The Long Noncoding Transcript HNSCAT1 Activates KRT80 and Triggers Therapeutic Efficacy in Head and Neck Squamous Cell Carcinoma

Head and neck squamous carcinoma (HNSC) is the most prevalent malignancy of the head and neck regions. Long noncoding RNAs (lncRNAs) are vital in tumorigenesis regulation. However, the role of lncRNAs in HNSC requires further exploration. Herein, through bioinformatic assays using The Cancer Genome Atlas (TCGA) datasets, rapid amplification of cDNA ends (RACE) assays, and RNA-FISH, we revealed that a novel cytoplasmic transcript, HNSC-associated transcript 1 (HNSCAT1, previously recognized as linc01269), was downregulated in tumor samples and advanced tumor stages and was also associated with favorable outcomes in HNSC. Overexpression of HNSCAT1 triggered treatment efficacy in HNSCs both in vivo and in vitro. More importantly, through high-throughput transcriptome analysis (RNA-seq, in NODE database, OEZ007550), we identified KRT80, a tumor suppressor in HNSC, as the target of HNSCAT1. KRT80 expression was modulated by lncRNA HNSCAT1 and presented a positive correlation in tumor samples (R = 0.52, p < 0.001). Intriguingly, we identified that miR-1245 simultaneously interacts with KRT80 and HNSCAT1, which bridges the regulatory function between KRT80 and HNSCAT1. Conclusively, our study demonstrated that lncRNA HNSCAT1 functions as a necessary tumor inhibitor in HNSC, which provides a novel mechanism of lncRNA function and provides alternative targets for the diagnosis and treatment of HNSC.


Introduction
Head and neck squamous cell carcinomas (HNSCs) are the most frequent malignancies of the head and neck regions and are a heterogeneous group of tumor entities covering several anatomical subsites in the head and neck [1]. Approximately half a million patients are diagnosed annually with HNSC, and it is a major contributor to cancerrelated mortality worldwide [2]. The pathogenesis of HNSC is multifactorial [3,4], including (a) an acute demand for protein synthesis that is often a result of oncogene activation, (b) inadequate oxygen and nutrient supply as well as chemo/radiotherapy exposure, and (c) immune surveillance escape [5][6][7]. as well as epigenetic alterations. For example, NOTCH, EGFR, and MET modifications enhance cell proliferation, migration, and survival via PI3K, RAS/RAF/ERK, and JAK/STAT signaling, and these pathways are frequently deregulated in HNSC [13]. p53 pathway disruption is associated with increased genomic instability [14]. Moreover, loss of epigenetic balance has been proven to be a driving event for the initiation and progression of HNSC [15,16]. For example, CDKN2A, a tumor suppressor gene, suppresses cyclin-dependent kinase expression as well as cell cycle progression. In HNSC cells, many promoters of tumor suppressors are often hypermethylated, contributing to the abnormal silencing of these tumor-inhibiting genes [17,18]. More importantly, long noncoding RNAs, which are transcripts (>200 nt) without coding capacity, have been proven to be associated with the pathogenesis of HNSC [19]. For instance, the lncRNA MIR31HG targets P21 and HIF1A to mediate head and neck cancer cell proliferation and tumorigenesis by stimulating cell cycle progression [20]. Moreover, an IFNα-initiated long noncoding RNA, lncMX1-215, exerts negative effects on immune suppression by interrupting the acetylation of H3K27 in HNSC [21]. However, the role of linc01269, a novel transcript with an unknown function, in the progression of HNSC remains to be further explored.
We thus aimed to determine the significance of abnormal lncRNA expression in the tumorigenesis of HNSC. Through bioinformatic assays and RACE assays, we found that a novel transcript, HNSCAT1 (previously identified as linc01269), was downregulated in HNSC and associated with favorable outcomes in the TCGA cohort. Moreover, overexpression of HNSCAT1 significantly inhibited tumor proliferation and migration in vitro and in vivo. Mechanistically, HNSCAT1 interacts with miRNA-1245 and promotes KRT80 expression. Our study elucidates the molecular biology of HNSCs and provides a basis for further studies on HNSC diagnosis and treatment.

Materials and Methods
2.1. Cell Culture. Primary keratinocytes were collected from eyelids during blepharoplasty. Specimens were washed in Dulbecco's phosphate-buffered saline containing 20 μg/mL gentamicin. The samples were cut into 5 mm 3 pieces, and the epidermis was isolated. Dispase (Roche) solution in Dulbecco's phosphate buffered saline with 5 μg/mL gentamicin was applied to the specimen pieces for 12 hours at 4°C. The epidermis was carefully separated and placed into a solution containing 0.05% trypsin-EDTA at 37°C for 15 minutes. The cells were then cultured in defined keratinocyte serum-free medium (Gibco). HaCaT, SCL-1, Cal27, and Colo16 cells were purchased from ATCC and cultured in DMEM (Gibco) supplemented with 10% fetal bovine serum (FBS; Gibco), streptomycin (100 mg/mL), and penicillin (100 U/mL). Incubation was performed at 37°C in a 5% CO 2 humid atmosphere. This study obtained ethical approval from the Independent Ethics Committee of Shanghai Ninth People's Hospital (registration no. 2015(72)).

Plasmid Construction and Rapid Amplification of cDNA
Ends. pLKO.1, pCDH, and pCMV were used in our study. ShRNA sequences were generated by PCR and then cloned into the pLKO.1 vector. The KRT80 and lncHNSCAT1 overexpression cassette was generated by PCR, cloned into the pCDH vector, and verified by DNA sequencing. A rapid amplification of cDNA ends (RACE) assay was carried out with the 5′ RACE System for Rapid Amplification of cDNA Ends (Invitrogen, #18374-058) and the 3 ′ RACE System for Rapid Amplification of cDNA Ends (Invitrogen, #18373-019) according to the manufacturer's protocols. The amplified DNA fragment was cloned into the pGEM-T Easy vector and validated by Sanger sequencing (Sangon Biotech, China).
2.3. CCK-8, Colony Formation, and Wound Healing Assays. Cell proliferation was assessed using CCK-8 assays (HY-K0301, MCE) according to the manufacturer's instructions. Briefly, cell seeding (3000 cells/100 mL) was performed in triplicate in 96-well plates. At various time points, the dye solution was added, followed by incubation at 37°C for 4 h. The absorbance was measured at 570 nm. For the colony formation assay, 1 mL of complete medium with 1,000 cells was added to each well of a six-well plate. After 1-2 weeks, the plate was stained using crystal violet (0.25%). A wound healing assay was performed after seeding 500,000 cells into a 6well plate. A wound was made by manually scraping the cell monolayer with a 200 μl pipet tip. Images were taken at the indicated times. The percentage wound healing was calculated using the formula: wound healing = ðinitial wound area − unhealed wound areaÞ/initial wound area × 100%.

Subcutaneous Xenograft
Study. The use of animals in this study was permitted by the Shanghai Jiao Tong University Animal Care and Use Committee. Experiments were performed following animal policies of Shanghai Jiao Tong University and the National Health and Family Planning Commission of China guidelines. Cell harvesting was performed by trypsinization, after which the cells were washed twice using PBS (GIBCO). Male BALB/c nude mice (aged 4 weeks) were used as models for experiments. Approximately 1 × 10 6 HNSC cells from each group were subcutaneously injected into the underarm region of the mice. At 14 days postinjection, the mice were euthanized and sacrificed for tumor resection. The formula V = ab 2 /2 was used to determine tumor volumes, with a and b denoting tumor length and width, respectively.
2.5. RNA-seq. The EZ-press RNA Purification Kit (B0004) was used for extraction of total RNA, whose integrity was confirmed using a 2100 Bioanalyzer (Agilent Technologies, USA). RNA concentrations were assessed using a Qubit 2.0 fluorometer and a Qubit RNA Assay Kit (Life Technologies, Carlsbad, CA, USA). Using total RNA (100 ng) and the Illumina TruSeq RNA Sample Prep Kit (San Diego, CA, USA) as instructed by the manufacturer, libraries were prepared. Sequencing and basic informatic assays were conducted by Kangcheng Biotech, Inc. (Shanghai, China). 2.11. Nuclear-Cytoplasmic Extraction. The cellular fraction was isolated as previously described [22]. In brief, 3 × 10 6 cells were harvested and resuspended in DEPC-PBS. The cell pellets were then resuspended in 2 ml of hypotonic buffer (10 mM Tris-HCl (pH 7.4), 20 mM MgCl 2 , and 4% Triton X-100), and the cells were gently ground by a Dounce homogenizer and incubated for 30 min on ice. Then, the cells were centrifuged for 15 min at 3000 rpm; the supernatant containing the cytoplasmic component and the pellet containing the nuclear fraction were kept for RNA extraction.
2.12. Statistical Analysis. We performed the statistical analyses with the GraphPad Prism 9 software. Data are presented as the mean ± SD or SEM. The differences between two groups were calculated by unpaired two-tailed Student's t -test as indicated. Survival plots were generated by Kaplan-Meier curves, and p values were calculated by the log-rank test. Biological triplicates were assessed where indicated. p < 0:05 was considered to indicate statistical significance ( * p < 0:05, * * p < 0:01, and * * * p < 0:001).

HNSCAT1 Expression Is Decreased in HNSC, and This
Downregulation Is Associated with a Favorable Outcome.
As the boundary of the long noncoding transcript is highly flexible [23], we then assessed whether linc01269 encodes a new transcript in HNSC cells. Based on the National Center for Biotechnology Information (NCBI) database, linc01269 lncRNA is 602 bp in length and has four exons. However, through rapid amplification of cDNA ends (RACE), we detected a novel isoform with different 5 ′ -and 3′-boundaries of linc01269 ( Figure 1(f)). This novel transcript was also identified to lack coding capacity ( Figure S2B, coding potential score = −1:19). Specifically, exons 2 and 3 were consistent with predicted exons;  BLCA  BRCA  CESC  CHOL  COAD  DLBC  GBM  HNSC  KICH  KIRC  KIRP  LAML  LGG  LIHC  LUAD  LUSC  MESO  OV  PAAD  PCPG  PRAD  PEAD  SARC  SKCM  STAD  TGCT  THCA  THYM  UCEC  UCS    Oxidative Medicine and Cellular Longevity however, exon 1 had an extra 166 bp fragment at its 5 ′ terminus, while exon 4 was extended by 303 bp. Compared to the predicted sequence, this transcript had an extra poly-A tail at its 3 ′ terminus ( Figure 1(g), Figure S2C). These findings imply that the linc01269 lncRNA isoform is a novel noncoding transcript in HNSC tumors; therefore, we named it HNSC-associated transcript 1 (HNSCAT1).

Overexpression of HNSCAT1 Attenuates HNSCs In Vitro
and In Vivo. To assess whether this novel HNSCAT1 was able to alter tumor behaviors, we first assessed lncHNSCAT1 expression levels in HNSC cell lines. Consistent with its expression level in TCGA samples, HNSCAT1 was significantly downregulated in HNSC cell lines, including SCL-1, Cal27, and Colo16 (Figure 2(a)), compared to normal keratinocyte cell lines (HaCaT) and primary keratinocytes (PKs). Through RNA-FISH and nuclear-cytoplasm separation assays, we found that HNSCAT1 was mainly distributed in the cytoplasm. In this assay, 18S RNA served as a cytoplasmic control, and U6 served as a nuclear RNA marker (Figures 2(b) and 2(c)). We further restored HNSCAT1 expression in HNSC cells by cloning full-length HNSCAT1 166bp extension 303bp extension  Oxidative Medicine and Cellular Longevity into an overexpression vector (Figure 2(d)). As expected, HNSCAT1-overexpressing HNSC cells presented an attenuated proliferation rate ( Figure S3) and formed smaller and fewer colonies (Figure 2(e)). Transwell assays (Figure 2(f)) and wound healing assays ( Figure S4A-D) demonstrated that HNSCAT1 overexpression resulted in impaired migration ability in HNSC cells. Moreover, a soft agar assay showed that tumorigenic capacity was inhibited after the restoration of HNSCAT1 expression (Figure 2(g)). Most importantly, in a xenograft mouse model, HNSCAT1-overexpressing cells formed smaller tumors, with a significant reduction in both size (Figure 2(h)) and weight (Figure 2(i)).   To further explore the mechanism underlying the tumor-inhibitory effect of HNSCAT1, RNA-seq analysis was performed in SCL-1 cells after overexpressing lncRNA HNSCAT1 (deposited in NODE database, https://www .biosino.org/node/login, OEZ007550). Through Gene Set Enrichment Analysis (GSEA), we found that cell junctionand adhesion complex-related pathways (KEGG_HSA04530, KEGG_HSA04510, and KEGG_HSA04520) were significantly activated after the restoration of HNSCAT1 expression (Figure 3(a)). In addition, through Circos analysis, we found that keratin 80 (KRT80), which encodes a type II keratin that is involved in terminal epithelial differentiation, was significantly upregulated after overexpression of HNSCAT1 (Figure 3(b)). We further found that in TCGA HNSC samples, KRT80 presented a strong positive correlation with HNSCAT1 (R = 0:52, p < 0:001, Figure 3(c)). We also further validated our finding in our RNA-seq data. We found that HNSCAT1-overexpressing HNSC cells presented enhanced reads in the KRT80 region (Figure 3(d)) and elevated expression levels at the mRNA ( Figure S6) and protein levels ( Figure 3(e)). Therefore, KRT80 is a potential downstream target of HNSCAT1.

KRT80 Is Downregulated in HNSC Samples.
Through a pancancer analysis of TCGA samples (Figure 4(a)), we found that KRT80 presented a similar expression pattern to HNSCAT1: both were significantly downregulated in HNSC and ESCA samples (Figure 4(b)). Moreover, we found that stage IV samples presented with lower expression than early-stage HNSC samples (Figure 4(c)). More importantly, elevated expression of KRT80 was associated with a more favorable outcome in terms of both DFS (Figure 4(d)) and OS (Figure 4(e)). We further validated the expression profiling of KRT80 in our cell lines. As expected, KRT80 was markedly downregulated in HNSC cells at the mRNA (Figure 4(f)) and protein (Figure 4(g)) levels.

Overexpression of KRT80 Attenuates HNSC Malignant
Behavior In Vitro and In Vivo. Since KRT80 presented with decreased expression in HNSC, we overexpressed KRT80 in SCL-1 and Cal27 cells. We found that the mRNA level of KRT80 was significantly upregulated by~5-fold after overexpression ( Figure 5(a)). Consistently, KRT80 protein was significantly enriched in KRT80-overexpressing cells ( Figure 5(b)). We were then interested in observing tumor behaviors in these cells. Notably, through a colony formation assay, we found that tumors formed smaller and fewer colonies after overexpression of KRT80 in both SCL-1 and Cal27 cell lines (Figures 5(c) and 5(d)). Moreover, cellular migration was also impaired in KRT80-overexpressing cells (Figures 5(e) and 5(f)). More importantly, we established a subcutaneous xenograft model and found that KRT80overexpressing SCL-1 cells formed smaller tumors, both in size ( Figure 5(g)) and weight ( Figure 5(h)). Taken together, these data indicate that both HNSCAT1 and KRT80 are tumor suppressor genes with decreased expression in HNSC.
3.6. miR-1254 Interacts with HNSCAT1 and KRT80. We were then curious about the molecular basis of the mechanism of the regulation of lncRNA HNSCAT1 and KRT80. Because lncHNSCAT1 is a cytoplasmic transcript, we hypothesized that lncRNA HNSCAT1 mainly functions through a competing endogenous RNA (ceRNA) network. First, we investigated potential HNSCAT1-binding miRNAs. Through TargetScan and miRDB, we found 268 miRNAs that may serve as binding candidates for HNSCAT1. Notably, 4 of them (miR-1254, miR-3150b-3p, miR-4505, and miR-149-3p) could also potentially interact with the 3 ′ UTR of KRT80 mRNA (Figure 6(a)). To further investigate which miRNA served as the regulator of HNSCAT1/KRT80, we  BLCA  BRCA  CESC  CHOL  COAD  DLBC  GBM  HNSC  KICH  KIRC  KIRP  LAML  LGG  LIHC  LUAD  LUSC  MESO  OV  PAAD  PCPG  PRAD  PEAD  SARC  SKCM  STAD  TGCT  THCA  THYM  UCEC  UCS  10 Oxidative Medicine and Cellular Longevity transfected these miRNA mimics into HNSCAT1overexpressing HNSC cells. Here, we found that miR-1254 significantly inhibited the expression of KRT80, while KRT80 expression remained unchanged in other miRNA groups (Figures 6(b) and 6(c), Figure S7). Consistently, the KRT80 protein signal also showed a significant reduction after exogenous miR-1254 transfection (Figure 6(d)). More importantly, we then established biotin-labeled miRNA-1254 and used it as a bait to capture miR-1254-interacting mRNAs (Figure 6(e)). Through a biotin pull-down assay, we identified that miR-1254 interacts with both KRT80 and HNSCAT1 mRNA, whereas the antisense miRNA showed a signal intensity similar to that of the negative control (only streptavidin beads) group (Figures 6(f) and 6(g)). Moreover, miRNA inhibition potencies were assessed by the dual-luciferase assay, whereby one of the reporter genes encoding Renilla luciferase was fused to wild-type and mutated 3 ′ UTR of KRT80. Here, we found that miR-1254 mimics significantly inhibited reporter activity, while the miRNA potential binding site-mutated group (lane 4-6) and miR-1254 inhibitor-treated group (lane 3) presented reporter signals similar to those of the control group ( Figure 6(h)). These data indicated that miR-1254 could bind with lncHNSCAT1 and KRT80 and inhibit KRT80 expression by interacting with its 3 ′ UTR.

miR-1254 Modulates HNSC Behaviors by Inhibiting
KRT80. Since miR-1254 bridges the regulation of HNSCAT1 and KRT80 expression in HNSC cells, we were interested in assessing the role of the HNSCAT1/miR-1254/KRT80 axis in HNSC tumorigenesis. In the colony formation assay, we found that HNSCAT1 overexpression-mediated tumor inhibition was largely compromised after introducing exogenous miR-1254 (Figures 7(a) and 7(b), lane 3), but the inhibitory effect could be rescued by adding miR-1254 inhibitors (Figures 7(a) and 7(b), lane 3). Moreover, for cell migration    analysis, we found that miR-1254 could also attenuate the effect of HNSCAT1 overexpression. Importantly, this inhibitory phenomenon could also be neutralized by miR-1254 inhibitors (Figures 7(c) and 7(d)). For cell growth analysis, we found that the cell growth inhibition effect after HNSCAT1 overexpression was largely restored when miR-1254 was reintroduced; however, the rescue efficacy was diminished after treatment with additional miR-1254 inhibitors in both SCL-1 (Figure 7(e)) and Cal27 cells (Figure 7(f)).
Thus, these results indicate that miR-1245 bridges the regulatory relationship between KRT80 and lncHNSCAT1.

Discussion
Globally, HNSC is the 6 th most prevalent human malignancy, accounting for more than 380 thousand mortalities annually worldwide [6,19,24]. In recent decades, the complexity and heterogeneity of HNSC have been recognized    Oxidative Medicine and Cellular Longevity [25]. In parallel, the development of high-throughput omics has given rise to a better picture of the behavior and characteristics of molecules during carcinogenesis [26]. lncRNAs are a class of functional RNA molecules that do not have translational capacities but play a vital role in the activities of transcription factors or regulate structural changes in chromatin. However, the function of lncRNAs in HNSC needs to be further explored.
To date, several lncRNAs have been recognized to have vital roles in cancer progression as well as metastasis [27][28][29][30][31]. For instance, HOTAIR is upregulated in HNSC, and its upregulation is related to an elevated rate of metastasis. More specifically, HOTAIR triggers the epithelialmesenchymal transition (EMT) process by recruiting EZH2 (an H3K27me3 modifier enzyme), which has been proven to be negatively associated with clinical outcomes in HNSC patients [32,33]. Most importantly, upregulation of the long noncoding RNA EGFR-AS1 mediates epidermal growth factor receptor addiction and modulates treatment resistance toward tyrosine kinase inhibitors (TKIs) in squamous cell carcinoma. EGFR-AS1 downregulation shifts splicing toward EGFR isoform D, leading to ligandmediated pathway activation [22]. Herein, we found that lncRNA linc01269 was one of the most downregulated lncRNAs in HNSC versus normal samples. Through RACE analysis, we found two linc01269-related transcripts with different 5 ′ and 3 ′ boundaries, presenting a novel transcript, which we named HNSC-associated transcript 1 (HNSCAT1). Moreover, overexpression of HNSCAT1 significantly inhib-ited tumor progression through HNSCAT1 interaction with miR-1254 and rescued KRT80 expression. Our study is the first to report existence and function of a novel transcript, HNSCAT1.
miRNA aberrations have been frequently identified in a wide variety of cancers [34]. For HNSC, previous studies have found 128 miRNAs were differentially expressed in HNSCC tissue in TCGA dataset. Moreover, several individual miRNAs have shown indicative regulatory functions [35]. For instance, miR-204-5p suppresses epithelialmesenchymal transition (EMT) and STAT3 signaling by targeting SNAI2, SUZ12, HDAC1, and JAK2 [36]. In addition, HPV+ HNSCC-derived exosomal miR-9 induces macrophage M1 polarization and increases tumor radiosensitivity, which provides a novel miRNA-based therapy [37]. However, the function of miR-01254 remains unexplored in HNSC. Herein, we found for the first time that miR-1254 interacts with both lncRNA HNSCAT1 and KRT80. After overexpression of HNSCAT1, the inhibition of miR-1254induced KRT80 degradation was partially rescued. The HNSCAT1/miR-1254/KRT80 signaling pathway serves as an important regulator in HNSC, and targeted treatments that address these epigenetic deficiencies might be promising strategies for HNSC.
Keratins have been revealed to play a vital role in the tumorigenesis of numerous cancer types [38][39][40]. In HNSC, highly keratinized squamous cell carcinoma exhibits an improved response to treatment and prognosis compared with weakly keratinized cancer [41]. Herein, we found that  Figure 7: miR-1254 promotes HNSC malignant behavior by inhibiting KRT80. (a, b) Colony formation assays were performed to determine proliferation capacity after overexpressing miRNA-1254 in SCL-1 and Cal27 cells. Experiments were conducted in triplicate, and the results are shown as the mean ± SEM. * p < 0:05. (c, d) Transwell assays showed that the migration ability was impaired after overexpressing miRNA-1254 in SCL-1 and Cal27 cells. All of the experiments were performed in triplicate and are presented as the mean ± SEM. * p < 0:05. (e, f) The CCK-8 assay indicated that the cell growth inhibition effect after HNSCAT1 overexpression was largely rescued when miR-1254 was reintroduced; however, the rescue efficacy was diminished after the addition of extra miR-1254 inhibitors in both (e) SCL-