FUT8-AS1 Inhibits the Malignancy of Melanoma Through Promoting miR-145-5p Biogenesis and Suppressing NRAS/MAPK Signaling

Melanoma is the major lethal skin malignancy. However, the critical molecular drivers governing melanoma progression and prognosis are still not clear. By analyzing The Cancer Genome Atlas (TCGA) data, we identified FUT8-AS1 as a prognosis-related long non-coding RNA (lncRNA) in melanoma. We further confirmed that FUT8-AS1 is downregulated in melanoma. Reduced expression of FUT8-AS1 is correlated with aggressive clinical factors and inferior overall survival. Using in vitro functional assays, our findings demonstrated that ectopic expression of FUT8-AS1 represses melanoma cell proliferation, migration, and invasion. FUT8-AS1 silencing promotes melanoma cell proliferation, migration, and invasion. Furthermore, in vivo functional assays demonstrated that FUT8-AS1 represses melanoma growth and metastasis. Mechanistically, FUT8-AS1 was found to bind NF90, repress the interaction between NF90 and primary miR-145 (pri-miR-145), relieve the repressive roles of NF90 on mature miR-145-5p biogenesis, and thus promote miR-145-5p biogenesis and upregulate mature miR-145-5p level. The expression of FUT8-AS1 is positively correlated with miR-145-5p in melanoma tissues. Via upregulating miR-145-5p, FUT8-AS1 reduces the expression of NRAS, a target of miR-145-5. FUT8-AS1 further represses MAPK signaling via downregulating NRAS. Functional rescue assays demonstrated that inhibition of miR-145-5p reverses the tumor suppressive roles of FUT8-AS1 in melanoma. The oncogenic roles of FUT8-AS1 silencing are also blocked by MAPK signaling inhibitor MEK162. In conclusion, these findings demonstrate that FUT8-AS1 exerts tumor suppressive roles in melanoma via regulating NF90/miR-145-5p/NRAS/MAPK signaling axis. Targeting FUT8-AS1 and its downstream molecular signaling axis represent promising therapeutic strategies for melanoma.


INTRODUCTION
Despite the total cancer incidence rate is declined since 1991, the incidence rate of skin melanoma is still increasing for both male and female since 1975 (1). Therefore, skin melanoma has been the fifth most common cancer in male, accounting for 7% of all site cancers according to 2020 cancer statistics of USA (1). In female, skin melanoma is the sixth most common cancer and accounts for 4% of all site cancers according to 2020 cancer statistics of USA (1). Although recent approval of molecule targeted therapy and immunotherapy for melanoma has greatly extended the survival of melanoma patients, most late melanomas particular for those with metastases are still incurable (2,3). Thus, uncovering the critical molecular drivers responsible for the progression and poor survival of melanomas would enable more effective therapy for melanoma.
In comparison, the knowledge of lncRNAs in melanoma is relative less (31). The contributions of lncRNAs to melanoma initiation and progression are only starting to be studied (32). Several cancer-related lncRNAs in melanoma were revealed, such as SLNCR1, OVAAL, EMICERI, THOR (33)(34)(35)(36). In our previous reports, we also identified three melanoma-related lncRNAs, including PVT1, ILF3-AS1, and MHENCR (37)(38)(39). The mechanisms of action of lncRNAs are complex and various. Some lncRNAs directly bind proteins and modulates the expression and/or functions of interacted proteins, such as OVAAL, THOR, and ILF3-AS1 (34,36,38). Some lncRNAs directly bind miRNAs and repress the roles of interacted miRNAs, such as MHENCR (39). Some lncRNAs may also directly modulate neighboring protein-coding genes, such as EMICERI (35). Despite several lncRNAs have been elucidated in melanoma, the contributions of most other lncRNAs to melanoma are still unclear. As transcriptomic sequencings have identified more than 58,000 lncRNAs in human (40), we could not preclude the contributions of other lncRNAs to melanoma.
To further identify the lncRNAs involved in melanoma, we analyzed The Cancer Genome Atlas (TCGA) Skin Cutaneous Melanoma (SKCM) dataset, and searched the genes correlated with outcome of melanoma patients. FUT8-AS1 was identified as one of the most significantly correlated genes with melanomas' prognosis. In this study, we further investigated the expression and clinical reverence of FUT8-AS1 in melanoma. In vitro and in vivo gain-and loss-of-function assays were performed to elucidate the biological roles of FUT8-AS1 in melanoma. Furthermore, the mechanism of action of FUT8-AS1 was explored and we identified a relative novel mechanism of action of FUT8-AS1, which is the promotion of miR-145-5p biogenesis.

Human Tissue Samples
A total of 68 malignant melanoma tissues and 36 age and gendermatched skin tissues with melanocytic nevi were acquired from patients who had undergone surgical resection at the 969 th Hospital of PLA (Hohhot, Inner Mongolia, China). All tissue samples were diagnosed by histopathological examination. The Review Board of the 969 th Hospital of PLA approved this study. Written informed consents were obtained from all patients.

Cell Culture and Treatment
Human melanoma cell lines CHL-1 and SK-MEL-2 were acquired from Cell Resource Center, Chinese Academy of Sciences. CHL-1 and SK-MEL-2 cells were cultured in DMEM and MEM medium respectively supplemented with 10% fetal bovine serum (Gibco, Thermo Fisher Scientific). Where indicated, cells were treated with 1 µM MEK162 (Selleck) for indicated time.

Cell Viability, Proliferation, Migration, and Invasion Assays
Glo cell viability assay was performed to measure cell viability as previously described (38). Briefly, 3,000 indicated melanoma cells per well were plated into 96-well plates. At the indicated time, the luminescence values were detected by the Cell Titer-Glo Luminescent Cell Viability Assay (Promega) to record cell viability. Ethynyl deoxyuridine (EdU) incorporation assay was performed using the EdU Kit (Roche) to measure cell proliferation as previously described (38). Transwell migration assay was performed to measure cell migration as previously described (38). Transwell invasion assay was performed to measure cell invasion as previously described (38).

Xenografts in Nude Mice
Athymic BALB/c nude mice were purchased from Chinese Academy of Sciences and maintained in pathogen-free condition. The use of animals was approved by the Review Board of the 969 th Hospital of PLA (Hohhot, Inner Mongolia, China). A total of 3.0 × 10 6 indicated melanoma cells were subcutaneously injected into the flanks of nude mice. Subcutaneous xenograft volumes were detected using caliper every 7 days and calculated following the formula V = 0.5 × LW2 (L, tumor length; W, tumor width). At the 28 th day after injection, subcutaneous xenografts were resected and weighed. Subcutaneous xenografts were further used to carry out immunohistochemistry (IHC) staining with the primary antibody against Ki67 (ab15580, 1 µg/ml, Abcam). Subcutaneous xenografts were also used to perform TdTmediated dUTP Nick-End Labeling (TUNEL) staining using the One Step TUNEL Apoptosis Assay Kit (Beyotime, Shanghai, China). To evaluate melanoma liver metastasis in vivo, 3.0 × 10 6 indicated melanoma cells were intrasplenically injected into nude mice to construct liver metastasis model. At the 28 th day after injection, the mice were sacrificed and the livers were resected. Hematoxylin-eosin (H&E) staining was performed using the livers. To evaluate melanoma lung metastasis in vivo, 3.0 × 10 6 indicated melanoma cells were injected into tail vein of nude mice to construct lung metastasis model. At the 28 th day after injection, the mice were sacrificed and the lungs were resected. H&E staining was performed using the lungs.

T T A T C T T G T A A G A G G A C A A T C C
ACATTCAC-3' (forward) and 5'-TTGTCCTCTTACAAGATA AAGTAGGGCTTCG-3' (reverse). Wild type and NF90 binding sites mutated FUT8-AS1 was in vitro transcribed and biotinylated from pSPT19-FUT8-AS1 and pSPT19-FUT8-AS1mut, respectively, using the Biotin RNA Labeling Mix (Roche) and Sp6 RNA polymerase (Roche). After purification, 3 µg of wild type or NF90 binding sites mutated FUT8-AS1 were incubated with 1 mg of whole-cell lysates from CHL-1 cells at 25°C for 1 h. The streptavidin agarose beads (Thermo Fisher Scientific) were used to enrich biotinylated RNA and interacted proteins. The proteins present in the pull-down material were detected using western blot.

RNA Immunoprecipitation (RIP)
RIP assays were performed in indicated melanoma cells using the Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore) and an antibody against NF90 (5 µg per reaction; ab131004, Abcam) in accordance with the manufacturer's manual.

Statistical Analysis
GraphPad Prism v6.0 was employed to perform all statistical analyses. For comparisons, log-rank test, Mann-Whitney test, two-tailed unpaired t test, one-way ANOVA followed by Dunnett's multiple comparisons test, Spearman correlation analysis, or one-way ANOVA followed by Tukey's multiple comparisons test was conducted as indicated in figure legends. P < 0.05 was considered as statistically significant.

Reduced Expression of FUT8-AS1 Is Correlated With Inferior Prognosis of Melanoma
To search the genes correlated with prognosis of melanoma, we analyzed The Cancer Genome Atlas (TCGA) Skin Cutaneous Melanoma (SKCM) dataset to retrieve the differentially expressed genes between different vital statuses (P-value < 0.01, fold change ≥2, Supplementary Table 1). Among these differentially expressed genes, we noted FUT8-AS1, which has a relative more significant P-value and bigger fold change. Further analysis of the TCGA-SKCM dataset revealed that reduced expression of FUT8-AS1 indicated inferior overall survival ( Figure 1A). Two in silico tools, namely the Coding Potential Assessment Tool (CPAT) (http://lilab.research.bcm. edu/cpat/index.php) and the Coding Potential Calculator (CPC) (http://cpc2.cbi.pku.edu.cn/) were employed to calculate the coding potential of FUT8-AS1. CPC and CPAT scores of FUT8-AS1 were equally low as well-known lncRNA HOTAIR (Supplementary Figures 1A, B), which indicated the noncoding nature of FUT8-AS1. Next, we measured FUT8-AS1 expression in our own cohort including 36 benign nevi and 68 melanomas using real-time PCR. The results indicated that FUT8-AS1 was downregulated in melanoma tissues compared with benign nevi ( Figure 1B). Correlation analyses between FUT8-AS1 expression and clinicopathological characters indicated that reduced expression of FUT8-AS1 was correlated with thickness, ulceration, and metastasis ( Figures 1C-E). Moreover, Kaplan-Meier survival analysis indicated that reduced expression of FUT8-AS1 was correlated with inferior overall survival among these 68 melanomas ( Figure 1F). Thus, these findings suggest that FUT8-AS1 is downregulated in melanoma and reduced expression of FUT8-AS1 is correlated with aggressive clinical factors and inferior overall survival.

FUT8-AS1 Inhibits Melanoma Cell Proliferation, Migration, and Invasion
In Vitro Due to the significant correlation between FUT8-AS1 expression and clinical characteristics of melanoma, we next investigated the potential roles of FUT8-AS1 in melanoma. We constructed CHL-1 and SK-MEL-2 cells stably overexpressing FUT8-AS1 via FUT8-AS1 overexpression lentivirus mediated transfection. The overexpression efficiencies were confirmed by real-time PCR (Figures 2A, B). Glo cell viability assays indicated that both CHL-1 and SK-MEL-2 cells overexpressing FUT8-AS1 had reduced cell viabilities compared with their control cells, respectively ( Figures 2C, D). EdU incorporation assays further indicated that both CHL-1 and SK-MEL-2 cells overexpressing FUT8-AS1 had slower cell proliferation rates compared with their control cells, respectively ( Figure 2E). Transwell migration assays indicated that both CHL-1 and SK-MEL-2 cells overexpressing FUT8-AS1 had less migrated cell numbers compared with their control cells, respectively ( Figure 2F). Transwell invasion assays indicated that both CHL-1 and SK-MEL-2 cells overexpressing FUT8-AS1 had less invasive cell numbers compared with their control cells, respectively ( Figure  2G). Thus, these findings suggest that FUT8-AS1 inhibits melanoma cell proliferation, migration, and invasion in vitro.
To further confirm the tumor suppressive roles of FUT8-AS1 in melanoma, we constructed FUT8-AS1 stably silenced CHL-1 and SK-MEL-2 cells via two independent FUT8-AS1 specific shRNAs lentivirus mediated transfection (Figures 3A, B). Glo cell viability assays indicated that FUT8-AS1 silenced CHL-1 and SK-MEL-2 cells both had increased cell viabilities compared with their control cells, respectively ( Figures 3C, D). EdU incorporation assays indicated that FUT8-AS1 silenced CHL-1 and SK-MEL-2 cells both had quicker cell proliferation rates compared with their control cells, respectively ( Figure 3E). Transwell migration assays indicated that FUT8-AS1 silenced CHL-1 and SK-MEL-2 cells both had more migrated cell numbers compared with their control cells, respectively ( Figure 3F). Transwell invasion assays indicated that FUT8-AS1 silenced CHL-1 and SK-MEL-2 cells both had more invasive cell numbers compared with their control cells, respectively ( Figure 3G). Collectively, these findings suggest that FUT8-AS1 silencing promotes melanoma cell proliferation, migration, and invasion in vitro, further supporting the tumor suppressive roles of FUT8-A1 in melanoma.

FUT8-AS1 Inhibits Melanoma Growth and Metastasis In Vivo
To investigate the roles of FUT8-AS1 in melanoma growth in vivo, FUT8-AS1 stably overexpressed and control CHL-1 cells were subcutaneously inoculated into nude mice. Tumor volumes were measured every 7 days and the tumors were excised and weighed at the 28 th day after inoculation. The results indicated that the tumors formed by CHL-1 cells overexpressing FUT8-AS1 had a slower growth rate and formed a smaller tumor compared with the tumors formed by control cells (Figures 4A, B). Proliferation marker Ki67 IHC staining indicated that the tumors formed by CHL-1 cells overexpressing FUT8-AS1 had less Ki67 positive cells compared with the tumors formed by control cells ( Figure 4C). TUNEL staining indicated that the tumors formed by CHL-1 cells overexpressing FUT8-AS1 had more apoptotic cells compared with the tumors formed by control cells ( Figure 4D).
To investigate the roles of FUT8-AS1 in melanoma metastasis in vivo, CHL-1 cells overexpressing FUT8-AS1 or control were intrasplenically injected to construct liver metastasis model. The results indicated that CHL-1 cells overexpressing FUT8-AS1 formed less liver metastases compared with control CHL-1 cells ( Figure 4E). Furthermore, CHL-1 cells overexpressing FUT8-AS1 or control were injected into the tail veins of nude mice to construct lung metastasis model. The results indicated that CHL-1 cells overexpressing FUT8-AS1 formed less lung metastases compared with control CHL-1 cells ( Figure 4F). Thus, these findings suggest that FUT8-AS1 inhibits melanoma growth and metastasis in vivo, further supporting the tumor suppressive roles of FUT8-AS1 in melanoma.

FUT8-AS1 Represses NRAS/MAPK Signaling
In our previous report, we have found that miR-145-5p activated MAPK signaling via directly targeting NRAS (29). Therefore, we further investigated the effects of FUT8-AS1 on NRAS/MAPK signaling. Our findings revealed that NRAS mRNA levels were decreased in FUT8-AS1 overexpressed cells and increased in FUT8-AS1 silenced cells ( Figures 6A, B). Consistently, NRAS protein levels were decreased in FUT8-AS1 overexpressed cells and increased in FUT8-AS1 silenced cells (Figures 6C, D). Next, the effects of FUT8-AS1 on MAPK signaling were investigated. As shown in Figures 6E, F, the phosphorylation levels of MEK1/ 2 and ERK1/2 were decreased in FUT8-AS1 overexpressed cells and increased in FUT8-AS1 silenced cells (Figures 6E, F). Thus, these data suggest that FUT8-AS1 represses NRAS/MAPK signaling in melanoma.
The Tumor Suppressive Roles of FUT8-AS1 Are Dependent on the Regulation of miR-145-5p/NRAS/MAPK Signaling Axis To elucidate whether the tumor suppressive roles of FUT8-AS1 in melanoma were dependent on the regulation of miR-145-5p/ NRAS/MAPK signaling axis, we inhibited miR-145-5p in FUT8-AS1 overexpressed CHL-1 cells ( Figure 7A). Glo cell viability assays indicated that the reduced cell viabilities caused by FUT8-AS1 overexpression were reversed by miR-145-5p inhibition ( Figure 7B). EdU incorporation assays indicated that the slower cell proliferation rates caused by FUT8-AS1 overexpression were reversed by miR-145-5p inhibition ( Figure  7C). Transwell migration assays indicated that the reduced migrated cell number caused by FUT8-AS1 overexpression were reversed by miR-145-5p inhibition ( Figure 7D). Transwell invasion assays indicated that the reduced invasive cell number caused by FUT8-AS1 overexpression was reversed by miR-145-5p inhibition ( Figure 7E). Collectively, these findings showed that the suppressive roles of FUT8-AS1 in melanoma cell proliferation, migration, and invasion were reversed by miR-145-5p inhibition.
To further elucidate whether the roles of FUT8-AS1 in melanoma were dependent on the regulation of MAPK signaling, we treated FUT8-AS1 silenced and control CHL-1 cells with MEK1/2 inhibitor MEK162. Glo cell viability assays indicated that MEK162 treatment abolished the increased cell viabilities caused by FUT8-AS1 silencing ( Figure 8A, compared with Figure 3C). EdU incorporation assays indicated that MEK162 treatment abolished the accelerated cell proliferation rate caused by FUT8-AS1 silencing ( Figure 8B, compared with Figure 3E). Transwell migration assays indicated that MEK162 treatment abolished the increased migrated cell number caused by FUT8-AS1 silencing ( Figure 8C, compared with Figure 3F). Transwell invasion assays indicated that MEK162 treatment abolished the increased invasive cell number caused by FUT8-AS1 silencing ( Figure 8D, compared with Figure 3G). Thus, these findings suggested that the roles of FUT8-AS1 in melanoma were dependent on the regulation of miR-145-5p/ NRAS/MAPK signaling.

DISCUSSION
Great progressions in molecule targeted therapy and immunotherapy have highlighted the therapeutic strategies based on molecular aberrations in melanoma (43). Current targeted therapies mainly focus on aberrant proteins. Recently, increasing evidences have revealed that except for proteins, ncRNAs also demonstrate important roles in cancers and would be developed as therapeutic targets (4,44,45). In this study, via analyzing public TCGA dataset, we identified a novel melanoma correlated lncRNA FUT8-AS1. FUT8-AS1 is locating at chromosome 14q23.3 and has only one exon. The knowledge of FUT8-AS1 is lacking with only one report showing that (E) Cell invasion of CHL-1 cells overexpressing FUT8-AS1 and concurrently silencing miR-145-5p were determined by the transwell invasion assay. Scale bars, 100 µm. Data are presented as mean ± SD. *P < 0.05, **P < 0.01, ***P < 0.001; ns, not significant, by one-way ANOVA followed by Tukey's multiple comparisons test.
FUT8-AS1 is correlated with outcome of a subpopulation of glioblastoma multiforme patients (46). In this study, we further investigated the expression, clinical significance, roles, and mechanism of action of FUT8-AS1 in melanoma.
FUT8-AS1 is reduced in melanoma tissues compared with benign nevi. Reduced expression of FUT8-AS1 is correlated with thickness, ulceration, metastasis, and inferior overall survival of melanomas. Gain and loss-of-function assays demonstrated that FUT8-AS1 suppresses melanoma cell proliferation, migration, and invasion in vitro. Furthermore, FUT8-AS1 represses melanoma growth and metastasis in vivo. Thus, FUT8-AS1 is identified as a tumor suppressive lncRNA in melanoma. FUT8-AS1 would also be a potential prognostic biomarker for melanoma. Enhancing FUT8-AS1 and/or its effects in downstream signaling would be potential therapeutic strategies for melanoma.
In addition, we identified a relative novel molecular mechanism of FUT8-AS1, which is the promotion of miR-145-5p biogenesis via competitively binding NF90. Previous reports about the effects of lncRNAs on miRNAs are mainly the competitive binding of miRNAs by lncRNAs, which further relieve the repressive roles of miRNAs on their mRNA targets (47). lncRNA-ATB is a classic example, which competitively binds miR-200s and upregulates miR-200s targets ZEB1 and ZEB2 in hepatocellular carcinoma (48). In our previous report, we have demonstrated that lncRNA MHENCR competitively binds miR-425 and miR-489, and therefore upregulates their targets IGF1 and SPIN1 in melanoma (39). In this study, we found that FUT8-AS1 does not bind miR-145-5p, but promotes miR-145-5p biogenesis.
Combining bioinformatic prediction and experimental verification, we found that FUT8-AS1 directly binds NF90. NF90 is a well-characterized dual strand RNA binding protein. Intriguingly, our previously identified tumor suppressive miR-145-5p was recently reported to be downstream target of NF90 by Zhuang et al. (42). NF90 directly binds pri-miR-145 and represses the biogenesis of mature miR-145-5p (42). In this study, we found that via competitively binding NF90, FUT8-AS1 represses the binding of NF90 to pri-miR-145. Therefore, FUT8-AS1 represses the effects of NF90 on miR-145-5p biogenesis. Lastly, FUT8-AS1 downregulates pri-miR-145 and upregulates mature miR-145-5p. The significant positive correlation between FUT8-AS1 and miR-145-5p expression in melanoma tissues supports the positive modulation of miR-145-5p by FUT8-AS1 in human melanoma. Furthermore, functional rescue assays showed that miR-145-5p inhibitors reverse the tumor suppressive roles of FUT8-AS1 in melanoma, supporting miR-145-5p as an important downstream target of FUT8-AS1. Several reports, including our own, have identified NRAS as a critical downstream target of miR-145-5p (29,49,50). Via targeting NRAS, miR-145-5p further represses MAPK signaling. MAPK signaling has been frequently shown to be aberrantly activated in melanoma (3). Therefore, in this study we further investigated the effects of FUT8-AS1 on NRAS/MAPK signaling via regulating miR-145-5p. In line with expectation, our results showed that FUT8-AS1 represses the expression of NRAS and further represses MAPK signaling in melanoma. MAPK signaling inhibitor MEK162 abolished the oncogenic roles of FUT8-AS1 silencing in melanoma, which further supports that miR-145-5p/NRAS/MAPK signaling is the critical mediator of the roles of FUT8-AS1 in melanoma. Therefore, we identified relative novel mechanism of action of lncRNA and also the detailed downstream molecular signaling cascade of FUT8-AS1 in melanoma.
In summary, our findings suggested FUT8-AS1 as an important tumor suppressive lncRNA in melanoma, which represses melanoma cell proliferation, migration, and invasion in vitro, and melanoma tumor growth and metastasis in vivo. FUT8-AS1 exerts its tumor suppressive roles via binding NF90, relieving the repressing roles of NF90 on miR-145-5p biogenesis, upregulating miR-145-5p, repressing NRAS, and lastly repressing MAPK signaling in melanoma. Furthermore, our data and TCGA data both suggested that FUT8-AS1 is correlated with inferior outcome of melanoma patients. Overall, these data provide novel molecular driver of melanoma progression and potential target for melanoma outcome prediction and therapy.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The studies involving human participants were reviewed and approved by the Review Board of the 969 th Hospital of PLA. The patients/participants provided their written informed consent to participate in this study. The animal study was reviewed and approved by the Review Board of the 969 th Hospital of PLA.