Research ArticleCANCER

Multi-omics characterization of molecular features of gastric cancer correlated with response to neoadjuvant chemotherapy

See allHide authors and affiliations

Science Advances  26 Feb 2020:
Vol. 6, no. 9, eaay4211
DOI: 10.1126/sciadv.aay4211
  • Fig. 1 Study overview.

    (A) Sample collection and multi-omics data generation. Our study included 35 patients with GC who received neoadjuvant chemotherapy before surgery. We collected pretreatment biopsy samples and posttreatment surgically resected tumor samples. On the basis of rigorously evaluated radiological and pathological evidence, patients were classified into a response group (n = 17) and a nonresponse group (n = 18). We obtained multi-omics data on the pretreatment samples and the nonresponse, posttreatment samples through whole-exome sequencing (WES), whole-genome sequencing (WGS), and RNA sequencing (RNA-seq). (B) The representative radiological and pathological images from responsive and nonresponsive patients. The yellow arrows indicate the lesion sites. (C) The necrosis rate distribution in the response and nonresponse groups. (D) Mandard tumor regression grading scores of the response and nonresponse groups.

  • Fig. 2 Mutation signatures in pretreatment GC samples.

    (A) Contributions of six possible substitution types at different nucleotide contexts. The relative weights of the COSMIC mutational signature 17 (B) and the microsatellite instability (MSI) scores (C) between the nonresponse and response groups. P value was based on Wilcoxon rank sum test. (D) Tumor mutation burden (TMB) distributions between the two groups. P value was based on one-tailed t test. (B to D) The middle line in the box is the median, the bottom and top of the box are the first and third quartiles, and the whiskers extend to 1.5× interquartile range of the lower and the upper quartiles, respectively.

  • Fig. 3 Significantly mutated genes in pretreatment GC samples.

    (A) Selected significantly mutated genes (SMGs) identified by MuSiC2 [false discovery rate (FDR) < 0.05] in pretreatment tumor samples. The bars on the top and on the right show the mutational rate observed for each patient and the composition of mutations in selected genes, respectively. Genes are ordered by their mutational frequencies, and different types of mutations are marked in different colors. (B) C10orf71 shows a significant mutation bias in the nonresponse samples (P < 0.045). The mutational sites are shown in the gene cartoon. (C) C10orf71 mutations are associated on the resistance to cisplatin in gastric cell lines; t test, P = 1.1 × 10−4. AUC, area under the curve. (D) Functional proteomic profiling of cell cycle based on eight protein markers in reverse-phase protein arrays. (E) C10orf71 mutations are associated on a lower cell cycle score in gastric cell lines; t test, P = 0.015. (F) A proposed mechanistic model in which C10orf71 mutations confer resistance to neoadjuvant chemotherapy through causing a less active cell cycle state.

  • Fig. 4 Significant SCNAs and their downstream signaling effects in pretreatment GC samples.

    (A) Amplification signals for SCNAs plotted for response versus nonresponse groups. Two cancer genes, MYC and CCNE1, reside in the unique peaks at 8q24.21 and 19q21, respectively, in the response group, while MDM2, a negative regulator of TP53, resides in the unique peak at 12q15 in the nonresponse group. MYC (B) and MDM2 (D) mRNA expression levels in the nonresponse and response groups. P values were based on one-tailed Wilcoxon rank sum test. The middle line in the box is the median, the bottom and top of the box are the first and third quartiles, and the whiskers extend to 1.5× interquartile range of the lower and the upper quartiles, respectively. Enrichment of MYC target genes (C) and DNA repair pathway (E) in the up-regulated genes in the response group relative to the nonresponse group. FDR was based on gene set enrichment analysis.

  • Fig. 5 Mutational evolution following neoadjuvant chemotherapy.

    (A) Mutational profiles in cancer genes before and after neoadjuvant chemotherapy. (B) The top subnetwork enriched in mutational alterations following the treatment. The size of the circle indicates the number of samples with a mutation in the network. (C) The mutational allele frequencies in the coding region of C10orf71 before and after treatment in four patients. P value was based on paired t test. Mutations in different patients are shown in different colors. (D and E) A schematic representation of the putative evolution of the acquired C10orf71 mutations in the two patients.

  • Fig. 6 Changes in gene expression and tumor-infiltrating immune cell following neoadjuvant chemotherapy.

    (A) Volcano plot showing differentially expressed genes between matched pre- and posttreatment samples. Significant genes are shown in red (fold change > 2, FDR < 0.05). (B) Pathways that are significantly down-regulated following neoadjuvant chemotherapy. Significantly differentially expressed genes were identified at fold change >2, FDR < 0.05. The bar color indicates the number of differentially expressed genes in the pathway. (C) A heatmap showing mRNA expression fold changes (posttreatment/pretreatment) of MYC target genes driven by the treatment in the MYC-amplified tumors. Genes with a significant differential expression (paired t test, P < 0.05) are marked in green (down-regulated) and red (up-regulated). (D) Differential expression of GC therapeutic targets in the pre- and posttreatment samples. P values were calculated on the basis of paired Wilcoxon rank sum test. (E) The fractions of neutrophil and dendritic cells in the pre- and posttreatment samples. P values were calculated on the basis of paired t test. (D and E) The middle line in the box is the median, the bottom and top of the box are the first and third quartiles, and the whiskers extend to 1.5× interquartile range of the lower and the upper quartiles, respectively.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/6/9/eaay4211/DC1

    Fig. S1. The age and stage distributions of patients with GC in this study.

    Fig. S2. Somatic base substitution summary.

    Fig. S3. SCNA signal profiles identified by GISTIC2.0 in the nonresponse and response groups.

    Fig. S4. Variant allele frequency in pre- and posttreatment samples based on WES data.

    Fig. S5. SCNA signal profiles identified by GISTIC2.0 in pre- and posttreatment samples.

    Fig. S6. Changes in gene expression and immune cell compositions before and after treatment.

    Table S1. The summary of patient clinical information and sequencing data.

    Table S2. Somatic mutations identified in this study.

    Table S3. SMGs identified in this study.

    Table S4. GC cell lines profiled for RPPAs.

  • Supplementary Materials

    The PDF file includes:

    • Fig. S1. The age and stage distributions of patients with GC in this study.
    • Fig. S2. Somatic base substitution summary.
    • Fig. S3. SCNA signal profiles identified by GISTIC2.0 in the nonresponse and response groups.
    • Fig. S4. Variant allele frequency in pre- and posttreatment samples based on WES data.
    • Fig. S5. SCNA signal profiles identified by GISTIC2.0 in pre- and posttreatment samples.
    • Fig. S6. Changes in gene expression and immune cell compositions before and after treatment.

    Download PDF

    Other Supplementary Material for this manuscript includes the following:

    • Table S1 (Microsoft Excel format). The summary of patient clinical information and sequencing data.
    • Table S2 (Microsoft Excel format). Somatic mutations identified in this study.
    • Table S3 (Microsoft Excel format). SMGs identified in this study.
    • Table S4 (Microsoft Excel format). GC cell lines profiled for RPPAs.

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article