The human H1299 lung cancer cells were purchased from Wuhan University (Hubei, China). The cells were maintained in high-glucose medium supplemented with 10% fetal bovine serum (Haoyang Biological Products Technology Co., Ltd.) and 1% penicillin/streptomycin (Beijing Solarbio Science & Technology Co., Ltd.), at 37°C in a humidified incubator with 5% CO₂.

Total RNA was extracted from the H1299 cells using TRIzol^® (Gibco; Thermo Fisher Scientific, Inc.), according to the manufacturer's protocol. A total of 4 ml medium, supplemented with dimethyl sulfoxide (Beijing Solarbio Science & Technology Co., Ltd.) was added to the cells in the control group for 24 h at 37°C, while 4 ml 10 µM daidzein (Shanxi Huike Plant Development Co., Ltd.) was added to the cells in the treatment group for 24 h at 37°C. The RNA was extracted from the cells in both groups after 24 h. The purity and concentration of the RNA was determined from optical density 260/280 ratio using a spectrophotometer (NanoDrop ND-1000; Thermo Fisher Scientific, Inc.). RNA integrity was determined by 1% formaldehyde denaturing gel electrophoresis.

cDNA labeled with fluorescent dyes (Cy5 and Cy3-dCTP) was produced using they Eberwine's linear RNA amplification method and subsequent enzymatic reaction, as previously described (20), which has also been improved using CapitalBio cRNA Amplification and Labeling kit (CapitalBio Technology, Inc.) to produce higher yields of labeled cDNA.

In detail, double-stranded cDNA (containing the T7 RNA polymerase promoter sequence) was synthesized from 1 µg total RNA using the CbcScript reverse transcriptase according to the manufacturer's instructions (CapitalBio Technology, Inc.), with T7 Oligo (dT) and T7 Oligo (dN). Following double-stranded cDNA synthesis using DNA polymerase and RNase H, the products were purified using a PCR NucleoSpin Extract II kit (Macherey-Nagel, GmbH and Co., KG) and eluted with 30 µl elution buffer. The eluted double-stranded cDNA products were then vacuum evaporated to 16 µl and reverse transcribed at 37°C for 14 h using a T7 Enzyme Mix. The amplified cRNA was purified using the RNA Clean-up kit (MN). The Klenow enzyme labeling method was used following RT using CbcScript II reverse transcriptase. Briefly, 2 µg amplified RNA was mixed with 4 µg random nanomer, denatured at 65°C for 5 min, then cooled on ice. Subsequently, 5 µl 4X first-strand buffer, 2 µl 0.1M DTT and 1.5 µl CbcScript II reverse transcriptase was added, then the samples were incubated at 25°C for 10 min, and 37°C for 90 min. The cDNA products were purified using a PCR NucleoSpin Extract II kit (Macherey-Nagel, GmbH and Co., KG), then vacuum evaporated to 14 µl. The cDNA was mixed with 4 µg random nanomer, heated to 95°C for 3 min, then snap cooled on ice for 5 min. Next, 5 µl Klenow buffer, dNTP and Cy5-dCTP or Cy3-dCTP (GE Healthcare) was added to a final concentration of 240 µM each for the dNTPs and 40 µM Cy-dCTP. A total of 1.2 µl Klenow enzyme was then added and the samples were incubated at 37°C for 90 min. The labeled cDNA was purified with a PCR NucleoSpin Extract II kit (Macherey-Nagel, GmbH and Co., KG) and resuspended in elution buffer. The labeled controls and test samples labeled with Cy5-dCTP and Cy3-dCTP were dissolved in 80 µl hybridization solution containing 3XSSC, 0.2% SDS, 5X Denhardt's solution and 25% formamide. The DNA in the hybridization solution was denatured at 95°C for 3 min prior to loading onto a microarray. The arrays were hybridized in an Agilent Hybridization Oven (Agilent Technologies, Inc.) overnight at a rotation speed of 20 rpm at 42°C then washed with two consecutive solutions (0.2% SDS and 2X SSC at 42°C for 5 min, then 0.2X SSC for 5 min at room temperature).

The lncRNA + mRNA array data was analyzed for data summarization, normalization and quality control using the GeneSpring software v13.0 (Agilent Technologies, Inc.). To select the differentially expressed lncRNAs and mRNAs fold-change (FC) ≥2 and ≤-2 and P<0.05 (unpaired t-test). The data was log₂ transformed and the median value was determined using the Adjust Data function in the CLUSTER v3.0 software (http://bonsai.hgc.jp/~mdehoon/software/cluster/software.htm), then further analyzed using hierarchical clustering with average linkage (21). Finally, tree visualization was performed using Java Treeview (Stanford University School of Medicine). Kyoto Encyclopedia of Genes and Genomes (KEGG) (https://www.genome.jp/kegg/) and Gene Ontology (GO) (http://www.geneongoloty.org/) analyses were performed to determine the function of differentially expressed mRNAs in biological pathways and GO terms (with a cut-off of P<0.05). After gene annotation function, STRING (Search Tool for the Retrieval of Interacting Genes/Proteins; http://string-db.org/) was used to identify the relationship between protein expression levels, in which a protein-protein interaction network was created.

Co-expression analysis is used to perform correlation analysis on the signal value trend of differential expressed lncRNA and mRNA in all samples (experimental group and control group) after comparison. Correlation, >0.99 or <-0.99 was used as the screening criteria, with P<0.05. The lncRNA/mRNA relationship pair, that met the aforementioned conditions was considered to have a co expression relationship. Transcription factor prediction (TFs_predict), which uses the Match-1.0 Public (http://gene-regulation.com/pub/programs.html%23match) transcription factor prediction tool, was used to predict the binding sites of TFs, 2,000 bp upstream and 500 bp downstream of the start site of each lncRNA based on the co-expression results, and network diagrams are created using Cytoscape (http://www.cytoscape.org/download.php) software.

Total RNA from the cell cultures was extracted using TRIzol^® (Invitrogen; Thermo Fisher Scientific, Inc.), reverse transcribed into cDNA using the EasyScript^® One-Step gDNA Removal and cDNA Synthesis SuperMix kit (TransGen Biotech Co., Ltd.) for 15 min at 42°C, 5 sec at 85°C, then stored at 4°C until further use. qPCR was performed on a CFX Connect Real-Time system (Bio-Rad Laboratories, Inc.) with TransStart^® Top Green qPCR SuperMix (TransGen Biotech Co., Ltd.), according to the manufacturer's protocol. GAPDH served as an internal control and was included in the same PCR reaction with differentially expressed lncRNAs and mRNAs for RT-qPCR. The primer sequences used for the qPCR of the lncRNAs and mRNAs are listed in Table I. Relative expression levels were quantified using the 2^−ΔΔCq method (22).

Statistical analysis was performed using SPSS 19.0 software (IBM Corp.). The data are presented as the mean ± SD. Differences between groups were assessed using a two tailed unpaired t-test. P<0.05 was considered to indicate a statistically significant difference.

The preliminary in vitro experiments showed that daidzein had the strongest inhibitory effect on the proliferation of H1299 cells (data not shown); therefore, these cells were used for further experiments. The soybean isoflavones were extracted, which are the active ingredients in legumes, and daidzein and its derivatives were synthesized and purified. The experimental flow chart of the present study is presented in Fig. 1.

The present study aimed to investigate the effect of daidzein on H1299 cells. Gene cluster analysis of the mRNAs and lncRNAs was performed following treatment with daidzein. The clustering map demonstrated similarities between the samples.

As presented in Fig. 2, a significant difference in both mRNAs and lncRNAs was observed between the experimental group of cells treated with daidzein and the control group of untreated with daidzein. Following the addition of daidzein, the expression level of eight lncRNAs was upregulated in the H1299 cells, while 111 lncRNAs were downregulated. In addition, five mRNAs were upregulated and 35 mRNAs were downregulated. In Fig. 2A and D, the top sample tree represents the similarity clustering relationship between the different samples. The top color block represents the expected grouping of the samples manually set prior to the cluster analysis, and the samples of the same color indicate that the experiment is expected to be a group. The red line in Fig. 2B and E, X2, is the threshold boundary line of the upregulated lncRNA/mRNAs, while the green line X (−2) is the threshold boundary line of the downregulated lncRNA/mRNAs, and the middle gray line is the fitted line of the overall expression amount. The equation in is the fitted line equation, and R represents the correlation coefficient between the two sets of samples.

Microarray analysis revealed that 40 mRNAs (5 upregulated and 35 downregulated) were significantly differentially expressed in the H1299 cell line treated with daidzein compared with that in the cells treated without daidzein (FC≥2.0; P<0.05). The top 20 significantly enriched KEGG pathway and disease terms were selected. The top three terms that were enriched for gene-affected diseases following daidzein treatment were: ‘Erythropoietic protoporphyrin’, ‘malignant pleural mesothelioma’ and ‘distal myopathy’. The addition of daidzein also affected the genes involved in diseases, including cancer of the lung and pleura, the results of which were consistent with the research model (Fig. 3A). KEGG pathway analysis revealed that ‘dilated cardiomyopathy’ (hsa05414), ‘hypertrophic cardiomyopathy’ (hsa05410), ‘long-term depression’ (hsa04730), ‘p53 signaling pathway’ (hsa04115) were included in the top 20 pathways (Fig. 3B).

As shown in Fig. 4, numerous differentially expressed mRNAs, compared with that to the background mRNAs, were enriched in biological processes, molecular function and cellular component categories, and the top 20 significantly enriched terms were selected from GO analysis. The results (Fig. 5) indicated that the most significantly enriched GO terms associated with the differentially expressed mRNAs were ‘detection of mechanical stimulus’ (biological processes; GO:0050982; P=0.0010488989), ‘condensed chromosome’ (cellular component; GO:0000793; P=0.0012408009), ‘α-actinin binding’ (molecular function; GO:0051393; P=0.0002475636). Several of the differentially expressed genes were localized in the cytoplasm and altered the expression of molecular functional genes associated with ‘binding’.

Co-expression analysis was performed to determine the association between lncRNA and mRNA pairs, with similar expression profiles, at the genomic level. The association between differentially expressed mRNAs and lncRNAs following the addition of daidzein is presented in Fig. S1A. Following the addition of daidzein, the co-expressed mRNAs with differentially expressed lncRNAs included: XLOC_I2_013457, LOC284825, CIQTNF3, FILP1, FANCC, SPATA4, FLJ41455, ANKRD12, SYCP2, RTKN2 and MTHFD2L.

The transcription factor was predicted (TFs_predict) using the Match-1.0 Public transcription factor prediction tool (Fig. S1B). Based on the co-expression results, the binding of transcription factors, 2,000 bp upstream and 500 bp downstream of each lncRNA initiation sites was predicted. The relevant transcription factors of differentially expressed lncRNAs, following the addition of daidzein to the H1299 cells, included Oct-1, HNF-1, HNF-4, Pax-4, COMP1, Pax-6, FOXD3 and FOXJ2.

A total of six differentially expressed lncRNAs and six mRNAs were randomly selected to validate the microarray data. RT-qPCR analysis was performed to determine the fold changes of selected lncRNAs and mRNAs, respectively (Fig. 6). The fold-change was positive when the expression was upregulated and negative when the expression was downregulated. Thus, the RT-qPCR results were consistent with the microarray data.

As presented in Fig. 7, the somatostatin C protein encoded by the insulin-like growth factor-1 (IGF1) gene interacted directly with the IGF1 receptor, IRS, AKT and caspase-9 proteins. Collectively, these results suggested that the IGF1 gene plays an important role in the development of tumors.