1. LADC data sets and preprocessing
The LADC data sets, including RNA expression quantification data and related clinical data were downloaded from TCGA database (https://portal.gdc.cancer.gov/). The expression of mRNAs and lncRNAs were obtained from HTSeq-counts files on Illumina platform. A total of 533 LADC samples and 59 adjacent normal samples were collected. The IDs of mRNAs and lncRNAs were converted using GENCODE Human Release 31 (GRCh38.p12; https://www.gencodegenes.org/human/).
2. Differential analysis
To identify DElncRNAs and DEmRNAs over LADC and adjacent normal samples, the R package DESeq2 (version 1.22.2; http://www.bioconductor.org/packages/release/bioc/vignettes/ DESeq2/inst/doc/DESeq2.html) [30]. DElncRNAs were screened with |log2 fold change (LFC)| > 1.5 and false discovery rate (FDR, or adjusted P-value) < 0.05. DEmRNAs were identified with |LFC| > 2 and FDR < 0.01. The R packages ggplot2 (version 3.1.1) and pheatmap (version 1.0.12) were used to plot the volcano plots and heatmaps of DElncRNAs.
3. Survival analysis
The R package survival (version 2.44-1.1) was used to evaluate the prognostic of DElncRNAs in LADC patients from TCGA. The expression data were normalized by the R package DESeq2. The LADC samples were divided into two groups: high expression (normalized counts > median) and low expression (normalized counts < median) in terms of each DElncRNA. The Log-Rank tests of survival probability in the two groups were conducted with P-value < 0.05 considered to be significant. The top-3 significant DElncRNAs related to overall survival were selected for further investigation. The survival curves of the significant DElncRNAs were calculated with Kaplan-Meier (KM) method and visualized using the R package survminer (version 0.4.4).
4. PCR experimental validation
LADC and adjacent normal tissues were obtained from 22 patients who suffered from LADC and treated in 2015 in Xiangya Hospital, Central South University. All subjects had written consents, and the protocol was approved by the ethics committee of Xiangya Hospital (No. 201503303). All procedures conducted in the study were in accordance with the Ethics Standards Institutions Research Committee and the 1964 Helsinki Declaration and its subsequent amendments or similar ethical standards. All patients had no history of other malignancies and never received radiotherapy or chemotherapy. All tissue samples were immediately frozen in liquid nitrogen and stored at -80 °C. All tumors and matched normal tissues were confirmed by two pathologists. Total RNA was extracted using the Trizol Reagent kit (Invitrogen) and was reversely transcribed into cDNAs by using HiScript II Q Select RT SuperMix for qPCR (Vazyme). Real-time PCR was performed on the cDNA templates using specific primers (TSINGKE) and the GeneCopoeia BlazeTaq™ SYBR® Green qPCR mix (GeneCopoeia). The lncRNAs relative expression levels were calculated as a ratio normalized to U6 expression. Comparative quantification was calculated with the 2−ΔΔCt method. The primers sequences used in this study are listed. lnc-YARS2-5:F, 5'- ggtaccagaagcagcacct-3'; and R, 5'- aaaagaactcggccaagctc-3'. lnc-NPR3-2: F, 5'- aagcaagcatactcgtccct-3'; and R, 5'- gagccaagacgtagaggagg-3'. LINC02310: F, 5'-gaggaggtgctttgcttctc-3'; and R, 5'-atgaaccgagtcctggagtc-3'.
5. Clinical analysis
The clinical analyses of the top-3 significant DElncRNAs were conducted based on the expression levels obtained from the above PCR experiments and the clinical tissues of the 22 patients of LADC. Moreover, LINC02310 expression and related LADC clinical data from TCGA were combined to verify the clinical relevance. Fisher's accurate test method was used to calculate the P-value.
6. CCK-8 assay
Two human lung cancer cell lines (H1299, PC-9) were purchased from Chinese Academy of Sciences Cell Bank (Shanghai, China). The cell lines H1299 and PC-9 (1×103) transfected with pLCDH-LINC02310 (OE), negative control (NC), si-LINC02310 (SI) and SI-NC were transferred into 96 well plates. Next, 100 ul of CCK-8 solution (Dojindo Laboratories, Kumamoto, Japan) was added to each well plate. Then, 40 min of incubation at 37 °C was involved. Finally, the absorbance was measured at 450 nm by automatic microplate reader (Tecan, NANOQUANT, Swizerland) on days 0, 1, 2, 3 and 4, respectively. The experiment was repeated three times.
7. Colony formation assay
For the colony formation assays, the cell line H1299 (1×103) transfected with OE, NC, SI and SI-NC were transferred into 6-well plates and then incubated at 37°C for 10 days. Colonies were dyed with dyeing solution containing 0.1% crystal violet and 20% methanol. Cell colonies were then counted and analyzed.
8. CeRNA network construction
MiRNAs binding to LINC02310 were predicted via lncRNASNP2 website (http://bioinfo.life.hust.edu.cn/lncRNASNP#!/) [31, 32]. The interactions between these binding miRNAs and DEmRNAs were predicted according to starBase (version 3.0; http://starbase.sysu.edu.cn/) [33]. The predicted program TargetScan, microT, miRanda, miRmap, PITA, RNA22 and PicTar were integrated into starBase. The ceRNA network was constructed with LINC02310, binding miRNAs and selected downstream DEmRNAs involved and visualized by the software Cytoscape (version 3.7.1).
9. GO functional annotation and KEGG pathway analysis
The R packages clusterProfiler (version 3.10.1) [34] and org.Hs.eg.db (version 3.7.0) were used for statistical analysis and visualization of functional profiles for the predicted downstream DEmRNAs. The Gene Ontology (GO) functional annotation, including molecular function (MF), biological process (BP) and cellular component (CC) [35], and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were conducted with P-value < 0.05. The mRNA IDs were converted for the preparation of KEGG according to HGNC database (https://www.genenames.org/).
10. PPI network
Interactions of the downstream DEmRNAs were predicted and organized according to the STRING database (version 11.0; https://string-db.org/). The interactions were screened with combined score > 0.7 and then divided into different modules using MCODE (version 1.5, an app for Cytoscape, http://apps.cytoscape.org/ apps/mcode). The module of highest combined score was selected to construct the PPI network, which was visualized using the software Cytoscape.