Abstract
Purpose: This study was implemented to investigate the associations between SNP in mature microRNA (miRNA) sequence and lung cancer prognosis and to verify the function of those SNP.
Experimental Design: Eight SNPs (rs3746444T>C in hsa-mir-499, rs4919510C>G in hsa-mir-608, rs13299349G>A in hsa-mir-3152, rs12220909G>C in hsa-mir-4293, rs2168518G>A in hsa-mir-4513, rs8078913T>C in hsa-mir-4520a, rs11237828T>C in hsa-mir-5579, and rs9295535T>C in hsa-mir-5689) were analyzed in a southern Chinese population with 576 patients with lung cancer, and the significant results were validated in two additional cohorts of 346 and 368 patients, respectively. A series of experiments were performed to evaluate the relevancies of those potentially functional SNPs.
Results: We found that the microRNA-499 rs3746444T>C polymorphism exhibited a consistently poor prognosis for patients with lung cancer in the discovery set [HR, 1.24; 95% confidence interval (CI), 1.02–1.49; P = 0.028], in the validation set I (HR, 1.31; 95% CI, 1.01–1.71; P = 0.048) and in the validation set II (HR, 1.45; 95% CI, 1.12–1.86; P = 0.004). The adverse effect of CT/CC variants was more remarkable in patients receiving platinum-based chemotherapy. Further functional assays demonstrated that the rs3746444C variant allele influences the expression of several cancer-related genes and affects lung cancer cells' proliferation and tumor growth in vivo and in vitro via the cisplatinum resistance.
Conclusion: Our findings suggested that the rs3746444T>C polymorphism in mature miR-499 sequence could contribute to poor prognosis by modulating cancer-related genes' expression and thus involve tumorigenesis and anti-chemotherapy, which may be a useful biomarker to predict lung cancer patients' prognosis. Clin Cancer Res; 21(7); 1602–13. ©2015 AACR.
Lung cancer remains a poor prognosis disease in recent decades, and novel, efficacious therapeutic biomarkers are needed. In the current study, we evaluated the associations of polymorphisms in mature microRNA sequence with lung cancer prognosis. We found that the microRNA-499 rs3746444T>C polymorphism contributed a harmful role on prognosis for patients with lung cancer. The adverse effect of the rs3746444CT/CC variant genotypes was more evident in patients receiving platinum-based chemotherapy. Functional experiments identified that the rs3746444C variant allele may influence the expression of several cancer-related genes and affect the lung cancer cells' proliferation and tumor growth in vivo and in vitro via the cisplatinum resistance. This finding would present an opportunity to increase our accuracy in forecasting therapeutic outcome in lung cancer, thereby helping refine therapeutic strategies in the treatment of lung cancer.
Introduction
Lung cancer is one of the most fatal human malignancies, which ranks the No. 1 incidence and cancer-related death among tumors in the world and China (1). Despite of much research effort in treatment for lung cancer in recent decades, the 5-year survival rate is still poor as less than 20% (2). Evidences revealed that patients with lung cancer with similar stages or histologic classifications have dramatically different responses to anticancer therapies and distinct survival outcomes, likely due to heterogeneity of gene/protein expression profiles (3). Established methods for predicting prognosis include the tumor, node, and metastasis (TNM) staging system; however, accumulating studies indicated that genetic biomarkers might play a vital role in cancer prognosis, according to their influences on cancer progression and treatment efficiency (4, 5). Therefore, seeking distinctive molecular biomarkers and genetic variants of the key genes involved tumor initiation and progression may promote improving survival outcomes for human cancers.
MicroRNAs (miRNA) are an abundant class of small noncoding, single-stranded RNAs of 21 to 24 nucleotides gene products, which has been conjectured miRNAs regulating the expression of approximately one third of human genes (6, 7), by distinctive mechanisms of perfect or imperfect base pairing with target mRNAs at the 3′-untranslated region(UTR) or protein-coding sequences(CDS), leading to mRNA cleavage or translational repression (8, 9). The biogenesis of miRNAs is a complex process and miRNAs are initially transcribed into a long noncoding RNA known as the primary miRNA (pri-miRNA) that are then processed in the nucleus into about 70-nt miRNA precursor (pre-miRNA) with hairpin-shaped. Pre-miRNAs are further processed into 21 to 24 nt mature miRNAs (10–12). Emerging evidences have indicated that miRNAs are involved in a wide diversity of biologic processes, including cell-cycle regulation, differentiation, proliferation, development, and apoptosis (13–15). Furthermore, miRNAs have been extensively associated with the etiology and clinical outcome of human malignancies, which influence tumorigenesis through their regulation of specific proto-oncogenes and tumor suppressor genes (6, 16, 17).
Genetic variants such as SNPs in miRNAs may affect the transcription of miRNA primary transcripts, the processing of miRNA precursors to mature miRNAs or miRNA–target interactions, resulting in multifarious functional consequences (12, 18). Recently, several reports have elucidated SNPs in mature miRNAs sequence to be associated with various types of malignancy and their survival (19–21). Nonetheless, few studies have well characterized the associations between the polymorphisms in the mature miRNAs and lung cancer prognosis. In this study, we studied the common [minor allele frequency (MAF) > 0.05] SNPs in mature miRNAs sequence and hypothesized that these SNPs were associated with the prognosis of lung cancer.
On the basis of 3 independent cohorts of 1,290 patients with lung cancer conducted in southern and eastern Chinese populations, we genotyped 8 common polymorphisms in 8 mature miRNAs sequence (rs3746444T>C in hsa-mir-499, rs4919510C>G in hsa-mir-608, rs13299349G>A in hsa-mir-3152, rs12220909G>C in hsa-mir-4293, rs2168518G>A in hsa-mir-4513, rs8078913T>C in hsa-mir-4520a, rs11237828T>C in hsa-mir-5579, and rs9295535T>C in hsa-mir-5689) and analyzed their associations with lung cancer prognosis. Further biochemical assays were performed to identify the biologic effects of those promising polymorphisms.
Materials and Methods
Study subjects and follow-up
All patients with histologically confirmed primary lung cancer including 1,056 patients in the discovery set, 503 cases in the validation set I, and 773 patients in the validation set II were consecutively recruited from Guangzhou City in southern China and Jiangsu Province in eastern China between March 2007 and June 2011 and were followed-up until December 2013 as previously described (22, 23). All participants were genetically unrelated ethnic Han Chinese and none had blood transfusions in the last 6 months. Having given a written informed consent, each participant was scheduled for an interview with a structured questionnaire to provide data on smoking status, alcohol use, and other factors including family history of cancer described previously (24, 25). We also reviewed the patients' medical records to collect clinical information including the date of diagnosis, histologic grade, pathologic stage, surgery, chemotherapy regimens, and radiotherapy status. Follow-up was performed every 3 months by telephone call from the time of enrollment until death or the last scheduled follow-up. Survival time was calculated from the date when patients first received confirmed diagnoses until the date of the last follow-up or death, and dates of death were obtained from medical records or from patients' families through telephone follow-up. Those patients who were lost to follow-up or had no accurate data on chemotherapy regimens were excluded. Finally, 576 patients from the discovery set, 346 patients from the validation set I, and 368 patients from the validation set II completed the follow-up and had intact survival data were included in this study. Moreover, there were no significant differences in clinical information, as well as survival data, between the included and excluded groups. The study was approved by the institutional review boards of Guangzhou Medical University (Ethics Committee of Guangzhou medical university: GZMC2007-07-0676) and Soochow University (Ethics Committee of Soochow University: SZUM2008031233).
SNP selection and genotyping
On the basis of the public database miRBase (http://www.mirbase.org, access to 1/4/2013), we performed bioinformatics analysis to search SNPs in miRNAs' mature sequences, with the strategy of blasting 1,527 miRNAs of their pre-microRNA sequences with the dbSNP database (http://www.ncbi.nlm.nih.gov/snp, access to 1/4/2013) and identified 218 SNPs located in mature miRNA sequences. We then found that of these SNPs there are only 8 common (MAF > 0.05) SNPs (i.e., rs3746444T>C in hsa-mir-499, rs4919510C>G in hsa-mir-608, rs13299349G>A in hsa-mir-3152, rs12220909G>C in hsa-mir-4293, rs2168518G>A in hsa-mir-4513, rs8078913T>C in hsa-mir-4520a, rs11237828 T>C in hsa-mir-5579, and rs9295535T>C in hsa-mir-5689) in Chinese populations (Supplementary Table S1). Therefore, we analyzed these SNPs in our study.
Genomic DNA was extracted from the peripheral blood lymphocytes of all the study subjects by using the DNA Blood Mini Kit (Qiagen), according to the manufacturer's protocol. The selected SNPs were genotyped using the TaqMan allelic discrimination assay on the ABI PRISM 7500 Sequence Detection Systems (Applied Biosystems). Primers and probes are described in Supplementary Table S2, which were designed by the Primer Express 3.0 (Applied Biosystems) and synthesized by the Shanghai GeneCore Biotechnologies. We further randomly selected 10% samples for each of the selected SNPs for regenotyping, and the results were 100% concordant (Supplementary Fig. S1).
Cell culture
A549 human lung adenocarcinoma cells and 293T human embryonic kidney cell lines were purchased from the Cell Bank of Type Culture Collection of the Chinese Academy of Sciences, Shanghai Institute of Cell Biology. A549 and 293T cells were cultured at 37°C in 90% humidity and 5% CO2 in DMEM (Gibco-BRL) with 10% FBS (Gibco-BRL) and penicillin (100 IU/mL)/streptomycin (100 mg/mL) in a Steri-Cult incubator (Thermo Fisher Scientific).
Plasmids construction, lentivirus package, and cell transfection
The lentiviral vector pLVX-IRES-neo (Clonetech Laboratories Inc.) was used to construct the pLV-miR-499T or pLV-miR-499C plasmid. A fragment of containing rs3746444T allele from the precursor sequence of miRNA-499 was amplified with the forward primer 5′-GACTCGAGCTTCACTTCCCTGCCAAATCC-3′ and reverse primer 5′-TCGCGGC CGCGCCCACAGAGCGACATTCC-3′ from a homozygous human genomic DNA sample. The amplified products were digested by XhoI and NotI (Fermentas) and then cloned into the lentiviral expression vector pLVX-IRES-neo. Mutation of miR-499 rs3746444C was induced by site-directed mutagenesis using the Quick Change XL site-directed mutagenesis kit (Stratagene). The resulting constructs were verified by direct sequencing. Replication-defective VSV-G pseudotyped viral particles were packaged in LentiX 293T human embryonic kidney cells (Clonetech) by using a 3-plasmid transient cotransfection method (Lenti-T HT packaging mix, Clonetech). Viruses were then harvested, cleared by centrifugation at 3,000 rpm at room temperature for 10 minutes, and passed through a 0.45-μm syringe filter after 48 hours. For transfection, A549 cells were infected with control lentivirus (an “empty” vector without the miR-499 fragment inserted), miRNA-499-T-allele lentivirus, and miRNA-499-C-allele lentivirus, respectively. After 48 hours of transduction, the cells were stably selected with G418 at 100 μg/mL (Gibco), and the drug-resistant cells were used for subsequent studies.
Cell proliferation assay
The MTT assay was performed to assess the effect of the rs3746444T>C polymorphism on cell viability via cisplatinum treatment. Briefly, the A549 cells infected with different allele lentivirus vectors or an “empty” lentivirus vectors were seeded at a density of 5 × 104 cells per well in a volume of 100 μL in 96-well microtiter plates for 24 hours. The cells were then incubated with a series of concentrations of cisplatinum, and cells treated with RPMI-1640 medium alone were used as controls. After incubation for another 24 or 48 hours, 20 μL MTT solutions (5 mg/mL, Sigma) were added into each well at 37°C in the dark for at least 4 hours. After that, the supernatant was removed and 150 μL DMSO was added into each well to dissolve the formazan crystals. The absorbance was then measured at 490 nm using a Plate Reader (Bio-Tec Instruments, Inc.). The MTT assay was also performed for above cells transfected with the 3 vectors every other day during a week without cisplatinum treatment to reveal the effect of the rs3746444T>C polymorphism on cell proliferation independent of cisplatinum. The cell proliferation rate was calculated as Atreated cells/Acontrol cells × 100%. All these experiments were repeated in triplicate.
Flow cytometry analysis of cell apoptosis
To quantify the effect on cell apoptosis of the rs3746444T>C polymorphism, the Annexin v-FITC (V-FITC)/propidium iodide (PI) double staining assay was conducted according to the manufacturer's instructions. Briefly, A549 cells infected with the different rs3746444T>C allele lentivirus vectors and an “empty” lentivirus vector were cultured for 48 hours without and with 200 μmol/L cisplatinum treatment. The cells were then harvested and washed twice with cold PBS and resuspended in 500 μL binding buffer at a concentration of 1 × 106 cells/mL. Then, 5 μL Annexin V-FITC solution and 5 μL PI (1 mg/mL) were added. The cells were incubated at room temperature for 15 minutes in the dark and analyzed by the flow cytometry within 1 hour. The number of apoptotic cells was counted and presented as a percentage of the total cell count.
Soft-agar colony formation assay
Colony formation ability was determined by the anchorage-independent soft agar assay on A549 cells treated with lentivirus vectors. A layer of agar containing 5 mL of 0.75% low melting agar (Bio-Rad) dissolved in growth media DMEM with 10% FBS was poured into per 60-mm culture dish and allowed to set at 4°C for 20 minutes. A second layer of 3 mL containing 0.35% agar dissolved in the same growth media containing A549 cells infected with different allele lentivirus (3 × 103 cells/mL) was added on top of the first layer. Two milliliters growth media containing different concentrations of cisplatinum (0, 100, and 200 μmol/L) were added on top of the second layer, and the cells were incubated in a humidified 5% CO2 incubator at 37°C for 14 days. During cell culture period, the medium accompanied with above referred concentrations of cisplatinum was changed every 3 days. At the end of the experiment, the colonies were stained with 0.04% crystal violet–2% ethanol in PBS. The colonies were photographed and colonies with at least one diameter of 100 μm within photographic fields were chosen. Independent triplicate experiments were done for each treated A549 cells.
Tumor xenografts and treatment
All experiments and procedures involving animals were conducted in accordance with guidelines approved by the Laboratory Animal Center of Guangzhou Medical University. Female nude mice (BALB/c), aged 4–5 weeks, were purchased from the Shanghai Laboratory Animal Center of Chinese Academy of Sciences. A549-miRNA-499-T, A549-miRNA-499-C, and A549-miRNA-499 empty cells were diluted to a concentration of 5 × 107/mL in physiologic saline. Nude mice were subcutaneously injected with 0.1 mL of the suspension into the dorsal flank (6 mice/group). When the tumors had reached the volume of approximately 200 mm3, animals were treated for 4 weeks with concentrations of cisplatinum (0, 0.5, and 5 mg/kg in 0.2 mL of normal saline per day, intraperitoneally; Sigma). The tumor dimensions was measured 3 days per time by a Vernier caliper along 2 perpendicular axes and calculated according to the following formula: volume = 1/2 × length × width2.
RNA extraction and microarray analysis
Total RNA was extracted from the cultured cells in both experimental groups (A549-miRNA-499-T and A549-miRNA-499-C) using RNeasy Mini Kits (Qiagen) following the manufacturer's protocol. RNA quantity and quality were assessed using a NanoDrop 2000 spectrophotometer (Thermo Scientific). The expression profiling was performed using human HT-12 Expression BeadChip arrays from Illumina (Illumina, Inc.) containing more than 34,600 unique polynucleotide probes. Microarray hybridization procedure was conducted according to the manufacturer guidelines. Arrays were scanned using the Illumina BeadStation Scanner to obtain gene expression levels and the data were processed using the Illumina GenomeStudio software, transformed by variance stabilization transformation (VST; ref. 26) and normalized by quantile normalization. Illumina custom model was used to detect differentially expressed changes among the study groups. To handle potential platform-related biases, the data were repeatedly calculated and triplicate samples were used for each group.
Quantitative real-time PCR analysis
To verify the microarray findings, we further evaluated differentially expressed genes by the quantitative real-time PCR (qRT-PCR) as described previously (24). Briefly, Total RNA was isolated from A549 processed with different lentivirus and reverse transcribed into cDNA following the manufacturer's protocols. Real-time PCR was performed in the ABI Prism 7500 sequence detection system (Applied Biosystems) with using the SYBR Premix Ex Taq (Perfect Real Time). The relative expression level of the target gene compared with β-actin used as an internal reference gene in each reaction was calculated by the 2−ΔCt method (27). The relative expression of precursor miRNA-499 in A549 cells was also quantified using SYBR-Green method and was calculated relative to the U6 small nuclear RNA. Each assay was performed in triplicate. All the primers used for PCR amplification are listed in Supplementary Table S3.
Statistical analysis
The χ2 test was applied separately to compare the distribution of selected demographic and clinical variables according to death status. The associations between genotypes and overall survival time were estimated using the Kaplan–Meier method and log-rank test. The Cox proportional hazards regression models with or without adjusted for confounders were used to evaluate the effect of polymorphisms on lung cancer survival. The predictive value of the SNPs or other factors was evaluated by time-dependent receiver–operator characteristic (ROC) curves for censored data and calculated the area under the curve (AUC) of the ROC curves (28). We calculated the AUC values for different scores by plotting (t, AUC [t]) for different values of follow-up time (t). The heterogeneity between subgroups was assessed with the χ2-based Q test. Multiplicative interactions were assessed by Cox regression. The Student t test was used to compare the difference in levels of target genes expression. The difference colonies number levels were compared by using one-way ANOVA test with a Bonferroni post hoc comparison. Repeated measure ANOVA test was performed to analyze the mean tumor volume of different group, as well as the different transfectants proliferation. All tests were 2-sided by using the SAS software (version 9.3; SAS Institute) and R software (version 3.0.2; The R Foundation for Statistical Computing). P < 0.05 was considered statistically significant.
Results
Patients' characteristics
The distributions of demographic variables and clinical information of the 3 datasets are described in Table 1. We found that factors, including age, smoking status, surgery, chemotherapy, radiotherapy and stages, had consistently significant influences on patient prognosis in the 3 datasets, and these factors were further adjusted for in the multivariate Cox regression to control possible confounding on the main effects of the studied microRNAs' polymorphisms on lung cancer prognosis.
Baseline demographic and clinical characteristics of study populations
. | Discovery set (Southern Chinese) . | Validation set I (Eastern Chinese) . | Validation set II (Southern Chinese) . | MST (months) . | Log-rank P . | |||
---|---|---|---|---|---|---|---|---|
Variables . | Cases, n (%) . | Death, n . | Cases, n (%) . | Death, n . | Cases, n (%) . | Death, n . | DS/VS.I/VS.II . | DS/VS.I/VS.II . |
Total | 576 | 512 | 346 | 266 | 368 | 298 | ||
Age, y | 0.045/0.006/0.018 | |||||||
≤60 | 264 (45.8) | 236 | 175 (50.6) | 128 | 189 (51.4) | 147 | 15/18/17 | |
>60 | 312 (54.2) | 276 | 171 (49.4) | 138 | 179 (48.6) | 151 | 11/12/11 | |
Sex | 0.056/0.524/0.437 | |||||||
Male | 415 (72.1) | 369 | 234 (67.6) | 183 | 255 (69.3) | 211 | 12/15/14 | |
Female | 161 (27.9) | 143 | 112 (32.4) | 83 | 113 (30.7) | 87 | 16/16/16 | |
Family history of cancer | 0.875/0.864/0.297 | |||||||
No | 520 (90.3) | 463 | 331 (95.7) | 253 | 354 (96.2) | 288 | 12/15/15 | |
Yes | 56 (9.7) | 49 | 15 (4.3) | 13 | 14 (3.8) | 10 | 13/15/20 | |
Family history of lung cancer | 0.694/0.717/0.061 | |||||||
No | 553 (96.0) | 492 | 340 (98.3) | 260 | 359 (97.5) | 293 | 12/15/14 | |
Yes | 23 (4.0) | 20 | 6 (1.7) | 6 | 9 (2.5) | 5 | 19/27/29 | |
Smoking status | 0.039/0.015/0.003 | |||||||
Never | 236 (41.0) | 213 | 196 (56.6) | 150 | 198 (53.8) | 156 | 15/17/16 | |
Ever | 340 (59.0) | 299 | 150 (43.4) | 116 | 170 (46.2) | 142 | 12/13/12 | |
Drinking status | 0.478/0.973/0.915 | |||||||
Never | 455 (79.0) | 407 | 300 (86.7) | 231 | 317 (86.1) | 258 | 13/15/15 | |
Ever | 121 (21.0) | 105 | 46 (13.3) | 35 | 51 (13.9) | 40 | 11/14/14 | |
Surgery | ||||||||
No | 358 (62.2) | 337 | 192 (55.5) | 159 | 213 (57.9) | 186 | 11/13/12 | 8.31 × 10−11/5.09 × 10−5/7.26 × 10−5 |
Yes | 218 (37.8) | 175 | 154 (44.5) | 107 | 155 (42.1) | 112 | 20/19/18 | |
Chemotherapy | 0.012/0.001/0.002 | |||||||
No | 218 (37.8) | 196 | 100 (28.9) | 78 | 96 (26.1) | 78 | 9/8/7 | |
Yes | 358 (62.2) | 316 | 246 (71.1) | 188 | 272 (73.9) | 220 | 15/17/17 | |
Chemotherapy regimens | 0.010/0.002/0.008 | |||||||
No | 218 (37.8) | 196 | 100 (28.9) | 78 | 96 (26.1) | 78 | 9/8/7 | |
Platinum-based treatmenta | 331 (57.5) | 292 | 210 (60.7) | 163 | 239 (64.9) | 192 | 15/17/17 | |
Other treatmentb | 27 (4.7) | 24 | 36 (10.4) | 25 | 33 (9.0) | 28 | 13/18/16 | |
Radiotherapy | 0.003/0.044/0.035 | |||||||
No | 304 (52.8) | 269 | 143 (41.3) | 106 | 167 (45.4) | 138 | 10/10/11/ | |
Yes | 272 (47.2) | 243 | 203 (58.7) | 160 | 201 (54.6) | 160 | 15/17/17 | |
Stages | 0.004/0.002/0.001 | |||||||
I + II | 100 (17.3) | 77 | 65 (18.8) | 45 | 89 (24.2) | 61 | 16/18/17 | |
III | 193 (33.7) | 176 | 108 (31.2) | 85 | 108 (29.4) | 96 | 12/15/14 | |
IV | 282 (48.9) | 259 | 173 (40.0) | 136 | 171 (46.4) | 141 | 11/12/12 | |
Histologic types | 0.066/0.941/0.948 | |||||||
Adenocarcinoma | 205 (35.6) | 182 | 162 (46.8) | 124 | 170 (46.2) | 138 | 13/15/14 | |
Squamous cell carcinoma | 189 (32.8) | 161 | 108 (31.2) | 83 | 115 (31.3) | 91 | 13/15/15 | |
Large cell carcinoma | 25 (4.3) | 24 | 17 (4.9) | 13 | 18 (4.9) | 14 | 14/10/11 | |
Small cell lung cancer | 71 (12.3) | 65 | 44 (12.7) | 34 | 46 (12.5) | 39 | 10/18/14 | |
Otherc | 86 (14.9) | 80 | 15 (4.3) | 12 | 19 (5.2) | 16 | 13/17/18 |
. | Discovery set (Southern Chinese) . | Validation set I (Eastern Chinese) . | Validation set II (Southern Chinese) . | MST (months) . | Log-rank P . | |||
---|---|---|---|---|---|---|---|---|
Variables . | Cases, n (%) . | Death, n . | Cases, n (%) . | Death, n . | Cases, n (%) . | Death, n . | DS/VS.I/VS.II . | DS/VS.I/VS.II . |
Total | 576 | 512 | 346 | 266 | 368 | 298 | ||
Age, y | 0.045/0.006/0.018 | |||||||
≤60 | 264 (45.8) | 236 | 175 (50.6) | 128 | 189 (51.4) | 147 | 15/18/17 | |
>60 | 312 (54.2) | 276 | 171 (49.4) | 138 | 179 (48.6) | 151 | 11/12/11 | |
Sex | 0.056/0.524/0.437 | |||||||
Male | 415 (72.1) | 369 | 234 (67.6) | 183 | 255 (69.3) | 211 | 12/15/14 | |
Female | 161 (27.9) | 143 | 112 (32.4) | 83 | 113 (30.7) | 87 | 16/16/16 | |
Family history of cancer | 0.875/0.864/0.297 | |||||||
No | 520 (90.3) | 463 | 331 (95.7) | 253 | 354 (96.2) | 288 | 12/15/15 | |
Yes | 56 (9.7) | 49 | 15 (4.3) | 13 | 14 (3.8) | 10 | 13/15/20 | |
Family history of lung cancer | 0.694/0.717/0.061 | |||||||
No | 553 (96.0) | 492 | 340 (98.3) | 260 | 359 (97.5) | 293 | 12/15/14 | |
Yes | 23 (4.0) | 20 | 6 (1.7) | 6 | 9 (2.5) | 5 | 19/27/29 | |
Smoking status | 0.039/0.015/0.003 | |||||||
Never | 236 (41.0) | 213 | 196 (56.6) | 150 | 198 (53.8) | 156 | 15/17/16 | |
Ever | 340 (59.0) | 299 | 150 (43.4) | 116 | 170 (46.2) | 142 | 12/13/12 | |
Drinking status | 0.478/0.973/0.915 | |||||||
Never | 455 (79.0) | 407 | 300 (86.7) | 231 | 317 (86.1) | 258 | 13/15/15 | |
Ever | 121 (21.0) | 105 | 46 (13.3) | 35 | 51 (13.9) | 40 | 11/14/14 | |
Surgery | ||||||||
No | 358 (62.2) | 337 | 192 (55.5) | 159 | 213 (57.9) | 186 | 11/13/12 | 8.31 × 10−11/5.09 × 10−5/7.26 × 10−5 |
Yes | 218 (37.8) | 175 | 154 (44.5) | 107 | 155 (42.1) | 112 | 20/19/18 | |
Chemotherapy | 0.012/0.001/0.002 | |||||||
No | 218 (37.8) | 196 | 100 (28.9) | 78 | 96 (26.1) | 78 | 9/8/7 | |
Yes | 358 (62.2) | 316 | 246 (71.1) | 188 | 272 (73.9) | 220 | 15/17/17 | |
Chemotherapy regimens | 0.010/0.002/0.008 | |||||||
No | 218 (37.8) | 196 | 100 (28.9) | 78 | 96 (26.1) | 78 | 9/8/7 | |
Platinum-based treatmenta | 331 (57.5) | 292 | 210 (60.7) | 163 | 239 (64.9) | 192 | 15/17/17 | |
Other treatmentb | 27 (4.7) | 24 | 36 (10.4) | 25 | 33 (9.0) | 28 | 13/18/16 | |
Radiotherapy | 0.003/0.044/0.035 | |||||||
No | 304 (52.8) | 269 | 143 (41.3) | 106 | 167 (45.4) | 138 | 10/10/11/ | |
Yes | 272 (47.2) | 243 | 203 (58.7) | 160 | 201 (54.6) | 160 | 15/17/17 | |
Stages | 0.004/0.002/0.001 | |||||||
I + II | 100 (17.3) | 77 | 65 (18.8) | 45 | 89 (24.2) | 61 | 16/18/17 | |
III | 193 (33.7) | 176 | 108 (31.2) | 85 | 108 (29.4) | 96 | 12/15/14 | |
IV | 282 (48.9) | 259 | 173 (40.0) | 136 | 171 (46.4) | 141 | 11/12/12 | |
Histologic types | 0.066/0.941/0.948 | |||||||
Adenocarcinoma | 205 (35.6) | 182 | 162 (46.8) | 124 | 170 (46.2) | 138 | 13/15/14 | |
Squamous cell carcinoma | 189 (32.8) | 161 | 108 (31.2) | 83 | 115 (31.3) | 91 | 13/15/15 | |
Large cell carcinoma | 25 (4.3) | 24 | 17 (4.9) | 13 | 18 (4.9) | 14 | 14/10/11 | |
Small cell lung cancer | 71 (12.3) | 65 | 44 (12.7) | 34 | 46 (12.5) | 39 | 10/18/14 | |
Otherc | 86 (14.9) | 80 | 15 (4.3) | 12 | 19 (5.2) | 16 | 13/17/18 |
NOTE: Bold type: statistically significant, P < 0.05.
Abbreviations: DS, discovery set; VS I, validation set I; VS II, validation set II.
aPatients receiving cisplatin plus gemcitabine or plus vinorelbine or plus docetaxel or plus etoposide or a combination of the above regimens.
bPatients not receiving cisplatin but with other chemotherapeutics.
cMixed-cell or undifferentiated carcinoma.
Association between the mature microRNAs' polymorphisms and lung cancer survival
The associations between the polymorphisms in mature microRNAs and prognosis of lung cancer are shown in Table 2. In the discovery set, we found that patients who carried the rs3746444CC variant genotype had a significantly shorter median survival time (MST; CC vs. TT, 9 vs. 13 months; log-rank test: P = 0.032) and had an increased death risk [adjusted HR, 1.54; 95% confidence interval (CI), 1.04–2.29; P = 0.033] than those with the rs3746444TT genotype, but patients carried the rs3746444CT genotype did not. The Cox model further showed that the SNP conferred a poor survival of lung cancer under the additive genetic model (CC vs. CT vs. TT: HR, 1.22; 95% CI, 1.04–1.42; P = 0.012). Consistently, the rs3746444C (CT + CC) variants were associated with poor survival (HR, 1.24; 95% CI, 1.02–1.49; P = 0.028) and a shorter MST (12 vs. 13 months; log-rank test: P = 0.029; Fig. 1A) than the rs3746444TT genotype for the patients with lung cancer. However, no significant associations of other SNPs with lung cancer survival were observed.
Kaplan–Meier (KM) curves and AUC for survival for patients with lung cancer carrying the different miR-499 rs3746444T>C genotypes. A, KM curves of patients from the discovery set. B, KM curves of patients from the validation set I. C, KM curves of patients from the validation set II. D, AUCs.
Kaplan–Meier (KM) curves and AUC for survival for patients with lung cancer carrying the different miR-499 rs3746444T>C genotypes. A, KM curves of patients from the discovery set. B, KM curves of patients from the validation set I. C, KM curves of patients from the validation set II. D, AUCs.
Associations between microRNA polymorphisms and lung cancer survival
Gene SNP . | Lung cancer patients, n (%) . | Deaths, n . | MST, mo . | Log-rank P . | Cox model, crude HR (95% CI) . | Cox model, adjusted HR (95% CI)a . |
---|---|---|---|---|---|---|
Discovery set | 576 | 512 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 389 (67.5) | 342 | 13 | 0.032 | 1.00 (ref.) | 1.00 (ref.) |
CT | 152 (26.4) | 143 | 12 | 1.17 (0.96–1.42) | 1.19 (0.98–1.45) | |
CC | 35 (6.1) | 27 | 9 | 1.55 (1.05–2.31) | 1.54 (1.04–2.29) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.21 (1.04–1.41) | 1.22 (1.04–1.42) | ||||
Dominant model | ||||||
TT | 389 (67.5) | 342 | 13 | 0.029 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 187 (32.5) | 170 | 12 | 1.21 (1.01–1.47) | 1.24 (1.02–1.49) | |
hsa-mir-608: rs4919510C>G | ||||||
CC | 179 (31.1) | 160 | 9 | 0.115 | 1.00 (ref.) | 1.00 (ref.) |
GC | 250 (43.4) | 223 | 14 | 0.80 (0.65–0.99) | 0.91 (0.73–1.12) | |
GG | 147 (25.5) | 129 | 13 | 0.84 (0.67–1.06) | 0.89 (0.70–1.12) | |
has-mir-3152: rs13299349G>A | ||||||
GG | 440 (76.4) | 397 | 12 | 0.122 | 1.00 (ref.) | 1.00 (ref.) |
AG | 115 (20.0) | 96 | 12 | 0.88 (0.70–1.10) | 0.92 (0.74–1.16) | |
AA | 21 (3.6) | 19 | 26 | 0.65 (0.42–1.05) | 0.88 (0.56–1.42) | |
hsa-mir-4293: rs12220909G>C | ||||||
GG | 321 (55.7) | 287 | 13 | 0.853 | 1.00 (ref.) | 1.00 (ref.) |
CG | 209 (36.3) | 186 | 12 | 1.02 (0.85–1.23) | 0.96 (0.80–1.16) | |
CC | 46 (8.0) | 39 | 12 | 0.93 (0.67–1.30) | 1.01 (0.72–1.42) | |
hsa-mir-4513: rs2168518G>A | ||||||
GG | 329 (57.1) | 296 | 13 | 0.305 | 1.00 (ref.) | 1.00 (ref.) |
AG | 220 (38.2) | 193 | 12 | 0.92 (0.77–1.10) | 0.92 (0.76–1.10) | |
AA | 27 (4.7) | 23 | 27 | 0.76 (0.49–1.16) | 0.77 (0.50–1.18) | |
hsa-mir-4520a: rs8078913T>C | ||||||
TT | 259 (45.0) | 235 | 12 | 0.164 | 1.00 (ref.) | 1.00 (ref.) |
CT | 220 (38.2) | 193 | 14 | 0.89 (0.74–1.08) | 0.90 (0.74–1.09) | |
CC | 97 (16.8) | 84 | 12 | 0.80 (0.62–1.03) | 0.79 (0.62–1.02) | |
hsa-mir-5579: rs11237828T>C | ||||||
TT | 240 (41.7) | 210 | 14 | 0.182 | 1.00 (ref.) | 1.00 (ref.) |
CT | 278 (48.3) | 249 | 12 | 1.18 (0.98–1.42) | 1.16 (0.96–1.40) | |
CC | 58 (10.1) | 53 | 13 | 1.08 (0.80–1.45) | 1.15 (0.84–1.56) | |
hsa-mir-5689: rs9295535T>C | ||||||
TT | 273 (47.4) | 253 | 12 | 0.062 | 1.00 (ref.) | 1.00 (ref.) |
CT | 258 (44.8) | 222 | 14 | 0.83 (0.69–1.01) | 0.89 (0.74–1.07) | |
CC | 45 (7.8) | 37 | 13 | 0.75 (0.53–1.07) | 0.77 (0.55–1.10) | |
Validation set I | ||||||
Total subjects | 346 | 266 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 243 (70.2) | 180 | 17 | 0.001 | 1.00 (ref.) | 1.00 (ref.) |
CT | 82 (23.7) | 68 | 13 | 1.28 (0.97–1.70) | 1.18 (0.88–1.57) | |
CC | 21 (6.1) | 18 | 6 | 2.43 (1.49–3.98) | 2.20 (1.33–3.64) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.23 (1.16–1.76) | 1.34 (1.08–1.65) | ||||
Dominant model | ||||||
TT | 243 (70.2) | 180 | 17 | 0.006 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 103 (29.8) | 86 | 12 | 1.42 (1.10–1.84) | 1.31 (1.01–1.71) | |
Validation set II | ||||||
Total no. of subjects | 368 | 298 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 257 (69.8) | 203 | 16 | 2.20 × 10−6 | 1.00 (ref.) | 1.00 (ref.) |
CT | 93 (25.3) | 77 | 12 | 1.33 (1.02–1.73) | 1.33 (1.01–1.75) | |
CC | 18 (4.9) | 18 | 5 | 2.87 (1.76–4.68) | 2.29 (1.40–3.76) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.50 (1.23–1.84) | 1.43 (1.16–1.75) | ||||
Dominant model | ||||||
TT | 257 (69.8) | 203 | 16 | 0.001 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 111 (30.2) | 95 | 11 | 1.47 (1.15–1.89) | 1.45 (1.12–1.86) | |
Merged set | ||||||
Total subjects | 1290 | 1076 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 889 (68.9) | 725 | 15 | 1.15 × 10−9 | 1.00 (ref.) | 1.00 (ref.) |
CT | 327 (25.4) | 288 | 12 | 1.23 (1.08–1.41) | 1.19 (1.04–1.37) | |
CC | 74 (5.7) | 63 | 7 | 2.05 (1.58–2.66) | 1.86 (1.44–2.42) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.33 (1.20–1.48) | 1.28 (1.15–1.42) | ||||
Dominant model | ||||||
TT | 889 (68.9) | 725 | 15 | 7.54 × 10−6 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 401 (31.1) | 351 | 11 | 1.33 (1.17–1.51) | 1.27 (1.12–1.45) |
Gene SNP . | Lung cancer patients, n (%) . | Deaths, n . | MST, mo . | Log-rank P . | Cox model, crude HR (95% CI) . | Cox model, adjusted HR (95% CI)a . |
---|---|---|---|---|---|---|
Discovery set | 576 | 512 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 389 (67.5) | 342 | 13 | 0.032 | 1.00 (ref.) | 1.00 (ref.) |
CT | 152 (26.4) | 143 | 12 | 1.17 (0.96–1.42) | 1.19 (0.98–1.45) | |
CC | 35 (6.1) | 27 | 9 | 1.55 (1.05–2.31) | 1.54 (1.04–2.29) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.21 (1.04–1.41) | 1.22 (1.04–1.42) | ||||
Dominant model | ||||||
TT | 389 (67.5) | 342 | 13 | 0.029 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 187 (32.5) | 170 | 12 | 1.21 (1.01–1.47) | 1.24 (1.02–1.49) | |
hsa-mir-608: rs4919510C>G | ||||||
CC | 179 (31.1) | 160 | 9 | 0.115 | 1.00 (ref.) | 1.00 (ref.) |
GC | 250 (43.4) | 223 | 14 | 0.80 (0.65–0.99) | 0.91 (0.73–1.12) | |
GG | 147 (25.5) | 129 | 13 | 0.84 (0.67–1.06) | 0.89 (0.70–1.12) | |
has-mir-3152: rs13299349G>A | ||||||
GG | 440 (76.4) | 397 | 12 | 0.122 | 1.00 (ref.) | 1.00 (ref.) |
AG | 115 (20.0) | 96 | 12 | 0.88 (0.70–1.10) | 0.92 (0.74–1.16) | |
AA | 21 (3.6) | 19 | 26 | 0.65 (0.42–1.05) | 0.88 (0.56–1.42) | |
hsa-mir-4293: rs12220909G>C | ||||||
GG | 321 (55.7) | 287 | 13 | 0.853 | 1.00 (ref.) | 1.00 (ref.) |
CG | 209 (36.3) | 186 | 12 | 1.02 (0.85–1.23) | 0.96 (0.80–1.16) | |
CC | 46 (8.0) | 39 | 12 | 0.93 (0.67–1.30) | 1.01 (0.72–1.42) | |
hsa-mir-4513: rs2168518G>A | ||||||
GG | 329 (57.1) | 296 | 13 | 0.305 | 1.00 (ref.) | 1.00 (ref.) |
AG | 220 (38.2) | 193 | 12 | 0.92 (0.77–1.10) | 0.92 (0.76–1.10) | |
AA | 27 (4.7) | 23 | 27 | 0.76 (0.49–1.16) | 0.77 (0.50–1.18) | |
hsa-mir-4520a: rs8078913T>C | ||||||
TT | 259 (45.0) | 235 | 12 | 0.164 | 1.00 (ref.) | 1.00 (ref.) |
CT | 220 (38.2) | 193 | 14 | 0.89 (0.74–1.08) | 0.90 (0.74–1.09) | |
CC | 97 (16.8) | 84 | 12 | 0.80 (0.62–1.03) | 0.79 (0.62–1.02) | |
hsa-mir-5579: rs11237828T>C | ||||||
TT | 240 (41.7) | 210 | 14 | 0.182 | 1.00 (ref.) | 1.00 (ref.) |
CT | 278 (48.3) | 249 | 12 | 1.18 (0.98–1.42) | 1.16 (0.96–1.40) | |
CC | 58 (10.1) | 53 | 13 | 1.08 (0.80–1.45) | 1.15 (0.84–1.56) | |
hsa-mir-5689: rs9295535T>C | ||||||
TT | 273 (47.4) | 253 | 12 | 0.062 | 1.00 (ref.) | 1.00 (ref.) |
CT | 258 (44.8) | 222 | 14 | 0.83 (0.69–1.01) | 0.89 (0.74–1.07) | |
CC | 45 (7.8) | 37 | 13 | 0.75 (0.53–1.07) | 0.77 (0.55–1.10) | |
Validation set I | ||||||
Total subjects | 346 | 266 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 243 (70.2) | 180 | 17 | 0.001 | 1.00 (ref.) | 1.00 (ref.) |
CT | 82 (23.7) | 68 | 13 | 1.28 (0.97–1.70) | 1.18 (0.88–1.57) | |
CC | 21 (6.1) | 18 | 6 | 2.43 (1.49–3.98) | 2.20 (1.33–3.64) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.23 (1.16–1.76) | 1.34 (1.08–1.65) | ||||
Dominant model | ||||||
TT | 243 (70.2) | 180 | 17 | 0.006 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 103 (29.8) | 86 | 12 | 1.42 (1.10–1.84) | 1.31 (1.01–1.71) | |
Validation set II | ||||||
Total no. of subjects | 368 | 298 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 257 (69.8) | 203 | 16 | 2.20 × 10−6 | 1.00 (ref.) | 1.00 (ref.) |
CT | 93 (25.3) | 77 | 12 | 1.33 (1.02–1.73) | 1.33 (1.01–1.75) | |
CC | 18 (4.9) | 18 | 5 | 2.87 (1.76–4.68) | 2.29 (1.40–3.76) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.50 (1.23–1.84) | 1.43 (1.16–1.75) | ||||
Dominant model | ||||||
TT | 257 (69.8) | 203 | 16 | 0.001 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 111 (30.2) | 95 | 11 | 1.47 (1.15–1.89) | 1.45 (1.12–1.86) | |
Merged set | ||||||
Total subjects | 1290 | 1076 | ||||
hsa-mir-499: rs3746444T>C | ||||||
TT | 889 (68.9) | 725 | 15 | 1.15 × 10−9 | 1.00 (ref.) | 1.00 (ref.) |
CT | 327 (25.4) | 288 | 12 | 1.23 (1.08–1.41) | 1.19 (1.04–1.37) | |
CC | 74 (5.7) | 63 | 7 | 2.05 (1.58–2.66) | 1.86 (1.44–2.42) | |
Additive model | ||||||
CC vs. CT vs. TT | 1.33 (1.20–1.48) | 1.28 (1.15–1.42) | ||||
Dominant model | ||||||
TT | 889 (68.9) | 725 | 15 | 7.54 × 10−6 | 1.00 (ref.) | 1.00 (ref.) |
CT + CC | 401 (31.1) | 351 | 11 | 1.33 (1.17–1.51) | 1.27 (1.12–1.45) |
NOTE: Bold type: statistically significant, P < 0.05.
aThe Cox regression analysis was adjusted for age, smoking, stage, surgery, chemotherapy, and radiotherapy status.
Only the significant results above were validated in the validation sets I and II, and the results were consistent. In those 2 datasets, when compared with patients with the TT genotype, patients carrying the rs3746444CC homozygous genotype had an increased death risk (validation set I: HR, 2.20; 95% CI, 1.33–3.64; P = 0.002; validation set II: HR, 2.29; 95% CI, 1.40–3.76; P = 0.001) and had a decreased MST (validation set I: 6 vs. 17 months; log-rank test: P = 0.001; validation set II: 5 vs. 16 months; log-rank test: P = 2.20 × 10−6). Those patients with rs3746444C (TC + CC) variants exerted a shorter MST (validation set I: 12 vs. 17 months, log-rank test: P = 0.006; Fig. 1B; validation set II: 11 vs. 16 months, log-rank test: P = 0.001; Fig. 1C) and a poorer survival outcome from multivariate Cox model regression (validation set I: HR, 1.31; 95% CI, 1.01–1.71; P = 0.048; validation set II: HR, 1.45; 95% CI, 1.12–1.86; P = 0.004). Similarly, the rs3746444T>C polymorphism had a harmful additive effect on lung cancer prognosis in the validation set I (HR, 1.34; 95% CI, 1.08–1.65; P = 0.007) and validation set II (HR, 1.43; 95% CI, 1.16–1.75; P = 0.001).
When merged all patients with lung cancer from 3 studies, we observed that the rs3746444CT heterozygous and CC variant homozygous were associated with 19% (HR, 1.19; 95% CI, 1.04–1.37; P = 0.013) and 86% (HR, 1.86; 95% CI, 1.44–2.42; P = 6.26 × 10−6) increased lung cancer–specific of death risk respectively, compared with the TT wild-type homozygote genotype. The rs3746444C (TC + CC) adverse genotypes conferred a 4-months decreased MST compared with TT genotype (11 vs. 15 months, log-rank test: P = 7.54 × 10−6) and had a 27% higher death risk (HR, 1.27; 95% CI, 1.12–1.45; P = 0.2 × 10−3). The rs3746444C variants also conferred a 28% higher death risk under the additive model (HR, 1.28; 95% CI, 1.15–1.42; P = 4.21 × 10−6). We further performed stratification analysis by selected demographic variables and clinical statuses. As shown in Table 3, although the strength of associations represented by the HR values between the rs3746444C (CT + CC) variants and lung cancer survival were different between a plurality of stratums, the heterogeneity test only showed that the difference was significant in subgroup of chemotherapy and chemotherapy regimens (P = 0.008 and 0.026), the adverse effect on cancer survival carried by the rs3746444C variants was more prominent in chemotherapy patients (HR, 1.49; 95% CI, 1.27–1.76; P = 8.32 × 10−7), especially in those with cisplatin chemotherapy (HR, 1.51; 95% CI, 1.28–1.79; P = 1.60 × 10−6). We also found significant interactions between the chemotherapy (P = 0.034) or chemotherapy regimens (P = 0.044) and the variants on lung cancer survival.
Stratification analysis of association between the rs374644T>C genotypes and lung cancer prognosis by selected variables
. | TT genotype . | CT + CC genotypes . | Adjusted HR (95% CI)a . | Phomob . | Pinterc . | ||
---|---|---|---|---|---|---|---|
Variables . | Patients, n . | Death, n (%) . | Patients, n . | Death, n (%) . | . | n (%) . | n (%) . |
Age, y | 0.327 | 0.573 | |||||
≤60 | 458 | 371 (81.0) | 170 | 140 (82.4) | 1.19 (0.97–1.45) | ||
>60 | 431 | 354 (82.1) | 231 | 211 (91.3) | 1.36 (1.14–1.62) | ||
Sex | 0.411 | 0.365 | |||||
Male | 621 | 514 (82.8) | 283 | 249 (88.0) | 1.31 (1.12–1.52) | ||
Female | 268 | 211 (78.7) | 118 | 102 (86.4) | 1.16 (0.91–1.49) | ||
Family history of cancer | 0.812 | 0.193 | |||||
No | 841 | 683 (81.2) | 364 | 321 (88.2) | 1.29 (1.13–1.48) | ||
Yes | 48 | 42 (87.5) | 37 | 30 (81.1) | 1.21 (0.73–2.02) | ||
Family history of lung cancer | 0.878 | 0.187 | |||||
No | 866 | 706 (81.5) | 386 | 339 (87.8) | 1.29 (1.13–1.47) | ||
Yes | 23 | 19 (82.6) | 15 | 12 (80.0) | 1.21 (0.54–2.72) | ||
Smoking status | 0.280 | 0.151 | |||||
Never | 436 | 352 (80.7) | 194 | 167 (86.1) | 1.17 (0.97–1.41) | ||
Ever | 453 | 373 (82.3) | 207 | 184 (88.9) | 1.35 (1.13–1.62) | ||
Drinking status | 0.113 | 0.111 | |||||
Never | 749 | 617 (82.4) | 323 | 279 (86.4) | 1.22 (1.05–1.40) | ||
Ever | 140 | 108 (77.1) | 78 | 72 (92.3) | 1.61 (1.18–2.20) | ||
Surgery | 0.733 | 0.725 | |||||
No | 519 | 454 (87.5) | 244 | 228 (93.4) | 1.30 (1.11–1.53) | ||
Yes | 370 | 271 (73.2) | 157 | 123 (78.3) | 1.24 (1.00–1.55) | ||
Chemotherapy | 0.008 | 0.034 | |||||
No | 267 | 223 (83.5) | 147 | 129 (87.8) | 1.03 (0.81–1.26) | ||
Yes | 622 | 502 (80.7) | 254 | 222 (87.4) | 1.49 (1.27–1.76) | ||
Chemotherapy regimens | 0.026 | 0.044 | |||||
No | 267 | 223 (83.5) | 147 | 129 (87.8) | 1.03 (0.81–1.26) | ||
Platinum-based treatmentd | 553 | 446 (80.7) | 227 | 201 (88.5) | 1.51 (1.28–1.79) | ||
Other treatmente | 69 | 56 (81.2) | 27 | 21 (77.8) | 1.37 (0.81–2.32) | ||
Radiotherapy | 0.234 | 0.075 | |||||
No | 422 | 340 (80.6) | 192 | 173 (90.1) | 1.38 (1.14–1.66) | ||
Yes | 467 | 385 (82.4) | 209 | 178 (85.2) | 1.18 (0.99–1.41) | ||
Stages | 0.830 | 0.297 | |||||
I + II | 182 | 123 (67.6) | 72 | 60 (83.3) | 1.35 (1.10–1.66) | ||
III | 296 | 259 (87.5) | 114 | 98 (86.0) | 1.23 (1.01–1.56) | ||
IV | 411 | 343 (83.5) | 215 | 193 (89.8) | 1.29 (1.07–1.54) | ||
Histologic types | 0.160 | 0.178 | |||||
Adenocarcinoma | 374 | 300 (80.2) | 163 | 144 (88.3) | 1.15 (1.00–1.41) | ||
Squamous cell carcinoma | 274 | 220 (80.3) | 138 | 115 (83.3) | 1.38 (1.09–1.74) | ||
Large cell carcinoma | 40 | 35 (87.5) | 20 | 18 (90.0) | 1.14 (0.88–2.01) | ||
Small cell lung cancer | 110 | 91 (82.7) | 51 | 47 (92.2) | 1.48 (1.02–2.15) | ||
Otherf | 91 | 79 (86.8) | 29 | 28 (96.5) | 1.67 (1.17–1.97) |
. | TT genotype . | CT + CC genotypes . | Adjusted HR (95% CI)a . | Phomob . | Pinterc . | ||
---|---|---|---|---|---|---|---|
Variables . | Patients, n . | Death, n (%) . | Patients, n . | Death, n (%) . | . | n (%) . | n (%) . |
Age, y | 0.327 | 0.573 | |||||
≤60 | 458 | 371 (81.0) | 170 | 140 (82.4) | 1.19 (0.97–1.45) | ||
>60 | 431 | 354 (82.1) | 231 | 211 (91.3) | 1.36 (1.14–1.62) | ||
Sex | 0.411 | 0.365 | |||||
Male | 621 | 514 (82.8) | 283 | 249 (88.0) | 1.31 (1.12–1.52) | ||
Female | 268 | 211 (78.7) | 118 | 102 (86.4) | 1.16 (0.91–1.49) | ||
Family history of cancer | 0.812 | 0.193 | |||||
No | 841 | 683 (81.2) | 364 | 321 (88.2) | 1.29 (1.13–1.48) | ||
Yes | 48 | 42 (87.5) | 37 | 30 (81.1) | 1.21 (0.73–2.02) | ||
Family history of lung cancer | 0.878 | 0.187 | |||||
No | 866 | 706 (81.5) | 386 | 339 (87.8) | 1.29 (1.13–1.47) | ||
Yes | 23 | 19 (82.6) | 15 | 12 (80.0) | 1.21 (0.54–2.72) | ||
Smoking status | 0.280 | 0.151 | |||||
Never | 436 | 352 (80.7) | 194 | 167 (86.1) | 1.17 (0.97–1.41) | ||
Ever | 453 | 373 (82.3) | 207 | 184 (88.9) | 1.35 (1.13–1.62) | ||
Drinking status | 0.113 | 0.111 | |||||
Never | 749 | 617 (82.4) | 323 | 279 (86.4) | 1.22 (1.05–1.40) | ||
Ever | 140 | 108 (77.1) | 78 | 72 (92.3) | 1.61 (1.18–2.20) | ||
Surgery | 0.733 | 0.725 | |||||
No | 519 | 454 (87.5) | 244 | 228 (93.4) | 1.30 (1.11–1.53) | ||
Yes | 370 | 271 (73.2) | 157 | 123 (78.3) | 1.24 (1.00–1.55) | ||
Chemotherapy | 0.008 | 0.034 | |||||
No | 267 | 223 (83.5) | 147 | 129 (87.8) | 1.03 (0.81–1.26) | ||
Yes | 622 | 502 (80.7) | 254 | 222 (87.4) | 1.49 (1.27–1.76) | ||
Chemotherapy regimens | 0.026 | 0.044 | |||||
No | 267 | 223 (83.5) | 147 | 129 (87.8) | 1.03 (0.81–1.26) | ||
Platinum-based treatmentd | 553 | 446 (80.7) | 227 | 201 (88.5) | 1.51 (1.28–1.79) | ||
Other treatmente | 69 | 56 (81.2) | 27 | 21 (77.8) | 1.37 (0.81–2.32) | ||
Radiotherapy | 0.234 | 0.075 | |||||
No | 422 | 340 (80.6) | 192 | 173 (90.1) | 1.38 (1.14–1.66) | ||
Yes | 467 | 385 (82.4) | 209 | 178 (85.2) | 1.18 (0.99–1.41) | ||
Stages | 0.830 | 0.297 | |||||
I + II | 182 | 123 (67.6) | 72 | 60 (83.3) | 1.35 (1.10–1.66) | ||
III | 296 | 259 (87.5) | 114 | 98 (86.0) | 1.23 (1.01–1.56) | ||
IV | 411 | 343 (83.5) | 215 | 193 (89.8) | 1.29 (1.07–1.54) | ||
Histologic types | 0.160 | 0.178 | |||||
Adenocarcinoma | 374 | 300 (80.2) | 163 | 144 (88.3) | 1.15 (1.00–1.41) | ||
Squamous cell carcinoma | 274 | 220 (80.3) | 138 | 115 (83.3) | 1.38 (1.09–1.74) | ||
Large cell carcinoma | 40 | 35 (87.5) | 20 | 18 (90.0) | 1.14 (0.88–2.01) | ||
Small cell lung cancer | 110 | 91 (82.7) | 51 | 47 (92.2) | 1.48 (1.02–2.15) | ||
Otherf | 91 | 79 (86.8) | 29 | 28 (96.5) | 1.67 (1.17–1.97) |
NOTE: Bold type: statistically significant, P < 0.05.
aHRs were adjusted for age, smoking, stage, surgery, chemotherapy, and radiotherapy status in a Cox regression model.
bP value of homogeneity test between strata for the related ORs of rs3746444T>C (rs3746444 CT + CC vs. TT genotype).
cP value of test for the multiplicative interaction between rs3746444T>C genotypes and selected variables on cancer death in Cox regression models.
dPatients receiving cisplatin plus gemcitabine or plus vinorelbine or plus docetaxel or plus etoposide or a combination of the above regimens.
ePatients not receiving cisplatin but with other chemotherapeutics.
fMixed-cell or undifferentiated carcinoma.
Indicative role of the rs3746444T>C polymorphism on lung cancer survival
As shown in the Fig. 1D, the values of AUC for the rs3746444T>C polymorphism alone were relatively low from 0.566 to 0.609 to predict lung cancer survival in different censor points from half a year to 5 years, indicating a low accuracy of this SNP alone on predicting lung cancer survival. The values were quite similar with those for the risk factors and clinical features including age, smoking, stage, surgery, chemotherapy, and radiotherapy, which are found to have significant effects on lung cancer survival. Interestingly, when we united the SNP, the risk factors, and clinical features, there is certain accuracy for the combination to predict lung cancer survival with the AUC values covered from 0.665 to 0.711, suggesting an indicative role of this SNP on predicting lung cancer survival.
Effects of miRNA-499 with rs3746444T or C allele on cell chemosensitivity upon cisplatinum treatment
As shown in Fig. 2, the MTT assay indicated that inhibition of A549 cells treated with cisplatinum occurred in a dose-dependent manner, and A549-miRNA-499-C cells had a significant higher survival rate than the A549-miRNA-499-T cells and the A549 empty cells in followed 24 and 48 hours after transfection (Fig. 2B; 24 hours: P = 1.49 × 10−6; 48 hours: P = 7.56 × 10−7). However, no significant difference in the proliferation rate between the A549 cells transfected with miRNA-499-T allele transfectants and the “empty” lentiviral vector with cisplatinum treatment (Fig. 2B; 24 hours: P = 0.321; 48 hours: P = 0.623) and between the cells infected with the 3 vectors without cisplatinum treatment (Supplementary Fig. S2). Meanwhile, the Annexin V–FITC/PI staining assay showed that under a 200 μmol/L cisplatinum treatment, A549 cells expressing miR-499-C had a significantly lower apoptosis portion than those cells expressing miR-499-T or empty vector (P = 0.008, Fig. 2D). In contrast, no significant difference in the apoptosis portion between the A549 cells transfected with miRNA-499-T transfectants and the cells with the “empty” lentiviral vector (P = 0.483, Fig. 2D). Also, there were no different apoptosis rates in those transfectants without cisplatinum treatment (P = 0.126, Fig. 2C).
Effect of the rs3746444T>C polymorphism on cell chemosensitivity after cisplatinum treatment. A, location of the rs3746444T>C SNP in the miR-499 sequence. B, the proliferation curves for 3 kinds of A549 cells transfected respectively with the miRNA-499-C vectors, miRNA-499-T vectors, or an “empty” lentivirus vector under the treatment of cisplatinum using the MTT assay. The main graph depicts the outcomes of a 48-hour exposure and the small inserts, the results of a 24-hour exposure to different levels of cisplatin. C, the Annexin V-FITC/PI apoptosis assay for the A549 cells transfected with different miR-499 transfectants without cisplatinum treatment. D, the Annexin V-FITC/PI apoptosis assay for the A549 cells transfected with different miR-499 transfectants after 200 μmol/L cisplatinum treatment. Each point represents mean ± SD of at least 3 independent measurements. ×, a statistical significance with P < 0.05 between the A549 cells transfected with the 3 different transfectants using the ANOVA test with a Bonferroni correction.
Effect of the rs3746444T>C polymorphism on cell chemosensitivity after cisplatinum treatment. A, location of the rs3746444T>C SNP in the miR-499 sequence. B, the proliferation curves for 3 kinds of A549 cells transfected respectively with the miRNA-499-C vectors, miRNA-499-T vectors, or an “empty” lentivirus vector under the treatment of cisplatinum using the MTT assay. The main graph depicts the outcomes of a 48-hour exposure and the small inserts, the results of a 24-hour exposure to different levels of cisplatin. C, the Annexin V-FITC/PI apoptosis assay for the A549 cells transfected with different miR-499 transfectants without cisplatinum treatment. D, the Annexin V-FITC/PI apoptosis assay for the A549 cells transfected with different miR-499 transfectants after 200 μmol/L cisplatinum treatment. Each point represents mean ± SD of at least 3 independent measurements. ×, a statistical significance with P < 0.05 between the A549 cells transfected with the 3 different transfectants using the ANOVA test with a Bonferroni correction.
Effects of miRNA-499 with rs3746444T or C allele on tumor growth after cisplatinum treatment in vitro and in vivo
To investigate the effect of rs3746444T>C genotypes on tumor growth with cisplatinum treatment in vitro and in vivo, we performed soft-agar colony formation assay and animal models. As presented in Fig. 3A, no significant difference of colony formations was observed between the in those A549 cell transfectants without cisplatinum treatment. However, while going through treatment with different concentrations of cisplatinum(100 or 200 μmol/L), there were significantly decreased numbers of colony formation in A549-miR- 499-T cells compared with those of A549-miR-499-C cells or A549-miR-499 empty cells. Furthermore, we evaluated the effects of rs3746444T>C polymorphism on tumor growth in different animal models. The results showed that no significant tumor growth inhibition was observed between A549-miRNA-499-T and A549-miRNA-499-C groups those without treatment or treated with 0.5 mg/kg cisplatinum (P = 0.097, Fig. 3C; P = 0.132, Fig. 3D, respectively). In contrast, tumors grew faster and larger in A549-miRNA-499-C xenograft than in A549-miRNA-499-T xenograft after treatment with 5 mg/kg cisplatinum(P = 6.72 × 10−8, Fig. 3E), suggesting that the presence of rs3746444C variant in miR-499 reduced the effect of cisplatinum antitumor drugs.
The effect of rs3746444T>C polymorphism on tumorigenicity in vitro and in vivo. A, colony formation assays for different A549 transfectants with indicated concentrations of cisplatinum treatment. The arrow indicated that the colonies were under the microscope with 10× magnification. B, colonies numbers for these A549 transfectants. Columns, mean from 3 duplicates; bars, SD. C–E, subcutaneously implanted A549-miR-499 empty vector, A549-miR-499-T, and A549-miR-499-C cells xenografted tumors were established. After all mice developed tumors of approximately 200 mm3, mice were assigned to receive 0.5 and 5 mg/kg cisplatinum, respectively, or control carrier solution alone, for a total of 4 weeks. Data points represent the mean ± SD of 6 mice per group. Representative images of the tumors were taken 4 weeks following cisplatinum treatment, and the tumor volumes were determined. Columns, mean; bars, SD. ×, indicated a statistical significance with P < 0.05 between the A549 cells transfected with the 3 different transfectants using the ANOVA test with a Bonferroni correction.
The effect of rs3746444T>C polymorphism on tumorigenicity in vitro and in vivo. A, colony formation assays for different A549 transfectants with indicated concentrations of cisplatinum treatment. The arrow indicated that the colonies were under the microscope with 10× magnification. B, colonies numbers for these A549 transfectants. Columns, mean from 3 duplicates; bars, SD. C–E, subcutaneously implanted A549-miR-499 empty vector, A549-miR-499-T, and A549-miR-499-C cells xenografted tumors were established. After all mice developed tumors of approximately 200 mm3, mice were assigned to receive 0.5 and 5 mg/kg cisplatinum, respectively, or control carrier solution alone, for a total of 4 weeks. Data points represent the mean ± SD of 6 mice per group. Representative images of the tumors were taken 4 weeks following cisplatinum treatment, and the tumor volumes were determined. Columns, mean; bars, SD. ×, indicated a statistical significance with P < 0.05 between the A549 cells transfected with the 3 different transfectants using the ANOVA test with a Bonferroni correction.
Effects of miRNA-499 with rs3746444T>C polymorphism on the gene expression profiles
To further explore the potential molecular mechanism underlying the rs3746444T>C polymorphism in miRNA-499–induced poor cancer prognosis, we conducted microarray assay to compare the gene expression profiles between miRNA-499-T-allele lentivirus–transfected and miRNA-499-C-allele lentivirus–transfected A549 cells. Stably overexpressed miRNA-499 in both transfectants were established as quantified by the qRT-PCR assay (Supplementary Fig. S3). By comparing RNA transcription levels between the A549-miRNA-499-C and A549-miRNA-499-T groups, we found that 46 genes were differentially expressed with P < 0.05 (Supplementary File S1). Quantitative real-time PCR was further performed to validate the above differently expressed genes, and 33 of the 46 genes were confirmed to have a significantly different expression between the 2 transfected cells (Supplementary File S2). We then performed bioinformatics analysis to infer the target genes of miR-499, no available evidences supported any one of these 33 genes to be known or predicted target genes of miR-499 based on published articles, public experimental databases, or results from bioinformatics software, such as the TargetScan human 6.2 (http://www.targetscan.org) and the MicroCosm Targets Version 5 (http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5). However, by using the Gene Ontology (GO) analysis (http://www.geneontology.org), these 33 genes were annotated to be cancer-related based on their biologic function, that is, immunity and defense genes, growth-related genes, tumor invasion and metastasis-related genes, cancer stem cell–related genes, and cell death–related genes. It suggests that these 33 genes might be the downstream genes influenced by miR-499 and involved in cancer cells' behavior.
Discussion
To our knowledge, this study is the first time to demonstrate that the rs3746444T>C polymorphism in miR-499 mature sequence is associated with unfavorable prognosis of patients with lung cancer treated with platinum-based chemotherapy. The SNP accompanying age, smoking, stage, surgery, chemotherapy, and radiotherapy has a definite increased accuracy on predicting the survival of patients with lung cancer. Results from functional bioassays performed in vitro and in vivo were also consistent with the epidemiologic findings. However, for other miRNA polymorphisms, no statistical evidence was found for any association to the prognosis of lung cancer. Taken together, the current study provides the first evidence that the functional rs3746444T>C polymorphism may be used as a novel candidate biomarker for lung cancer prognosis.
Recently, accumulating evidences have demonstrated a strong link between miRNAs and cancer prognosis (29, 30). Altered structures or expressions of mature miRNAs have been reported to be involved in cancer biology, including carcinogenesis, progression, invasion, and metastases (31–33), which play vital roles as characters of specific proto-oncogenes or tumor suppressors (34). Furthermore, genetic variants in miRNAs are investigated to be associated with cancer susceptibility and prognosis in the past few decades (35–38). Variations within the mature miRNAs sequence may either weaken or reinforce the binding between miRNAs and their targets, thus leading to a corresponding regulation of the target mRNA translation (39) and resulting in various functional endings (12, 18). For instance, a study has shown that an SNP in the mature region of miR-125a significantly blocks the processing of pri-miRNA to pre-miRNA, in addition to reducing the miRNA-mediated translational suppression (18). Another study demonstrated that inherited mutations or rare SNPs in the primary transcripts of has-mir-15a and has-mir-16-1 are linked to familial chronic lymphocytic leukemia and familial breast cancer (40), which was further supported in a spontaneous mouse model of chronic lymphocytic leukemia (41). Consistently, a series of epidemiologic studies have evaluated the associations between the rs3746444T>C polymorphism in mature miR-499 and various human disease risk and prognosis (42–45). However, the results are controversial and inconclusive, and the precise biologic functions of this SNP in cancer progression have not yet been well elucidated. In the current study, our results indicated that the rs3746444T>C polymorphism was associated with a poor prognosis in patients with lung cancer receiving platinum-based chemotherapy.
The prevailing consensus is that critical miR “seed sequences” (bases 2–8) in mature region largely determine miR–mRNA interactions (46, 47). The genetic variation rs3746444T>C is appropriately located in the “seed sequences” of miR-499, so we postulated that naturally occurring variation in mature miR-499 sequences might change their targets or complementary affinity, thus altering suppression of target mRNAs and resulting in subversive biologic effects. Our microarray and subsequent qRT-PCR findings showed that the rs3746444T>C polymorphism may influence the expression of a series cancer-related genes that might be annotated to immunity and defense, cell growth, tumor invasion and metastasis, cancer stem cell, and cell death. Although no available evidences supported any genes to be known or predicted target genes of miR-499, it can be inferred that a part of these genes might be downstream genes that are influenced by miR-499. Thus, the SNP can influence the prognosis of lung cancer by affecting the expressions of these genes, which have consequences directly related to cancer cell survival and tumor growth.
Emerging evidences suggest that the effect of polymorphisms in miRNAs on prognosis may be dependent on clinical characteristics of the disease. Here, we found that the association between the rs3746444CT/CC genotypes and poor prognosis of lung cancer was more prominent in patients with receiving platinum-based chemotherapy. The A549-miRNA-499-C–transfected cells had inferior cell chemosensitivity including higher proliferation rate and lower apoptosis portion upon cisplatinum treatment than A549-miRNA-499-T–transfected cells, whereas without cisplatinum treatment, the above difference was not significant. Also, the rs3746444C allele had an ability of highly aggressive human lung cancer A549 cells to form colonies in soft agar and to form greater tumors in nude mice in vivo and in vitro, while undergoing cisplatinum treatment. These functional experiments were convincingly verified by our previous findings that rs3746444C variants contributed an adverse role of prognosis in those patients with platinum-based chemotherapy. In addition, one study has reported that miR-499 had the potential to regulate the DNA damage signaling (48). Also, DNA repair–related genes and cisplatinum resistance genes, such as AKR1C3 (49, 50), have higher expression in A549-miRNA-499-C cells than in A549-miRNA-499-T cells. All these indicated that rs3746444C variants conferred an unfavorable effect on lung cancer survival in those patients receiving platinum-based chemotherapy via modulating a series of genes and thus affect lung cancer therapeutic effectiveness, which seems relatively plausible.
Some limitations and strengths in our study need to be advertent. Our study is a hospital-based case-only study, and the populations were ethnic Han Chinese, so inherent selection bias cannot be completely excluded. Also, because of the technical limitation, we did not reveal any direct target genes of miR-499 with respect to the SNP rs3746444T>C, which can help to clarify the concrete molecular mechanism about the SNP on influencing cancer prognosis. However, we performed the study with large sample size, a total of 1,290 patients with lung cancer included in 3 survival cohorts consistently confirmed the association of rs3746444T>C polymorphism and lung cancer survival, which provided an improved statistical power and reduced the probability of false positives. Furthermore, the results of our functional findings were also consistent with the significant association for mir-499 rs3746444T>C and lung cancer prognosis. Nevertheless, further validations of the association of the SNP with lung cancer prognosis in other ethnic groups by larger prospective studies and functional assay to reveal target gene of the SNP rs3746444T>C of miR-499 were warranted.
In conclusion, our results provide the first evidence that rs3746444T>C polymorphism in has-mir-499 mature sequence had a poor prognosis of patients with lung cancer treated with platinum-based chemotherapy, which may be a useful predictive biomarker applied to improve prognosis in patients with lung cancer.
Disclosure of Potential Conflicts of Interest
No potential conflicts of interest were disclosed.
Authors' Contributions
Conception and design: J. Lu
Development of methodology: F. Qiu, J. Lu
Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): F. Qiu, R. Yang, X. Yang, W. Fang, J. Lu
Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): F. Qiu, X. Ling, J. Lu
Writing, review, and/or revision of the manuscript: F. Qiu, L. Yang, Y. Zhou, J. Lu
Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): F. Qiu, C. Xie, D. Huang, J. Lu
Study supervision: J. Lu
Other (participated in genotyping, quantitative real-time PCR assay, and animal experiment and collaborated with data interpretation): R. Yang
Other (participated in genotyping, MTT and colony formation assay): X. Yang
Other (coordinated the genotyping polymorphism and collaborated with flow cytometry analysis): L. Zhang
Other (coordinated the genotyping and the quantitative real-time PCR experiment): W. Fang
Acknowledgments
The authors thank Dr. Zhanhong Xie, Wanmin Zeng, and Ling Liu for their assistance in recruiting the subjects and Di Wu, Jiansong Chen, and Zhihuang Chen for their laboratory assistance.
Grant Support
This study was supported by the National Natural Scientific Foundation of China grants 30671813, 30872178, 81072366, 81273149, 81473040 (J. Lu) and partly by 81001278, 81171895, 81472630 (Y. Zhou); 81402753, 81102159, 81102061 (L. Yang); and Guangdong Provincial High Level Experts Grants 2010-79 (J. Lu).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.