Research ArticleCANCER

Remote modulation of lncRNA GCLET by risk variant at 16p13 underlying genetic susceptibility to gastric cancer

See allHide authors and affiliations

Science Advances  20 May 2020:
Vol. 6, no. 21, eaay5525
DOI: 10.1126/sciadv.aay5525
  • Fig. 1 The procedures of gene-set analysis of gastric cancer risk.

    (A) Flowchart of the study design. (B) Global map of the distribution of genetic variants by gene types, including protein-coding genes (red), noncoding RNAs (ncRNAs) (blue), the combination (orange), and intergenic regions (green). (C) Application of the gastric cancer GWAS. (D) Distribution-based gene-set analysis, including intervals of protein-coding genes, ncRNA genes, the combination, and intergenic regions. The top panel shows the SNP numbers, and the bottom panel shows the P value. (E) Gene-based gene-set analysis. The top panel shows the SNP numbers, and the bottom panel shows the P value. (F) The linkage disequilibrium plots based on Asian populations (Han Chinese in Beijing and Japanese in Tokyo from 1000 Genomes Project; left) and functional annotation with scores from RegulomeDB and HaploReg (right) for the third novel gene set. DNase, deoxyribonuclease.

  • Fig. 2 The expression pattern of the lncRNA GCLET and genetic effect of rs3850997.

    (A) Both public and in-house datasets showed significantly down-regulated GCLET expression in gastric cancer tissues [public databases: GSE13911, GSE37023, and GSE29272; in-house datasets: 51 pairs of tumor and normal tissues analyzed using reverse transcription polymerase chain reaction (PCR) and 20 pairs from one unpublished dataset detected using an SBC microarray]. Red dots, tumor (T) tissues; blue dots, normal (N) tissues. (B) The constitutive expression of GCLET in gastric cancer cells (BGC-823, SGC-7901, and MGC-803) is significantly lower than that in normal gastric cell (GES-1). (C) The fractionation of gastric cell lysates reveals the cytoplasmic expression of GCLET. The U6 RNA served as a positive control for nuclear gene expression. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. (D) An eQTL analysis for the association of rs3850997 and GCLET expression was conducted using 51 tumor tissues from Nanjing-1 and 280 tumor tissues from Nantong. P values were calculated using an unpaired t test, paired t test, Wilcoxon signed-rank test, and linear regression analysis. (E) The mediation model for the genetic effect of rs3850997 (exposure variable) on the risk of gastric cancer (dependent variable) through GCLET expression (mediator variable). The arrows indicate the direction of the mediation model.

  • Fig. 3 Allele-specific effect of rs3850997.

    (A) GCLET expression displayed allele specificity, with a preference for the G allele of rs3850997. The ratio of signal density at rs3850997 was calculated using Sanger sequencing in both the genomic DNAs (gDNAs) and complementary DNAs (cDNAs) derived from 32 gastric cancer tissues carrying the TG genotype. P values were calculated using the Wilcoxon signed-rank test. siRNA, small interfering RNA. (B) In dual-luciferase reporter assays, a higher transcriptional activity of the GCLET promoter was observed with the rs3850997 G allele than with the T allele. (C) In electrophoretic mobility shift assays (EMSAs), the G allele of rs3850997 showed a lower binding affinity for the negative regulatory TF CTCF than the T allele. (D) ChIP and allele-specific quantitative polymerase chain reaction (qPCR) indicated that CTCF preferentially bound to the T allele at rs3850997. Notably, the PCR products shown in lanes 1 to 4 were amplified using primers specific for the human H19 imprinting control region, which served as positive controls (lane 1, 100- to 600-bp ladder; lane 2, CTCF binding product; lane 3, immunoglobulin G (IgG) binding product as a negative control; lane 4, input product), and the products shown in lanes 5 to 9 were amplified with primers specific for rs3850997 alleles (lane 5, 100- to 600-bp ladder; lanes 6 and 8, CTCF binding product for T and G alleles, respectively; lanes 7 and 9, IgG binding product for T and G alleles, respectively). (E) The left diagram showed the relative positions of the long-range regulation of rs3850997 on GCLET transcription. Target and Anchor were two primers for hybrid fragments; the blue blocks refer to exons in the GCLET gene; the colorful pies represent multiple TFs, and the red line indicates the Pst1 restriction sites. TSS, transcription start site. The right plots showed the physical interactions of the rs3850997 intronic region with the GCLET promoter (top) as determined using PCR. The products from Pst1-digested cross-linked chromatin without ligation and noncross-linked gDNA with or without ligation were used as negative controls. We used the PCR products from gDNA that were not cleaved by any restriction enzyme as the loading controls (bottom).

  • Fig. 4 Association of GCLET expression with clinical features of gastric cancer and effect of GCLET on cellular phenotypes and molecular regulation.

    (A) Significant association of GCLET expression with the survival of gastric cancer, as determined using the log-rank test. (B to D) Significant associations of GCLET expression with clinical features, including tumor invasion (B), lymph node metastasis (C), and tumor-node-metastasis classification (D), as determined using a linear regression analysis. (E and F) Prominent effect of GCLET on gastric cancer cell proliferation, as assessed using the CCK8 assay (E) and colony-forming assay (F). OD, optical density. (G) Remarkable effects of GCLET on gastric cancer cell invasion and migration. (H and I) GCLET attenuated the effect of miR-27a-3p on the expression of its target gene FOXP2. Luciferase expression was reduced if the reporter plasmid contained the wild-type FOXP2 3′ untranslated region (3’UTR), and this effect was more notable when the reporter plasmids were cotransfected together with the miR-27a-3p mimics (H). GCLET independently increased the level of the FOXP2 protein and further rescued the inhibitory effect of miR-27a-3p on FOXP2 expression (I).

  • Fig. 5 Graphical representation of our findings.

    SNP rs3850997 was significantly associated with a decreased risk of gastric cancer, most likely by altering the CTCF binding affinity and mediating a long-range regulatory effect on the expression of the lncRNA GCLET. In addition, GCLET competed with miR-27a-3p to regulate FOXP2 expression, tumor growth, invasion, and migration.

  • Table 1 Association between rs3850997 and gastric cancer risk in GWAS and validation stages.

    MAF, minor allele frequency; HWE, Hardy-Weinberg equilibrium. The covariates of age, sex, and study design were adjusted in GWAS stage; the covariates of age and sex were adjusted in the Nanjing-1, Nanjing-2, Yixing, Nantong, and Jilin stages; and sex was adjusted in both the Nanjing-3 and European stages.

    StagesPopulationCases/controlsrs3850997 genotypes (TT/TG/GG)MAF (G allele)PHWEOR (95% CI)*P*
    CasesControlsCase/control
    GWASShanxi/Linxian1625/2100661/720/241706/1052/3420.371/0.4130.1310.83 (0.75–0.91)9.29 × 10−5
    Validation-1Nanjing-11275/1436473/612/170489/715/2280.379/0.4090.2130.87 (0.78–0.98)1.73 × 10−2
    Validation-2Nanjing-21021/1203398/424/163442/552/2060.381/0.4010.1360.92 (0.81–1.03)1.50 × 10−1
    Validation-3Nanjing-3550/1155205/256/69400/555/1750.372/0.4000.4580.88 (0.76–1.03)1.09 × 10−1
    Validation-4Yixing940/1061366/416/145339/518/1640.381/0.4140.1470.87 (0.76–0.99)3.02 × 10−2
    Validation-5Nantong471/652167/218/65225/319/970.387/0.4000.3530.96 (0.80–1.14)6.18 × 10−1
    Validation-6Jilin1096/1178402/510/155405/541/1990.384/0.4100.4280.89 (0.79–1.01)6.19 × 10−2
    Combined6978/87850.87 (0.84–0.92)2.43 × 10−8
    European254/283529/99/103222/1027/11860.660/0.6981.0000.74 (0.59–0.93)9.26 × 10−3
    Combined7232/11,6200.87 (0.83–0.91)2.13 × 10−9

    *The additive genetic model used in the logistic regression analysis was constructed to detect the association between rs3850997 and gastric cancer risk.

    †A meta-analysis was performed for the combined group with Pheterogeneity = 0.820 for Asian populations and 0.669 for all populations.

    Supplementary Materials

    • Supplementary Materials

      Remote modulation of lncRNA GCLET by risk variant at 16p13 underlying genetic susceptibility to gastric cancer

      Mulong Du, Rui Zheng, Gaoxiang Ma, Haiyan Chu, Jiafei Lu, Shuwei Li, Junyi Xin, Na Tong, Gang Zhang, Weizhi Wang, Fulin Qiang, Weida Gong, Qinghong Zhao, Guoquan Tao, Jinfei Chen, Zhifang Jia, Jing Jiang, Guangfu Jin, Zhibin Hu, Hongbing Shen, Meilin Wang, Zhengdong Zhang

      Download Supplement

      This PDF file includes:

      • Supplementary Materials and Methods
      • Tables S1 to S11
      • Figs. S1 to S11

      Files in this Data Supplement:

    Stay Connected to Science Advances

    Navigate This Article