Genetic Polymorphisms of lncRNA LINC00673 as Predictors of Hepatocellular Carcinoma Progression in an Elderly Population

Long noncoding (lnc)RNAs are reported to be key regulators of tumor progression, including hepatocellular carcinoma (HCC). The lncRNA long intergenic noncoding RNA 00673 (LINC00673) was indicated to play an important role in HCC progression, but the impacts of genetic variants (single-nucleotide polymorphisms, SNPs) of LINC00673 on HCC remain unclear. A TaqMan allelic discrimination assay was performed to analyze the genotypes of three tagging SNPs, viz., rs9914618 G > A, rs6501551 A > G, and rs11655237 C > T, of LINC00673 in 783 HCC patients and 1197 healthy subjects. Associations of functional SNPs of LINC00673 with HCC susceptibility and clinicopathologic variables were analyzed by logistic regression models. After stratification by confounding factor, we observed that elderly patients (≥60 years) with the LINC00673 rs9914618 A allele had an increased risk of developing HCC under a codominant model (p = 0.025) and dominant model (p = 0.047). Moreover, elderly patients carrying the GA + AA genotype of rs9914618 exhibited a higher risk of having lymph node metastasis compared to those who were homozygous for the major allele (p = 0.013). Genotype screening of rs9914618 in HCC cell lines showed that cells carrying the AA genotype expressed higher LINC00673 levels compared to the cells carrying the GG genotype. Further analyses of clinical datasets from the Cancer Genome Atlas (TCGA) showed that LINC00673 expressions were upregulated in HCC tissues compared to normal tissues, and were correlated with advanced clinical stages and poorer prognoses. In conclusions, our results suggested that the LINC00673 rs9914618 polymorphism may be a promising HCC biomarker, especially in elderly populations.


Introduction
Hepatocellular carcinoma (HCC), accounting for over 90% of all primary hepatic malignancies, is ranked as the fourth leading cause of cancer deaths globally [1]. HCC initiation has been majorly correlated with environmental and genetic factors. For example, obesity, aflatoxin intake, alcohol and cigarette consumption, and hepatitis B or C virus (HBV or HCV) infection are usually recognized as environmental factors contributing to HCC development [1]. The HCC incidence varies geographically. In Taiwan, HBV infection is still a major risk factor for HCC patients [2]. As for genetic factors, emerging singlenucleotide polymorphisms (SNPs) in some noncoding genes such as micro (mi)RNA and long noncoding (lnc)RNAs were reported to be associated with HCC risk, development, and prognosis [3][4][5]. SNPs are some of the most common heritable mutations that induce DNA sequence polymorphisms at the gene level and were shown to have potential value for HCC prediction.
LncRNAs are a novel class of noncoding RNAs (>200 nt) which were shown to play significant roles in influencing the tumorigenesis, metastasis, and prognosis of cancers [6]. Recently, genome-wide association studies (GWASs) showed that only a very small portion of SNPs associated with complex diseases such as cancers was located in protein-coding regions. In contrast, the remaining part (>90%) was located in noncoding regions [7,8]. To the present, there are still few studies that focused on lncRNA polymorphisms as predictive biomarkers for HCC risk and progression.
LINC00673, also known as SRA-like noncoding RNA (SLNCR), is an lncRNA located on chromosome 17q24.3. Based on the Ensembl genome browser, the locus of LINC00673 largely overlaps with the locus of LINC00511, and 106 transcripts of LINC00673 were found to be derived from LINC00511, including LINC00673-v1-5 [9]. Recent studies indicated that LINC00673 acts as an oncogene or tumor suppressor gene in the occurrence and development of several cancer types including gastric [10], breast [11], pancreatic [12], lung [13], and prostate cancers [14]. In HCC, LINC00673 was reported to compete with and absorb miR-205 and promote progression of HCC [15]. As to the impact of LINC00673 SNPs on cancer, current studies mainly focused on rs11655237, a common noncoding transcript variant of LINC00673. This SNP was shown to increase the susceptibility of populations to gastric cancer [16], cervical cancer [17], pancreatic cancer [18], and neuroblastomas [19]. Moreover, the rs11655237 polymorphism was reported to be correlated with a risk of hepatoblastomas (HBs), the most common childhood hepatic malignancy, in Chinese children [20]. However, the roles of functional SNPs of LINC00673 within the context of HCC in adult populations have not yet been investigated.
Herein, rs11655237, together with two SNPs (rs6501551 and rs9914618) located in LINC00673 with a RegulomeDB score of <3, were chosen as tagSNPs. We investigated their associations with the risk and clinical characteristics of HCC in an adult Taiwanese population sample.

Study Population Characteristics
Demographic characteristics of recruited subjects are shown in Table 1. This study group comprised 783 pathologically confirmed HCC patients (542 males and 241 females) and 1197 cancer-free controls (837 males and 360 females). No significant differences between HCC patients and healthy controls were observed in terms of the distributions of age <60 and ≥60 years (p = 0.383), gender (p = 0.739), or smoking status (p = 0.284). Consistent with findings from other studies [21], significantly higher frequencies of HCC patients, compared to healthy controls, had a habit of alcohol consumption (14.1% vs. 35%; p < 0.001) and were positive for the hepatitis B surface antigen (HBsAg) (12.2% vs. 34%; p < 0.001). In HCC patients, higher proportions were diagnosed as being at early clinical (72.8%) and T stages (73.6%), with liver cirrhosis (59%) or without lymph node (97.2%), distal metastasis (96.2%), or vascular invasion (64.1%).

Association Studies of LINC00673 Genetic Polymorphisms and HCC Risks
To investigate possible associations of LINC00673 gene polymorphisms with the risk of developing HCC, genotype frequencies of selected tagSNPs (rs9914618, rs6501551, and rs11655237) were evaluated in all recruited cohort ( Table 2). Genotypic frequencies of these tagSNPs in the healthy control group conformed to Hardy-Weinberg equilibrium (rs9914618: p = 0.639; rs6501551: p = 0.284; rs11655237: p = 0.868). After adjusting for potential confounding factors including age, gender, alcohol consumption, and cigarette smoking, we observed no significant correlations of these LINC00673 variants with the occurrence of HCC between HCC patients and controls ( Table 2). The recruited cohort was further divided by age, and we observed that elderly patients (≥60 years) with the GA/AA genotypes of LINC00673 rs9914618 had an increased risk of HCC under the codominant model (GA vs. GG: adjusted odds ratio (AOR), 1.328; 95% confidence interval (CI), 1.036-1.703; p = 0.025) and dominant model (GA + AA vs. GG: AOR, 1.129; 95% CI, 1.002-1.273; p = 0.047) ( Table 3).

Relationships of LINC00673 Genetic Polymorphisms with Clinicopathological Features in HCC Patients
To further investigate the impacts of LINC00673 genetic polymorphisms on HCC progression, several clinicopathological features such as primary tumor size, clinical stage, tumor vascular invasion and metastases, hepatitis viral infection, and liver cirrhosis were chosen and are shown in Table 4. We observed that patients carrying at least one minor allele of rs9914618 (GA and AA) were prone to develop lymph node (LN) metastasis, compared to their corresponding wild-type genotype (GG) (OR, 2.346; p = 0.073). The HCC population was further divided into younger (<60 years) and elderly (≥60 years) groups and differences between LINC00673 SNPs and HCC in clinicopathological features were determined for these two groups. The results showed only elderly HCC patients harboring at least one minor allele of rs9914618 (GA and AA) had a significantly (p = 0.013) 3.970-fold higher risk (95% CI, 1.227-12.845) of developing LN metastasis compared to those homologous to the major allele (Table 5).

Upregulation of LINC00673 Is Observed in HCC Tissues and Correlated with Tumor Progression and a Poor Prognosis
Considering the potential effects of LINC00673 polymorphic genotypes on LINC00673 expression levels [18], correlations of LINC00673 expression levels with clinical significance and survival rates in HCC patients were further analyzed by examining cases of HCC from the TCGA dataset. According to the GEPIA2 website, we observed the prognostic value of LINC00673 in 33 different cancer types and found that high expression of LINC00673 showed poor prognostic impacts on four cancer types, including adrenocortical carcinoma (ACC), kidney renal clear cell carcinoma (KIRC), thymoma (THYM), and HCC ( Figure 1A, left panel) and Kaplan-Meier curves for overall survival (OS) of patients with HCC are shown in the right panel of Figure 1A. We also observed significantly higher LINC00673 transcripts in HCC compared to noncancerous tissues ( Figure 1B). Furthermore, patients with advanced clinical stages (II or III) showed significantly higher LINC00673 expression in tumors compared to patients at an early clinical stage (I) ( Figure 1C). The clinical data mentioned above suggest that LINC00673 genetic variants may affect LINC00673 expression levels and subsequently modulate the formation and progression of HCC.

The Correlations of LINC00673 Genetic Variants with LINC00673 Expression Levels
We next examined the correlations between LINC00673 rs9914618 genotypes and LINC00673 expression levels among six HCC cell lines (Mahlavu, PLC5, HCC36, SK-HEP-1, Huh7, and HepG2). We observed that Mahlavu, PLC5, and HCC36 cells carried the AA genotype of rs9914618 compared to Huh7 and HepG2 cells, which carried the GG genotype ( Figure 2, lower panel). From the results of RT-qPCR, we found that Mahlavu, PLC5, and HCC36 cells harboring the AA genotype expressed higher LINC00673 levels than Huh7 and HepG2 cells harboring the GG genotype (Figure 2, upper panel).

The Correlations of LINC00673 Genetic Variants with LINC00673 Expression Levels
We next examined the correlations between LINC00673 rs9914618 genotypes and LINC00673 expression levels among six HCC cell lines (Mahlavu, PLC5, HCC36, SK-HEP-1, Huh7, and HepG2). We observed that Mahlavu, PLC5, and HCC36 cells carried the AA genotype of rs9914618 compared to Huh7 and HepG2 cells, which carried the GG genotype ( Figure 2, lower panel). From the results of RT-qPCR, we found that Mahlavu, PLC5, and HCC36 cells harboring the AA genotype expressed higher LINC00673 levels than Huh7 and HepG2 cells harboring the GG genotype ( Figure 2, upper panel).

The Correlations of LINC00673 Genetic Variants with LINC00673 Expression Levels
We next examined the correlations between LINC00673 rs9914618 genotypes and LINC00673 expression levels among six HCC cell lines (Mahlavu, PLC5, HCC36, SK-HEP-1, Huh7, and HepG2). We observed that Mahlavu, PLC5, and HCC36 cells carried the AA genotype of rs9914618 compared to Huh7 and HepG2 cells, which carried the GG genotype ( Figure 2, lower panel). From the results of RT-qPCR, we found that Mahlavu, PLC5, and HCC36 cells harboring the AA genotype expressed higher LINC00673 levels than Huh7 and HepG2 cells harboring the GG genotype ( Figure 2, upper panel).

Discussion
Although HCC is recognized as one of the most prevalent forms of cancer, but the pathophysiology and underlying causes of HCC are less well understood. Therefore, identifying useful biomarkers for surveillance and early diagnosis of HCC is still deficient. Serum alpha fetal protein (AFP) is a common and clinically used tumor biomarker for HCC surveillance; however, recent reports indicated that the specificity and sensitivity of AFP for early diagnosis of HCC are not satisfactory [22]. Thus, there is still a need to search for novel biomarkers for early HCC detection. Accumulating evidence has manifested that several serum lncRNAs are potential biomarkers for predicting the occurrence, progression, and prognosis of HCC [23]. For example, Xu et al. found that serum levels of LINC00635 and ENSG00000258332.1 were upregulated in HCC patients and correlated with poor prognoses [24]. Wang et al. indicated that serum levels of lncRNA uc007biz.1 (LRB1) were positively correlated with tumor stages and negatively associated with OS in patients with HCC [25]. Moreover, Zheng et al. indicated that upregulation of urothelial cancerassociated (UCA) 1 in serum of HCC patients was associated with advanced TNM stages [26]. LINC00673 is a recently discovered lncRNA, and the oncogenic roles of LINC00673 in HCC were previously reported, including functions such as the promotion of proliferation and metastasis of HCC through negatively regulating miR-205 [15]. In the present study, we also observed that LINC00673 was upregulated in HCC and correlated with advanced clinical stages and poor prognoses of HCC patients. Although the oncogenic roles of LINC00673 in HCC have been studied, knowledge of the clinical relevance of LINC00673 SNPs in HCC, which might affect the functional changes and expression of LINC00673, is still lacking. Herein, we first identified that the LINC00673 SNPs play critical roles in influencing the occurrence and clinicopathological features of HCC in a Taiwanese population.
The present study demonstrates that individuals older than 60 years with the mutant base A of rs9914618 had a significantly higher risk of HCC occurrence and LN metastasis under a dominant model (GA + AA). These results were similar to our previous findings, which indicated that the LINC00673 rs9914618 SNP was linked to the lymphatic spread of oral cancer [27]. Moreover, Zhao et al. also indicated that the LINC00673 rs9914618 SNP was significantly associated with susceptibility to gastric cancer [16]. SNP variants of an lncRNA were shown to affect its expression or function due to structural changes and to further contribute to cancer progression [28]. Previous reports indicated that rs9914618 is located within the enhancer region containing a CCAAT box [27], the putative binding motif of transcription factors (TFs), including CCAAT/enhancer-binding proteins (C/EBPs) [29], and nuclear transcription factor Y (NF-Y) [30]. NF-Y and C/EBPs were reported to respectively play oncogenic and tumor-suppressive roles in HCC [31,32]. We suggest that rs9914618 polymorphisms may influence interactions with NF-Y and C/EBPs, thereby regulating HCC progression, but this issue should be further investigated in our future work. To further determine the effects of these variations on TF binding, we used variation annotation databases including RegulomeDB and VARAdb [33], and rs9914618-associated TF binding information was based on CHIP-seq data. Both databases showed that rs9914618 affected the binding of the TF termed structure-specific recognition protein 1 (SSRP1), in HCC cells (Figure 3). SSRP1 was reported to promote the proliferation and metastasis of HCC cells, and its upregulation in HCC tissues was correlated with higher T stages and shorter OS times [34], suggesting that rs9914618 variants may impact SSRP1 binding to modulate the progression of HCC. Actually, our present study has indicated that HCC cells carrying rs9914618 AA genotype expressed higher LINC00673 levels compared to cells carrying the GG genotype, suggesting that the A allele of rs9914618 may produce an increase in LINC00673 levels in HCC to promote its progression. that HCC cells carrying rs9914618 AA genotype expressed higher LINC00673 levels compared to cells carrying the GG genotype, suggesting that the A allele of rs9914618 may produce an increase in LINC00673 levels in HCC to promote its progression. Rs11655237 is a common noncoding transcript variant of LINC00673, and this genetic variation was investigated in different cancer types, but the results are still controversial. For example, studies showed that the LINC00673 rs7214041 polymorphism was significantly associated with the development of pancreatic cancer, neuroblastomas, and hepatoblastomas in Chinese populations [18][19][20]. In contrast, this SNP was not correlated with the susceptibility to pediatric gliomas or Wilms tumors in the same ethnic group [35,36]. In addition to Chinese populations, two GWASs demonstrated that rs11655237 SNP could increase the risk of pancreatic cancer in North American, Central European, Australian, and American Jewish populations, but a GWAS of women of European and African ancestry showed that this SNP was not correlated with susceptibility to breast cancer [37][38][39]. These results implied that different clinical impacts of rs11655237 on cancers may be due to different cancer types or ethnicities. In the present study, we observed that rs11655237 SNPs were not correlated with the predisposition to HCC in a Taiwanese population, but further exploration of this genetic factor in relation to HCC will require a larger sample size to verify the current findings.

Study Populations, Ethics, and Consent
HCC patient samples (N = 783) were collected from the National Biobank Consortium of Taiwan (NBCT) and Chung Shan Medical University Hospital (Taichung, Taiwan). In total, 1197 age-, gender-, and ethnicity-matched healthy controls were randomly selected Rs11655237 is a common noncoding transcript variant of LINC00673, and this genetic variation was investigated in different cancer types, but the results are still controversial. For example, studies showed that the LINC00673 rs7214041 polymorphism was significantly associated with the development of pancreatic cancer, neuroblastomas, and hepatoblastomas in Chinese populations [18][19][20]. In contrast, this SNP was not correlated with the susceptibility to pediatric gliomas or Wilms tumors in the same ethnic group [35,36]. In addition to Chinese populations, two GWASs demonstrated that rs11655237 SNP could increase the risk of pancreatic cancer in North American, Central European, Australian, and American Jewish populations, but a GWAS of women of European and African ancestry showed that this SNP was not correlated with susceptibility to breast cancer [37][38][39]. These results implied that different clinical impacts of rs11655237 on cancers may be due to different cancer types or ethnicities. In the present study, we observed that rs11655237 SNPs were not correlated with the predisposition to HCC in a Taiwanese population, but further exploration of this genetic factor in relation to HCC will require a larger sample size to verify the current findings.

Study Populations, Ethics, and Consent
HCC patient samples (N = 783) were collected from the National Biobank Consortium of Taiwan (NBCT) and Chung Shan Medical University Hospital (Taichung, Taiwan). In total, 1197 age-, gender-, and ethnicity-matched healthy controls were randomly selected from the Taiwan Biobank Project. All HCC patients had been pathologically confirmed and clinically staged according to the tumor, node, metastasis (TNM) staging system of the American Joint Committee on Cancer (AJCC). Through interviewer-administered questionnaires, we obtained the information about the history of smoking and alcohol consumption from all the recruited subjects. Before collecting venous blood, written informed consent was obtained from each participant, and the investigation protocol was approved by the Institutional Review Board of Chung Shan Medical University Hospital (IRB no. CS2-19133).

Genomic DNA Extraction from Blood
Whole-blood samples from all recruited subjects were collected and placed in ethylenediaminetetraacetic acid (EDTA)-containing tubes. Blood samples were immediately centrifuged to separate genomic DNA from buffy coats, which were isolated by using a QI-Aamp DNA Blood Mini Kit (Qiagen, Valencia, CA, USA) as previously described [40]. The quality of the final extracted DNA was checked using a Nanodrop-2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA) and preserved at −20 • C [41].

Selection of LINC00673 SNPs
We selected rs11655237 because this SNP was reported to be correlated with the risks of different cancer types [42]. Moreover, two other SNPs (rs6501551 and rs9914618) located in LINC00673 were selected based on their functional potential with RegulomeDB scores of <3 obtained from the RegulomeDB database.

Extraction of RNA and Reverse-Transcriptase Quantitative Polymerase Chain Reaction (RT-qPCR)
Total RNA was isolated from HCC cell lines using TRIzol reagent (Thermo Fisher Scientific) and amplified as described previously [44]. RT-qPCR was carried out using LINC00673-specific primers (forward: AATATTAAACGGTCCAGTCCTACAA; reverse: TAGGACTGCCCATTACAGAGGA) and Hot Firepol EvaGreen qPCR Mix Plus (Solis BioDyne, Tartu, Estonia), according to the manufacturer's instruction. Fluorescence data of detected genes were normalized to the expression of actin using the 2 −∆∆CT method.

Bioinformatics Analysis
RNA sequencing analysis and the visualization platform, Gene Expression Profiling Interactive Analysis 2 (GEPIA2), were applied to determine the prognostic effects of LINC00673/LINC00511(ENSG00000227036) in different cancer types including HCC. Correlations of LINC00673/LINC00511 with prognoses were calculated using the median cutoff. GEPIA2 performs data mining based on The Cancer Genome Atlas (TCGA) data. The expression level of LINC00637, which also refers to ENSG00000227036, and related clinical parameters in HCC patients were obtained from TCGA cohort, which was downloaded using UCSC Xena.

Statistical Analysis
Significant differences in categorical variables and demographic characteristic distributions between HCC patients and the healthy controls were determined using the Mann-Whitney U-test. Associations of LINC00673 genotypes with HCC susceptibility were determined using multiple logistic regression methods and were adjusted for potential confounders such as gender, age, cigarette smoking, and alcohol consumption. Differences in LINC00673 levels between normal and HCC tissues or in different clinical stages of HCC tissues obtained from TCGA were compared by an independent t test. The Statistical Analytical System (SAS Institute, Cary, NC, USA) software (vers. 9.1) was used to analyze all data, and a p value of <0.05 was considered significant.

Conclusions
At present, most HCC patients will eventually develop advanced disease and the treatment outcomes in these patients remain unsatisfactory. However, we still do not have reliable tools to predict who those patients are. Our present study indicated that elevated LINC00673 expression levels contribute to the development of advanced stages in HCC patients. We found HCC cells carrying rs9914618 AA genotype may cause an increase level of LINC00673. We first identified the diverse allelic effects of LINC00673 SNPs (rs9914618) which contribute to the susceptibility and LN metastasis of HCC in a Taiwanese population. These findings contribute to a better understanding of the risks and early detection of HCC.