The prognostic value of long non-coding RNA PlncRNA-1 in patients with cancers: a systematic review and meta-analysis

Background We performed this meta-analysis to elucidate whether the expression of PlncRNA-1 might serve as an effective prognostic marker for various cancers. Methods We conducted a database search of PubMed, ScienceDirect, Embase, Web of Science and CNKI database (up to Oct 31, 2019). The pooled hazard ratio (HR), odds ratio (OR) and 95% confidence interval (CI) were used to estimate the strength of the relationship between PlncRNA-1 expression and the clinical prognosis of cancer patients. Results The results showed that elevated PlncRNA-1 expression predicted a poor OS with pooled HRs of 1.43 (95% CI: 1.25-1.63, I 2 =63.1%, P=0.004). Likewise, we found that advanced tumour stages were associated with upregulated PlncRNA-1 expression in various cancer types (III–IV vs I–II: OR=2.79, 95% CI: 1.76-4.41, I 2 =0%, P=0.822),patients with high PlncRNA-1 expression might have an increased risk of large tumours (OR=2.03, 95% CI: 1.31-3.14, I 2 =67.1%, P=0.028). tool for various


Background
Long non-coding RNAs (lncRNAs) are a group of transcribed RNAs longer than 200 nt, that cannot be translated into proteins [1,2]. LncRNAs are being continually discovered by high-throughput sequencing [3,4]. Emerging evidence has identified lncRNAs in various human tissues and has shown lncRNAs to play an important role in the occurrence and development of diseases [5][6][7][8]. For example, lncRNAs participate in tumourigenesis with oncogenic or tumour-suppressive effects by regulating gene expression at the transcriptional and post-transcriptional levels [9]. LncRNAs regulate gene expression in many ways, such as by interfering with the promoters of genes, inducing histone modification, chromatin reorganization, the regulation of subcellular localization and the production of endogenous siRNAs [10,11]. The dysregulation of lncRNAs has been shown to be related to tumour formation, progression, invasion and metastasis in various types of cancer, such as breast cancer, colorectal cancer and ovarian cancer [12][13][14]. Some lncRNAs have been used as potential biomarkers for cancer diagnosis and as therapeutic and prognostic targets for cancer treatment [15,16].
However, the majority of the biological functions of lncRNAs are still unknown.
Accumulating evidence has demonstrated that PlncRNA-1 acts as a transcriptional regulator to modify various developmental processes. For example, upregulated PlncRNA-1 expression can modulate apoptosis and proliferation and induce epithelial-mesenchymal transition [17][18][19][20][21][22][23][24][25]. More recently, a study reported that the overexpression of PlncRNA-1 predicted unfavourable prognosis and promoted tumourigenesis in osteosarcoma [22], suggesting that PlncRNA-1 may serve as a prognostic biomarker for patients with osteosarcoma. In these studies, PlncRNA-1 was identified as a prognostic biomarker in malignant tumour patients. However, due to limitations of small patient samples and discrete outcomes, the evidence to prove the relationship between PlncRNA-1 and cancers remains insufficient. Therefore, to elucidate whether the expression of PlncRNA-1 serves as an effective prognostic marker for various cancers, we performed this meta-analysis by comprehensively analysing all previously published data.

Methods 2.1 Literature retrieval strategy
The selected publications were identified by using up-to-date electronic databases, including PubMed, ScienceDirect, Embase, Web of Science and CNKI database. The literature search included all relevant studies published until Oct 31, 2019. The following key words were used in combination for the search: "PlncRNA-1", "CBR3-AS1", "lncRNA CBR3-AS1", and "cbr3-as1".

Selection criteria
The inclusion criteria were as follows: (1) PlncRNA-1 expression was assessed in human cancer tissues or blood. (2) According to the expression levels of PlncRNA-1, the patients were divided into low-and high-expression groups. (3) The hazard ratios (HRs) and 95% confidence intervals (CIs) for survival time were available or could be calculated from the survival curve.

Quality assessment
The assessment was performed by two authors, who had already reached an agreement on all items assessed. The quality of the papers was assessed as previously reported [27,28]. Briefly, the assessment system consisted of four items: scientific design, laboratory methodology, generalizability and results analysis. Each part was scored as follows: 2 points (if it was clearly defined in the article), 1 point (if its description was incomplete or unclear) and 0 point (if it was not defined or was inadequate). The final quality score was calculated using the sum of the total points divided by 44 and multiplied by 100. Half of the investigated studies defined 85% of the quality score as the cut-off point. Higher scores represented high methodological quality. The work was reported in line with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) and AMSTAR (Assessing the methodological quality of systematic reviews) Guidelines.

Data extraction
Two investigators independently extracted data from the included studies. Any problems were discussed between them. If these problems could not be overcome, a third investigator was consulted. For each eligible study, the following information was extracted: first author, year of publication, cancer type, sample size, PlncRNA-1 detection method, number of low-PlncRNA-1expression groups and high-PlncRNA-1-expression groups, tumour stage, cut-off value, follow-up duration, and results of multivariate or univariate analyses, HR with the corresponding 95% CI for overall survival (OS) and HR retrieval method. The HRs were extracted by two methods. In method 1, the HRs were obtained directly from the corresponding articles. In method 2, the HRs were extracted from Kaplan-Meier curves [27].

Statistical analysis
The extracted data were analysed with STATA software version 12.0 (STATA Corporation, College Station, TX, USA). HRs with 95% CIs were used to estimate the strength of the relationship between PlncRNA-1 and the clinical prognosis of cancer patients. In this meta-analysis, some HRs and their 95% CIs could not be extracted directly from original texts; thus, we calculated HRs with Kaplan-Meier curves using Engauge Digitizer version 4.1. Other HRs and their 95% CIs were collected from eligible articles. In this study, the heterogeneity among the included studies was quantified by the chisquared test and I 2 statistics [29]. If I 2 >50% or P < 0.1, a random-effects model was applied (it indicated strong heterogeneity across studies). If P > 0.1 and I 2 < 50%, a fixed-effects model was applied (it indicated nonsignificant heterogeneity among studies). Subgroup analyses were carried out according to cancer type, quality score, sample size, HR estimation method and follow-up duration. Begg's funnel plots were employed to evaluate the publication bias. A Galbraith radial plot was used to illustrate the sources of heterogeneity across the studies. As usual, P<0.05 was considered statistically significant.

Study characteristics
A total of seven studies and 829 patients were finally included in the meta-analysis based on the screening criteria. The mean patient sample size was 118 (range 70-318). The flowchart of the meticulous process of study retrieval is shown in Fig1 (PRISMA 2009 Flow Diagram). The published period of eligible studies ranged from 2016 to 2019, suggesting that the prognostic value of PlncRNA-1 is a novel field of research. Among the seven articles, six studies were from China, while the other study was from Iran. In this meta-analysis, patients with eight types of cancer were enrolled, including digestive system malignancies (colorectal cancer, hepatocellular carcinoma and gastric cancer), neurologic system tumours (glioma and glioblastoma multiforme) and other system carcinomas (osteosarcoma, breast cancer, and lung adenocarcinoma). According to PlncRNA-1 expression, the patients were divided into two groups, namely, high-and low-PlncRNA-1-expression groups. The main features of and data from the eligible studies are summarized in Table1. For the subgroups based on cancer type, we classified all tumours into two categories (digestive system malignancies and other system malignancies). The results showed that there was a significant association between poor OS and digestive system malignancies (HR = 2.28, 95% CI: 1.50-3.45), with no significant heterogeneity (I 2 = 7.5%, P = 0.356), while there was high heterogeneity (I 2 = 66.7%, P = 0.010) in other system malignancies (HR = 1.47, 95% CI: 1.10-1.95) (Fig3b).
Next, we evaluated the relationship between the quality of the selected paper in the studies and OS.
We found that the scores did not influence the result of the estimated HR (score ≥85%: HR = 2.00,

Association between PlncRNA-1 expression and tumour size
Four studies with a total of 370 patients were used to estimate the correlations between PlncRNA-1 expression levels and tumour size. A strong relationship was observed between high PlncRNA-1 expression and large tumours with high heterogeneity (OR = 2.03, 95% CI: 1.31-3.14, I 2 = 67.1%, P = 0.028) (Fig5). Therefore, the result also showed that patients with high PlncRNA-1 expression might have an elevated risk of large tumours.

Sensitivity analysis and publication bias
We conducted sensitivity analyses to investigate the reliability of our pooled estimates by omitting one study at a time. The results showed that the pooled HRs of OS were reliable, regardless of which study was excluded, and the significance of the HRs did not change (Fig6). To illustrate the sources of heterogeneity across the studies, we employed a Galbraith radial plot. As shown in Fig7, the study by 4. Discussion role in the development and progression of tumours [30,31]. LncRNAs have become a hot topic as diagnostic markers of diseases. The transcriptional level of PlncRNA-1 is significantly elevated in multiple malignancies, and PlncRNA-1 acts as a transcriptional regulator of various developmental processes [17][18][19][20][21][22][23][24][25]. In several human tumours, overexpression of PlncRNA-1 is related to poor OS, high TNM classification, advanced clinical stage, low histological differentiation, and poor vital status [20,22,24,[32][33][34]. The present meta-analysis was conducted to clarify the prognostic value of PlncRNA-1 expression in all cancer types and to examine its correlation with the main clinicopathological characteristics. and AR contributed to prostate cancer pathogenesis [18]. This conclusion was consistent with the results of another study showing that PlncRNA-1 could regulate a feed-forward loop (PlncRNA-1 protected AR from microRNA-mediated inhibition by sponging AR-targeting microRNAs) to contribute to the development of prostate cancer [21]. Yang et al investigated the function of PlncRNA-1 and discovered that PlncRNA-1 could regulate the cell cycle and cyclinD1 levels and could also affect apoptosis and proliferation in prostate cancer cells through the Her-2 pathway [19]. Song et al demonstrated that PlncRNA-1 promoted colorectal cancer cell progression by regulating the PI3K/Akt signalling pathway [20]. Another study noted that the upregulation of PlncRNA-1 indicated poor prognosis and promoted glioma progression by activating the Notch signalling pathway [24]. In this meta-analysis, as shown in Table3, we systematically analysed data regarding PlncRNA-1 and its potential targets, related microRNAs and pathways to provide a reference for exploring its mechanism in cancer and targeted therapy.
Although we tried our best to conduct a comprehensive study, this meta-analysis still has many limitations. First, some of the HRs were calculated indirectly by reconstructing Kaplan-Meier survival curves rather than being extracted directly from the corresponding articles. This may have resulted in bias and heterogeneity. Second, the conclusion may be weak due to the relatively small number of included studies and sample size. Third, PlncRNA-1 might display different biological functions in different malignant tumours, and due to the limited number of included articles, we could not pool the results according to a single type of tumour. Although we performed subgroup analyses, heterogeneity was still unavoidable. Fifth, the majority of the included studies were from Asia which might reduce the applicability of the results across different ethnicities. In addition, the tendencies of positive outcomes in the publications might give rise to potential selection and publication bias.

Conclusions
In summary, this meta-analysis revealed that high expression of lncRNA PlncRNA-1 represents a significant risk factor for survival outcomes in the development of tumours in patients with different types of cancer and could develop as an independent factor for predicting the prognosis of cancer patients.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests

Funding
This study was funded by grants from Projects of medical and health technology development program in Shandong province 2018WS316 2018WS317 .

Authors' contributions
Guarantor of integrity of entire study: HJ Yang       Galbraith radial plot analysis to illustrate the sources of heterogeneity across the studies Figure 8 Begg's funnel plot of publication bias