Upregulation of the novel lncRNA U731166 is associated with migration, invasion and vemurafenib resistance in melanoma

Abstract Our previous work using a melanoma progression model composed of melanocytic cells (melanocytes, primary and metastatic melanoma samples) demonstrated various deregulated genes, including a few known lncRNAs. Further analysis was conducted to discover novel lncRNAs associated with melanoma, and candidates were prioritized for their potential association with invasiveness or other metastasis‐related processes. In this sense, we found the intergenic lncRNA U73166 (ENSG00000230454) and decided to explore its effects in melanoma. For that, we silenced the lncRNA U73166 expression using shRNAs in a melanoma cell line. Next, we experimentally investigated its functions and found that migration and invasion had significantly decreased in knockdown cells, indicating an essential association of lncRNA U73166 for cancer processes. Additionally, using naïve and vemurafenib‐resistant cell lines and data from a patient before and after resistance, we found that vemurafenib‐resistant samples had a higher expression of lncRNA U73166. Also, we retrieved data from the literature that indicates lncRNA U73166 may act as a mediator of RNA processing and cell invasion, probably inducing a more aggressive phenotype. Therefore, our results suggest a relevant role of lncRNA U73166 in metastasis development. We also pointed herein the lncRNA U73166 as a new possible biomarker or target to help overcome clinical vemurafenib resistance.

such as Australia and Northern America, mainly due to public policy campaigns and new therapies respectively. [3][4][5] However, in 2021, considering only the United States, there is still a prediction of 101,280 new cases of melanoma in situ, with an estimation of 7180 melanoma-associated deaths. 6 Cutaneous melanoma can be classified according to mutations that patients harbour in their cells, and this is known as the Genomic Classification of Cutaneous Melanoma. 7 This molecular subtype stratification of patients is very relevant because about 50% of melanoma patients have a mutation in the BRAF gene. [8][9][10] Another reason for BRAF relevance in melanoma is that from all BRAF-mutants individuals, about 90% harbour the specific V600E mutation [11][12][13] which activates the MAPK-ERK pathway constitutively, allowing cells to become self-sufficient in growth signals and leading to tumour formation. 14,15 The BRAFV600E relevance in melanoma is also reflected in drugs that have been investigated to target this mutation. The first widely used BRAF-mutant inhibitor is known as vemurafenib and it initially demonstrated promising results, reducing the risk of death and tumour progression by 63% and 74% respectively. 11 However, the main obstacles regarding vemurafenib exclusive treatment are the primary resistance that accounts for 20% of the cases and acquired drug resistance. 16,17 These mechanisms of drug resistance in melanoma are mainly due to the MAPK reactivation with or without the PI3K/AKT pathway activation. 18 It is well known that mutations in melanoma oncogenes and tumour suppressor genes may result in rupture of diverse pathways involved in cell signalling. 19 Currently, the clinicians can base their decision in patient's management according to the clinical aspects they present, specific subtype mutations or other relevant biological features to define the melanoma treatment strategies. 20,21 Considering metastatic melanomas, a combination of strategies can be utilized more effectively than a single therapeutics. 21 In the last decades, increasing interest in biomarkers and gene therapy using nucleic acids to treat melanoma has revealed its potential application to therapeutics. Small non-coding RNAs (sncRNAs) were the leading topic of research for many years. [22][23][24][25] This research field using sncRNAs was improved by developing technologies like microarray profiling, real-time PCR array and next-generation sequencing (NGS) technologies. 26 More recently, long noncoding RNAs (lncRNAs) have begun to be extensively studied and demonstrated a high potential to be used as biomarkers in many cancers and can be helpful for cancer patient management. 27 The lncRNAs are generally defined as RNA molecules that are 200 nucleotides long and have no proteincoding ability. 28 They are expressed in a tissue-specific manner, participate in a myriad of critical cellular functions and have been implicated as mediators in distinct disease pathogenesis. 29 In cancer, lncRNAs have initially been pointed out as deregulated transcripts whose expression levels impact normal processes, but recently, they have been associated with functionally relevant alterations in critical cancer processes and pathways. 30 Therefore, lncRNAs demonstrated an enormous potential to be explored and used as a biomarker in cancer research. 31,32 An increasing number of novel lncRNAs have been revealed in melanoma, demonstrating their contributions to tumour development. Notably, many of them have great potential to be used as biomarkers and even as therapeutic targets. 33 They have been implicated in diverse cancer aspects in melanoma as proliferation, 34 invasion, 35 metastasis, 36 migration, 37 apoptosis 35 and other tumoural processes. Another reason for the importance of this class of transcripts in melanomagenesis is that they can be more expressed in specific melanoma samples, making them putative biomarker molecules of specific biological aspects regarding melanoma development and staging. 38 This study explored our previous RNA-Seq results and found that the novel lncRNA U73166 was deregulated and associated with an invasive profile in melanoma. Moreover, we experimentally verified these findings, and we found that silencing lncRNA U73166 impacts migration, invasion and proliferation in melanoma cells. Our further analysis found an association between the lncRNA U73166 expression and acquired resistance to vemurafenib, suggesting that this lncRNA may play an essential role in melanoma resistance. To obtain vemurafenib-resistant cell lines, we used A375, SK-MEL-28 and WM164 cell lines seeded at a low cell density (1×10 4 cells) in a 60 mm plate, according to Sandri et al., (2016). 39 Then, these cell lines were treated with increasing doses of vemurafenib in the range 0.5-6 µM every 3 days for a maximum of 6 weeks. It is important to mention that the vemurafenib resistance was previously validated in these cell lines by western blotting and it was confirmed the MAPK pathway reactivation-MEK and ERK phosphorylation levels in vemurafenib presence. 39,40 These cells were named naïve and resistant cells. The resistant cell lines were continuously refilled with 6 µM vemurafenib every 2-3 days.

| Melanocytic cell lines
All cells were incubated at 37℃ in a humidified atmosphere with 5% CO 2 and were tested for mycoplasma infection.

| Analysis of RNA-Seq Data and Gene Set Variation Analysis for invasive and proliferative enrichment scores
In our previous work, 41 we performed Gene Set Variation Analysis (GSVA) with specific gene expression signatures to score our melanoma cell lines and melanoma tissue samples from the TCGA database for invasive and proliferative phenotypes based on their RNA expression profiles. 41 Here, we evaluated the correlation between the lncRNA U73166 expression level and each of these enrichment scores (proliferative and invasive) in the same set of samples. The expression pattern of lncRNA U73166 in several TCGA tumours and GTEx normal tissues was obtained via gene expression profiling interactive analysis (GEPIA2). 42 Also, GEPIA2 was utilized to explore gene expression correlations between lncRNA U73166 and RBFOX2, HNRNPA2B1 and SRSF1. To identify RNA-binding proteins interacting with lncRNA U73166, according to experimental evidence, we utilized CLIP-Seq data obtained from the starBase v2.0 platform. 43

| RNA extraction and reverse transcriptionquantitative PCR
Total RNA was isolated from cell lines using miRNeasy Mini Kit

| Subcellular fractionation
To obtain RNA from A375 cell line subcellular fractions, we utilized the Ambion PARIS kit (Thermo Fisher Scientific) following instructions provided by the manufacturer. Briefly, melanoma cells were trypsinized, and a total of 1x10 6 cells was pelleted in a microfuge tube. These cells were submitted to cell fractionation and centrifugation at 4°C, resulting in partitioned nuclear and cytoplasmic fractions. The supernatant was relocated to another microfuge tube, and the pellet remained in the same tube. The following steps included each subcellular lysate being submitted to column binding and washing, and each fraction of nuclear RNA and cytoplasmic RNA was collected in different tubes. The subsequent steps included DNase treatment, RNA quality control, RNA quantification and cDNA production, as mentioned above.

| Generation of shRNA constructs and lentiviral transduction
First, we used the lncRNA U73166 FASTA sequence with the best ENSEMBL support to design three shRNAs using the web-based tool for siRNA selection from Whitehead Institute (http://sirna.wi.mit.

| Transwell migration and invasion assays
In vitro cell migration was performed using Thincert Cell Culture Insert For 24 Well Plates (Greiner Bio-One). Briefly, a volume of 600 µl of cell suspension (1 × 10 4 cells) in DMEM serum-free medium was added into the upper chambers. A volume of 600 µL DMEM medium with 10% FBS without antibiotic was added in the lower chamber to induce cell migration. After 24 h incubation, the medium was removed, and migrated cells were fixed with 4% paraformaldehyde and stained using 0,5% crystal violet solution. The non-invading cells were removed from the insert's upper surface using a cotton swab, and five random fields were photographed using the inverted microscope IX71 system (Olympus). Images were later processed, quantified and analysed using ImageJ software.

In vitro cell invasion was conducted using BioCoat Matrigel
Invasion Chamber assay (Corning). The invasion chamber was removed from the freezer and rehydrated with DMEM medium at 37°C. DMEM was added to the insert's interior and the bottom of wells 2 h before plating the cells. The following steps were performed as described above for the migration assay. All the experiments were performed in triplicate.

| Wound healing assay
A375 melanoma cells were plated in 6-well plates at a density of

| Statistical and image analysis
Statistical analyses were carried out using the R platform and the PRISM software package (version V.6.01, GraphPad Software).
Values of p ≤ 0.05 were considered statistically significant. ImageJ software was used to process and perform measurements in wound healing experiments, invasion and transwell migration assays.

| The novel lncRNA U73166 is upregulated in melanoma and its expression is correlated with an invasiveness signature
Our previous work using a melanoma progression model composed of melanocytes, primary and metastatic melanoma samples indicated several s deregulated genes, including a few known lncR-NAs. 41 Further analysis was carried out to discover novel lncRNAs impacting melanoma development, and candidates were prioritized due to their potential association with invasiveness. In this sense, we identified the intergenic lncRNA U73166 (ENSG00000230454) located in the region between the protein-coding genes SEMA3B and GNAI2 ( Figure 1A), and this transcript was selected to be further investigated. We observed that lncRNA U73166 was upregulated in

| The lncRNA U73166 is expressed in a melanoma-testis pattern
We searched the expression profile of the lncRNA U73166 in a public data platform, 42 and noted that its expression is higher in testis than in any other normal or tumoural tissues ( Figure S1A and 1B). We decided to experimentally validate these results assessing lncRNA U73166 expression in several melanocytic cell lines and in a panel of human RNA samples from normal tissues. We confirmed that lncRNA U73166 is more expressed in normal testis in comparison to other normal tissues ( Figure 2C) and, that it is highly expressed in melanoma cells compared to melanocytes, although not presenting a specific trend of enrichment to a particular melanoma progression phase ( Figure 2F). In addition, according to the data available in TCGA repository, the expression level of lncRNA U73166 is not significantly different among molecular subgroups of melanoma samples ( Figure   S2A) nor between primary and metastatic melanomas ( Figure S2B).

| Downregulation of lncRNA U73166 impacts proliferation and invasion in melanoma cells
To

| lncRNA U73166 silencing impacts cell migration
Further analysis was conducted using different assays to evaluate if lncRNA U73166 was associated with migration. Using transwell migration assay, we found that the number of migrating control cells was significantly higher than for silenced cells (Figure 4). This result indicates that lncRNA U73166 may induce the migratory phenotype of melanoma cells. So, we decided to test if lncRNA U73166 reduced expression levels could also influence collective cell migration, and for that, we performed the wound healing ('scratch') assay.
Interestingly, we found that control cells presented higher migration rates than silenced cells ( Figure 5). This result indicates that collective melanoma cell migration could also be affected by reduced lncRNA U73166 gene expression. Therefore, we could infer that lncRNA U73166 gene expression can impact melanoma cells in their cell migration abilities.

| Subcellular fraction analysis demonstrates nuclear enrichment of the lncRNA U73166
It is well known that many lncRNAs accumulate into specific subcellular compartments. 44 Thus, we decided to check if the lncRNA U73166 is enriched in cytoplasmic or nuclear compartments. This analysis was first performed using public data derived from ENCODE and available in lncATLAS. 45 The  Figure 6C). It is important to mention that we utilized the same lncRNA gene markers for subcellular compartment enrichment as the public database.

| The lncRNA U73166 expression level is associated with vemurafenib resistance in BRAFV600E mutants
To check if the lncRNA U73166 could be associated with melanoma drug resistance, we utilized three melanoma cell lines that harbour the BRAFV600E mutation. These original cell lines (naïve) were treated with increasing concentrations of vemurafenib to induce them towards acquired drug resistance, and then they were labelled as resistant. Our results demonstrated that the resistant cell lines express significantly higher levels of the lncRNA U73166 than the naïve cells ( Figure 7A). This result indicates that vemurafenib resistance may be associated with increased lncRNA U73166 transcript levels. Moreover, we used public RNA expression data 46,47 from a patient treated with vemurafenib who underwent melanoma biopsy before and after acquiring vemurafenib resistance. We compared the data from this patient with five other patients treated with different anti-melanoma drugs. The results showed that in all the other five patients submitted to treatments not including vemurafenib the lncRNA U73166 expression levels did not correlate with resistance ( Figure 7B). However, for the patient who developed resistance after vemurafenib treatment lncRNA U73166 gene expression was 10 times higher in melanoma cells after than before treatment.
( Figure 7C). These results could implicate a possible correlation between melanoma resistance and higher levels of lncRNA U73166.

| lncRNA U73166 interacts with a diverse group of proteins and is a potential mediator of deregulated RNA processing in cancer
To gain additional insight regarding how lncRNA U73166 modifies BRAFV600 expression, we probed for protein with which U73166 may interact at the transcriptional level in the nucleus using the CLIP-Seq public database. 43 We found many RNA-binding proteins (RBP) that directly interact with lncRNA U73166 (Table S1).
We ranked all these RBP interacting partners, and the top 10 well-

| DISCUSS ION
Recently, many potential regulatory lncRNAs have been identified, and this field shows many transcripts with a wide range of functions. In cancer-related studies, lncRNAs have several roles associated with many aspects of carcinogenesis. 30 Due to the specific expression pattern that lncRNAs demonstrate, their importance is being mentioned in an increasing number of studies in the last decades. Indeed, some of them have already been described as specific biomarkers, 48 and other cancer-specific lncRNAs probably will be revealed in the following years. The studies in this area will enable researchers to tackle many challenges in cancer research that are difficult to address with current knowledge and approaches.
Our findings revealed, for the first time, that the lncRNA U73166 is expressed in a melanoma-testis pattern, which resembles the antigen-testis pattern, particularly interesting in the search for molecules as potential biomarkers and targetable molecules in the tumour. 49,50 Thus, the novel lncRNA U73166 can be a valuable transcript to be explored in melanoma assessment and treatment in the future.
Using knockdown experiments, we could test if our previous bioinformatics findings associating U73166 levels with an invasive expression profile would be experimentally validated. Our results demonstrated that cells presented reduced invasive potential after shRNA-mediated silencing of the lncRNA U73166. It is worthy of mentioning that the melanoma invasive state is mediated by key regulators, as AP1 and TEADs, and alterations in the expression of genes involved in invasion are also associated with an increase in patient's therapy resistance, including BRAF inhibitors. 51 Moreover, using an insert-based migration experiment, we found a higher level of cells migrating in control samples than cells in which lncRNA U73166 had been silenced. We also found that in wound healing assay, silenced cells presented decreased ability to cover the scratch area, demonstrating that lncRNA U73166 also affects collective cell migration.
As the expression of lncRNA U73166 was positively associated with an invasiveness expression profile (bioinformatics supported), and with migration, and invasion (experimentally supported), our findings suggest a role of lncRNA U73166 in melanoma invasive and migratory phenotypes.
Surprisingly, the proliferation rate-that was negatively cor-   cer. [52][53][54][55][56] It is well known that in the nuclear compartment, lncRNAs may act as a scaffold and be associated with proteins working in the gene regulation process. According to our analysis potential protein partners interacting with lncRNA U73166 included HNRNPA2B1, SRSF1 and RBFOX2, which demonstrated a robust positive expression level correlation with lncRNA U73166. While insufficient information regarding melanoma was found in the literature, we found that RBFOX2 and HNRNPA2B1 have been established to regulate tumour development-primarily through EMT-related processesand in pancreatic cancer, HNRNPA2B1 acts through the ERK/Snail pathway. 54,57,58 It is essential to mention that ERK1/2 are critical regulators of the MAPK signalling pathway and act as both downstream targets and upstream regulators (negative feedback) of the A/B/C-RAF kinases. 59 Therefore, a direct interaction between ln-cRNA U73166 and RBPs such as HNRNPA2B1 and their correlated expression may indicate a role for lncRNA U73166 in the regulation of the MAPK signalling pathway. However, confirmation of this hypothesis will require further investigation. If confirmed, the relationship between lncRNA U73166 and BRAFV600E mutants may represent an important finding as this transcript may be an essential mediator in this cancer signalling pathway. Additionally, to our results, further analysis in the future should include the silencing of lncRNA U73166 using ASO (antisense oligos) to assess if the phenotypic changes remain or are more pronounced.
Our findings showing that silencing of lncRNA U73166 impacts melanoma tumoural processes and that vemurafenib-resistant cells have significantly higher levels of lncRNA U73166 could be helpful for patient assessment and therapeutic management. In the future, these results may be beneficial to melanoma research and therapeutics, contributing, for example, to expand the field of lncRNAs biomarkers for this tumour or to indicate better approaches for drug-resistance monitoring and potential molecular targets for improvement of melanoma therapy. Rentmeister from WWU-Germany that would collaborate with this work, but due to the pandemic, the experiments had to be interrupted.

CO N FLI C T O F I NTE R E S T
All the authors declare no competing interests.