitumour T-cell responses in patients with melanoma. Due to its association with multiple tumour-related pathways, MIR205HG-1 was also investigated, particularly in epidermis development and the immune response (39,40). The USP30-mediated stabilisation of dynamin-related protein 1 has been reported to serve a critical role in the development of hepatocellular carcinoma (41). As Lv et al (42) reported, a prognostic model based on the 7 lncRNAs, including AL365361, could effectively predict early recurrence after surgical resection of hepatocellular carcinoma. AC022706.1 has also been investigated as a useful biomarker for promoting antitumour immunity and the immunotherapy response in gastric cancers (43). In the present study, a signature of prognosis-related lncRNAs in patients with melanoma was successfully constructed, which was demonstrated to be associated with melanoma progression. The internal validation also demonstrated the accuracy of the present signature. Univariate and multivariate Cox regression analyses demonstrated that the present signature could act as an independent predictor for prognosis prediction of patients with melanoma. Time-dependent ROC curve analysis suggested that the present risk model had the best sensitivity and specificity compared with three other clinicopathological parameters. A novel nomogram based on several crucial clinical features and the present risk score was successfully constructed and validated. These results suggested that the present risk model could be an independent predictor of the survival period in patients with melanoma.

Based on the prediction of different databases and hypoxia-related mRNAs, a co-expression RNA network, including 17 miRNAs and 54 mRNAs, was constructed. Using GSEA, it was revealed that the gene sets enriched in the high-risk group included the activation of immune response, regulation of immune system process, adaptive immune response, development of immune system, negative regulation of cell cycle, positive regulation of cell death, positive regulation of immune response, positive regulation of the Wnt signalling pathway and regulation of the Wnt signalling pathway. Immunotherapy is an important mean of systemic treatment of melanoma, which could improve the prognosis of melanoma, and thus, the present study investigated the relationship between the present risk model and TIICs. The results demonstrated that M0 macrophages, resting mast cells, plasma cells, CD8+ T cells, resting memory CD4+ T cells, activated memory CD4+ T cells, follicular helper T cells, NK cells and M1 macrophages exhibited markedly different infiltration patterns in melanoma. The present results suggested that the 7-hypoxia-related-lncRNA signature could partly reflect the immune infiltration level of melanoma and provide valuable information for immunotherapy.

To validate the prediction of the present signature, the expression levels of HLA, MIR205HG and TRBV11 were detected in 20 clinical samples. The results demonstrated high expression levels of MIR205HG and low expression levels of HLA and TRBV11 in melanoma tissues. The increased MIR205HG expression in melanoma tissues attracted our attention. To investigate the relationship between poor prognosis and aberrant MIR205HG expression, siRNA was used to silence its expression, resulting in suppressed proliferation of melanoma A375 cells. These results suggested that MIR205HG is a target that can promote melanoma proliferation. The dysregulation of cell cycle control is an obstacle to the suppression of melanoma growth (44). Agents targeting the G1/S and G2/M checkpoints have shown promising preclinical activity in patients with melanoma (45,46). Because si-MIR205HG-1 suppressed proliferation in A375 cells, cell cycle arrest was investigated using flow cytometry in si-MIR205HG-1 cells. Cell aggregation that decreased at the S phase but increased at the G2/M phase was observed after treatment with si-MIR205HG-1. These results demonstrated that silencing MIR205HG may inhibit melanoma cell proliferation at the G2/M phase, which could partly explain the relationship between the present risk model based on the expression of 7 hypoxia-related lncRNAs and poor prognosis of patients with melanoma.

The aberrant activation of canonical Wnt/β-catenin signalling is observed in multiple malignant tumours and is recognised as an attractive therapeutic target for molecular drugs (47). It has been reported that Wnt serves a crucial role in tumorigenesis and malignancy, which contributes to tumour proliferation and differentiation, survival, stress response and resistance (48). To elucidate the molecular mechanisms of the effect of si-MIR205HG-1 on cell proliferation, regulatory proteins of the Wnt/β-catenin signalling pathway were examined in A375 cells using western blot analysis. Increased p-GSK3-β levels in A375 cells decreased β-catenin and c-Myc expression, resulting in low si-MIR205HG-1 expression. These results demonstrated that the silencing of MIR205HG-1 inhibited the proliferation of melanoma via the canonical Wnt/β-catenin signalling pathway. LncRNA can influence tumor cell migration and invasion. Chen et al (16) revealed that NORAD may play critical roles in tumorigenesis and progression of malignant melanoma by regulating of the MIR205-EGLN2 pathway. Xu et al (49) revealed that lncRNA CRNDE can promote the migration and invasion of melanoma by sponging MIR205 and releasing CCL18. And in the present study, it was primarily identified that the overexpression of MIR205HG could inhibit the proliferation of melanoma cells through the canonical Wnt/β-catenin signalling pathway. In order to render the present study more thorough, it will be investigated in future studies how the dose of our 7 hypoxia-related-lncRNA influences melanoma by cell migration and invasion.

The present study has several limitations. The present prognostic model should be further validated using a larger clinical sample size with adequate follow-up duration. In addition, the underlying mechanisms of how these 7 lncRNAs or some of them influence the proliferation of melanoma should be more comprehensively investigated. The detailed mechanisms of how MIR205HG inhibits the proliferation of melanoma cells via the canonical Wnt/β-catenin signalling pathway require more exploration and validation. Furthermore, the present study lacks animal experiments, which will be verified in further exploration.

At present, there were kinds of lncRNA signatures used to predict the prognosis of melanoma. Although hypoxia is an important circumstance for melanoma cells surviving, the study focused on hypoxia-related-lncRNA signature has not been reported yet. Moreover, MIR205HG played an important role in influencing the pluripotency, proliferation (50), and dissemination of malignant melanoma (51). Nevertheless, few of them paid attention to investigate and verify that MIR205HG could be a prognostic biomarker and an actionable target for inhibiting proliferation of melanoma.

In conclusion, a novel 7-hypoxia-related-lncRNA signature was preliminarily constructed and validated to promote the prognostic prediction of patients with melanoma. Based on this risk model, a nomogram and ceRNA network were successfully constructed to provide a new sight of investigating the underlying molecular mechanism of melanoma. Among these lncRNAs, supressing the overexpression of MIR205HG inhibited the proliferation of melanoma cells via the canonical Wnt/β-catenin signalling pathway, which could partly explain the relationship between the present risk model and poor prognosis of patients with melanoma.

 Supplementary Material Supporting Data Supporting Data Supporting Data Supporting Data Acknowledgements

Not applicable.

 Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

 Authors' contributions

YL and AJC designed and implemented experiments. THL and HGZ collected data and performed statistical analysis. YL and THL helped perform the analysis with constructive discussions and revised the manuscript carefully. All authors read and approved the final manuscript. YL and AJC confirm the authenticity of all the raw data.

 Ethics approval and consent to participate

This article meets the requirements of TCGA and other websites for publication and does not contain any studies involving animals performed by any of the authors. The study was approved (approval no. 2021-487) by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University. All individuals provided written informed consent for the use of their samples for clinical research.

 Patient consent for publication

Not applicable.

 Competing interests

The authors declare that they have no competing interests.

 References LeonardiGCFalzoneLSalemiRZanghìASpandidosDAMccubreyJACandidoSLibraMCutaneous melanoma: From pathogenesis to therapy (Review)Int J Oncol5210711080201829532857 DaveyMGMillerNMcInerneyNMA review of epidemiology and cancer biology of malignant melanomaCureus13e15087202134155457 CarrSSmithCWernbergJEpidemiology and risk factors of melanomaSurg Clin North Am100112202010.1016/j.suc.2019.09.00531753105 LuXKangYHypoxia and hypoxia-inducible factors: Master regulators of metastasisClin Cancer Res1659285935201010.1158/1078-0432.CCR-10-136020962028 LiuYCiottiGEEisinger-MathasonTSKHypoxia and the tumor secretomeAdv Exp Med Biol11365769201910.1007/978-3-030-12734-3_431201716 GrahamKUngerEOvercoming tumor hypoxia as a barrier to radiotherapy, chemotherapy and immunotherapy in cancer treatmentInt J Nanomedicine1360496058201810.2147/IJN.S14046230323592 JingXYangFShaoCWeiKXieMShenHShuYRole of hypoxia in cancer therapy by regulating the tumor microenvironmentMol Cancer18157201910.1186/s12943-019-1089-931711497 QiuHChenFChenMMicroRNA-138 negatively regulates the hypoxia-inducible factor 1α to suppress melanoma growth and metastasisBiol Open8bio042937201910.1242/bio.04293731371307 DongLYouSZhangQOsukaSDeviNSKaluzSFergusonJHYangHChenGWangBArylsulfonamide 64B inhibits hypoxia/HIF-induced expression of c-Met and CXCR4 and reduces primary tumor growth and metastasis of uveal melanomaClin Cancer Res2522062218201910.1158/1078-0432.CCR-18-136830563937 QuinnJJChangHYUnique features of long non-coding RNA biogenesis and functionNat Rev Genet174762201610.1038/nrg.2015.1026666209 JarrouxJMorillonAPinskayaMHistory, discovery, and classification of lncRNAsAdv Exp Med Biol1008146201710.1007/978-981-10-5203-3_128815535 LiuMZhongJZengZHuangKYeZDengSChenHXuFLiQZhaoGHypoxia-induced feedback of HIF-1α and lncRNA-CF129 contributes to pancreatic cancer progression through stabilization of p53 proteinTheranostics947954810201910.7150/thno.3098831367258 PiaoHYLiuYKangYWangYMengXYYangDZhangJHypoxia associated lncRNA HYPAL promotes proliferation of gastric cancer as ceRNA by sponging miR-431-5p to upregulate CDK14Gastric Cancer254463202210.1007/s10120-021-01213-534247316 YuZWangYDengJLiuDZhangLShaoHWangZZhuWZhaoCKeQLong non-coding RNA COL4A2-AS1 facilitates cell proliferation and glycolysis of colorectal cancer cells via miR-20b-5p/hypoxia inducible factor 1 alpha subunit axisBioengineered1262516263202110.1080/21655979.2021.196983334477476 LiuYHeDXiaoMZhuYZhouJCaoKLong noncoding RNA LINC00518 induces radioresistance by regulating glycolysis through an miR-33a-3p/HIF-1α negative feedback loop in melanomaCell Death Dis12245202110.1038/s41419-021-03523-z33664256 ChenYCaoKLiJWangASunLTangJXiongWZhouXChenXZhouJLiuYOverexpression of long non-coding RNA NORAD promotes invasion and migration in malignant melanoma via regulating the MIR-205-EGLN2 pathwayCancer Med817441754201910.1002/cam4.204630843652 WuLLiuGHeYWChenRWuZYIdentification of a pyroptosis-associated long non-coding RNA signature for predicting the immune status and prognosis in skin cutaneous melanomaEur Rev Med Pharmacol Sci2555975609202134604952 ChenHPanYJinXChenGIdentification of a four hypoxia-associated long non-coding RNA signature and establishment of a nomogram predicting prognosis of clear cell renal cell carcinomaFront Oncol11713346202110.3389/fonc.2021.71334634386428 LivakKJSchmittgenTDAnalysis of relative gene expression data using real-time quantitative PCR and the 2(−Delta Delta C(T)) methodMethods25402408200110.1006/meth.2001.126211846609 RaimondiSSuppaMGandiniSMelanoma epidemiology and sun exposureActa Derm Venereol100adv00136202010.2340/00015555-349132346751 PavriSNCluneJAriyanSNarayanDNarayan, malignant melanoma: Beyond the basicsPlast Reconstr Surg138330e340e201610.1097/PRS.000000000000236727465194 ZhangHQinCLiuHWGuoXGanHAn effective hypoxia-related long non-coding RNAs assessment model for prognosis of clear cell renal carcinomaFront Oncol11616722202110.3389/fonc.2021.61672233692953 GongPJShaoYCHuangSRZengYFYuanXNXuJJYinWNWeiLZhangJWHypoxia-associated prognostic markers and competing endogenous RNA Co-expression networks in breast cancerFront Oncol10579868202010.3389/fonc.2020.57986833344235 ChiYWangDWangJYuWYangJLong non-coding RNA in the pathogenesis of cancersCells81015201910.3390/cells809101531480503 ZhangJLiuHZhangWLiYFanZJiangHLuoJIdentification of lncRNA-mRNA regulatory module to explore the pathogenesis and prognosis of melanomaFront Cell Dev Biol8615671202010.3389/fcell.2020.61567133392203 SunLSunPZhouQYGaoXHanQLong noncoding RNA MALAT1 promotes uveal melanoma cell growth and invasion by silencing of miR-140Am J Transl Res839393946201627725873 YuXZhengHChanMTWuWKKBANCR: A cancer-related long non-coding RNAAm J Cancer Res717791787201728979803 GaoYZhuHMaoQExpression of lncRNA FGD5-AS1 correlates with poor prognosis in melanoma patientsJ Gene Med22e3278202010.1002/jgm.327832997827 YuXZhengHTseGChanMTWuWKLong non-coding RNAs in melanomaCell Prolif51e12457201810.1111/cpr.1245729582492 KaragkouniDKaravangeliAParaskevopoulouMDHatzigeorgiouAGCharacterizing miRNA-lncRNA interplayMethods Mol Biol2372243262202110.1007/978-1-0716-1697-0_2134417757 LaiXEberhardtMSchmitzUVeraJSystems biology-based investigation of cooperating microRNAs as monotherapy or adjuvant therapy in cancerNucleic Acids Res4777537766201910.1093/nar/gkz63831340025 ZhangTChoiSZhangTChenZChiYHuangSXiangJZDuYNmiR-431 promotes metastasis of pancreatic neuroendocrine tumors by targeting DAB2 interacting protein, a ras GTPase activating protein tumor suppressorAm J Pathol190689701202010.1016/j.ajpath.2019.11.00731953039 FengJLiKLiuGFengYShiHZhangXPrecision hyperthermia-induced miRNA-409-3p upregulation inhibits migration, invasion, and EMT of gastric cancer cells by targeting KLF17Biochem Biophys Res Commun549113119202110.1016/j.bbrc.2021.02.06333667708 SalminenAKaarnirantaKKauppinenAOjalaJHaapasaloASoininenHHiltunenMImpaired autophagy and APP processing in Alzheimer's disease: The potential role of Beclin 1 interactomeProg Neurobiol106-1073354201310.1016/j.pneurobio.2013.06.00223827971 SuJMorganiSMDavidCJWangQErEEHuangYHBasnetHZouYShuWSoniRKTGF-β orchestrates fibrogenic and developmental EMTs via the RAS effector RREB1Nature577566571202010.1038/s41586-019-1897-531915377 ZhangBTangBGaoJLiJKongLQinLA hypoxia-related signature for clinically predicting diagnosis, prognosis and immune microenvironment of hepatocellular carcinoma patientsJ Transl Med18342202010.1186/s12967-020-02492-932887635 SouriZWierengaAPAMulderAJochemsenAGJagerMJHLA expression in uveal melanoma: An indicator of malignancy and a modifiable immunological targetCancers (Basel)111132201910.3390/cancers1108113231394860 SeranaFSottiniACaimiLPalermoBNataliPGNisticòPImbertiLIdentification of a public CDR3 motif and a biased utilization of T-cell receptor V beta and J beta chains in HLA-A2/Melan-A-specific T-cell clonotypes of melanoma patientsJ Transl Med721200910.1186/1479-5876-7-2119317896 GuoJGanQGanCZhangXMaXDongMLncRNA MIR205HG regulates melanomagenesis via the miR-299-3p/VEGFA axisAging (Albany NY)1352975311202110.18632/aging.20245033535182 LiuNLiuZLiuXChenHComprehensive analysis of a competing endogenous RNA network identifies seven-lncRNA signature as a prognostic biomarker for melanomaFront Oncol9935201910.3389/fonc.2019.0093531649871 GuLZhuYLinXLiYCuiKProchownikEVLiYAmplification of glyceronephosphate O-acyltransferase and recruitment of USP30 stabilize DRP1 to promote hepatocarcinogenesisCancer Res7858085819201810.1158/0008-5472.CAN-18-034030143522 LvYWeiWHuangZChenZFangYPanLHanXXuZLong non-coding RNA expression profile can predict early recurrence in hepatocellular carcinoma after curative resectionHepatol Res4811401148201810.1111/hepr.1322029924905 HeYWangXIdentification of molecular features correlating with tumor immunity in gastric cancer by multi-omics data analysisAnn Transl Med81050202010.21037/atm-20-92233145269 XuWMcArthurGCell cycle regulation and melanomaCurr Oncol Rep1834201610.1007/s11912-016-0524-y27106898 BarnabaNLaRocqueJRTargeting cell cycle regulation via the G2-M checkpoint for synthetic lethality in melanomaCell Cycle2010411051202110.1080/15384101.2021.192280633966611 AfrangNHonardoostMCell cycle regulatory markers in melanoma: New strategies in diagnosis and treatmentMed J Islam Repub Iran3396201931696090 NishiyaNScreening for chemical suppressors of the Wnt/β-catenin signaling pathwayYakugaku Zasshi1371331362017(In Japanese)10.1248/yakushi.16-00229-228154320 XueGRomanoEMassiDMandalàMWnt/β-catenin signaling in melanoma: Preclinical rationale and novel therapeutic insightsCancer Treat Rev49112201610.1016/j.ctrv.2016.06.00927395773 XuLZhangYZhaoZChenZWangZXuSZhangXLiuTYuSThe long non-coding RNA CRNDE competed endogenously with miR-205 to promote proliferation and metastasis of melanoma cells by targeting CCL18Cell Cycle1722962308201810.1080/15384101.2018.152660230257602 SahranavardfardPMadjdZEmami RazaviANGhanadanARFirouziJKhosravaniPGhavamiSEbrahimieEEbrahimiMAn integrative analysis of the Micro-RNAs contributing in stemness, metastasis and B-Raf pathways in malignant melanoma and melanoma stem cellCell J23261272202134308569 Sánchez-SendraBSernaENavarroLGonzález-MuñozJFPorteroJRamosAMurguiAMonteagudoCTranscriptomic identification of miR-205 target genes potentially involved in metastasis and survival of cutaneous malignant melanomaSci Rep104771202010.1038/s41598-020-61637-432179834

Construction of the hypoxia-related-lncRNA signature. (A) Overlap between two hypoxia-related gene sets (KRIEG and WINTER) and mRNA expression levels in melanoma. (B) Multivariate Cox regression analysis was used to identify hypoxia-related lncRNAs. (C) Heatmap demonstrating hypoxia-related-lncRNA signature expression in patients. (D) Patients sorted into high- and low-risk groups according to the median risk score. (E) Survival status and survival time of each patient. The SR was calculated in both the high- and low-risk group. (F) Survival curve analysis of the 7-hypoxia-related-lncRNA signature in training cohort. (G) Time-dependent ROC curve analysis in training cohort. (H) Survival curve of the 7-hypoxia-related-lncRNA signature in sub-validation cohort 1. (I) Time-dependent ROC curves analysis in sub-validation cohort 1. (J) Survival curve of the 7-hypoxia-related-lncRNA signature in sub-validation cohort 2. (K) Time-dependent ROC curves analysis in sub-validation cohort 2. lncRNA, long non-coding RNA; ROC, receiver operating characteristic; SR, survival rate.

Association of the 7-hypoxia-related-long non-coding RNA signature with clinicopathological features in melanoma. Association between the signature risk scores and the clinicopathological features, including (A) age (>65 vs. ≤65; unpaired Student's t-test, P=0.9469), (B) sex (female vs. male; unpaired Student's t-test, P=0.2866), (C) American Joint Committee on Cancer stage (stage I–II vs. stage III–IV; unpaired Student's t-test, P=0.0246), (D) M stage (stage 0 vs. stage 1; unpaired Student's t-test, P=0.6898), (E) N stage (stage <1 vs. stage ≥1; unpaired Student's t-test, P=0.0104) and (F) T stage (stage 0–2 vs. stage 3–4; unpaired Student's t-test, P<0.001).

Hypoxia-related-lncRNA signature is an independent factor of prognostic indicators of melanoma. (A and B) Univariate and multivariate Cox regression analysis demonstrated the association between overall survival and clinicopathological parameters and the risk score. (C) Receiver operating characteristic curve analysis indicated the prognostic accuracy of clinicopathological parameters and the hypoxia-related lncRNA prognostic risk score. lncRNA, long non-coding RNA.

PCA. PCA between the high-risk group and the low-risk group based on (A) the hypoxia-related-lncRNA signature, (B) hypoxia-related lncRNAs, (C) hypoxia-related genes and (D) all gene sets. The red and green dots represent high- and low-risk genes, respectively. lncRNA, long non-coding RNA; PCA, principal component analysis.

Construction and validation of the prognostic nomogram. (A) Construction of the nomogram was based on traditional clinical variables and the risk score. (B and C) Calibration plot for the internal validation of the nomogram at 3 and 5 years. (D) Time-dependent receiver operating characteristic curves indicating the area under the curve of the nomogram.

Co-expression RNA regulation network construction and functional enrichment analysis. (A) Construction of a hypoxia-related competing endogenous RNA regulation network based on predicted microRNAs and hypoxia-related mRNAs. (B-K) Gene Ontology-Biological Process gene sets enriched in the high-risk group, including activation of immune response (NES=1.75; P=0.012), regulation of immune system process (NES=1.63; P=0.002), adaptive immune response (NES=2.05; P<0.001), development of immune system (NES=1.82; P<0.001), positive regulation of immune response (NES=1.94; P<0.001), positive regulation of cell death (NES=1.77; P=0.002), negative regulation of cell cycle (NES=1.57; P=0.028), positive regulation of Wnt signalling pathway (NES=1.85; P<0.001) and regulation of Wnt signalling pathway (NES=1.75; P=0.004). NES, normalized enrichment score.

Melanoma is associated with a variety of TIICs. (A) Related percentage of 22 types of TIICs calculated by the CIBERSORT algorithm shown in a bar plot. (B) Differences in infiltration levels of TIICs between the high- and low-risk groups based on the expression of 7 hypoxia-related long non-coding RNAs (unpaired Student's t-test). *P<0.05, **P<0.01 and ***P<0.001. TIICs, tumour-infiltrating immune cells.

Silencing of MIR205HG inhibits proliferation of melanoma via the canonical Wnt/β-catenin signaling pathway. (A-C) RT-qPCR to detect the expression levels of HLA-DQB1-AS1, MIR205HG-1 and TRBV11-2 in melanoma or paired tissues from 20 patients (Paired Student's t-test; *P<0.05 and **P<0.01). (D) RT-qPCR to detect the relative silencing levels of lncMIR205HG (Data are presented as the mean ± SD; n=3; one-way ANOVA and Dunnett's test; *P<0.05 and **P<0.01). (E) Flow cytometric analysis of the cell cycle distribution in A375 cells after transfection with NC or MIR205HG siRNA. The percentages of cells in G0/G1, S and G2/M phase are shown in the bar graph (n=3; unpaired Student's t-test; *P<0.05 and **P<0.01). (F) Proliferation of A375 cells transfected with NC or si-MIR205HG-1 was assessed using a Cell Counting Kit-8 assay (n=3; one-way ANOVA and Dunnett's test; **P<0.01). (G) β-catenin, p-β-catenin, c-Myc, GSK3-β and p-GSK3-β levels in A375 cells transfected with siRNAs determined by western blotting, ratio of phosphorylated protein to total protein (p/t, phosphorylated/total) about GSK3-β and β-catenin were calculated (n=3; unpaired Student's t-test; *P<0.05 and **P<0.01). MIR205HG, MIR205 host gene; NC, negative control; p-, phosphorylated; RT-qPCR, reverse transcription-quantitative PCR; siRNA/si, small interfering RNA; ns, non-significant.

Sequences of small interfering RNAs and primer sequences for reverse transcription-quantitative PCR.

Designation Gene/siRNA Sequences (5′-3′) Organism
Primer HLA-DQB1-AS1 F: AGCAGAGTCCAGGGTGTATTG Homo sapiens
R: GCTGCTAGTGGTCGGGAAG
TRBV11-2 F: GTGACCCTGATTGGGCAAAG
R: TATCTGGGAGACTGGGCAAC
MIR205HG F: AGGAGTCATTTCTGTTCCGCA
R: CAAATAGTGTCCCAGCCACCT
GAPDH F: CAGTGGCAAAGTGGAGATTGTTG
R: TCGCTCCTGGAAGATGGTGAT
siRNA MIR205HG-siRNA-1 F: UCUCCUUCAAUUCCACUUUTT Homo sapiens
R: AAAGUGGAAUUGAAGGAGATT
MIR205HG-siRNA-2 F: GAGACAGCCAGAGAGAAAUTT
R: AUUUCUCUCUGGCUGUCUCTT

F, forward; HLA-DQB1-AS1, HLA-DQB1 antisense RNA 1; MIR205HG, MIR205 host gene; R, reverse; siRNA, small interfering RNA; TRBV11-2, T cell receptor β variable 11-2.

Hypoxia-related long non-coding RNAs associated with the overall survival of patients with melanoma.

Gene symbol Aliases Ensemble ID Location P-value HR
TRBV11-2 Lnc-MGAM2-1 ENSG00000241657 chr7:142,433,895-142,434,394 0.000361432 1.330058039
HLA-DQB1-AS1 HSALNG0049432 ENSG00000223534 chr6:32,659,879-32,660,729 0.012388052 0.726654702
AL365361.1 lnc-KCNA3-3 ENSG00000259834 chr1:110653560-110657040 0.018247458 0.675041646
MIR205HG NPC-A-5 ENSG00000230937 chr1:209,425,755-209,433,438 0.003259287 1.270371188
AC004847.1 Lnc-MYO1G-1 ENSG00000260997 chr7:44,958,998-44,960,909 0.032147891 1.76373733
USP30-AS1 HSALNG0093872 ENSG00000256262 chr12:109,051,790-109,054,033 0.0000617 0.481152446
AC022706.1 Lnc-SLFN5-1 ENSG00000267364 chr17:33,640,521-33,651,919 0.000622238 1.508349735

HLA-DQB1-AS1, HLA-DQB1 antisense RNA 1; HR, hazard ratio; MIR205HG, MIR205 host gene; TRBV11-2, T cell receptor β variable 11-2; USP30-AS1, USP30 antisense RNA 1.