Prognostic significance of frequent CLDN18-ARHGAP26/6 fusion in gastric signet-ring cell cancer

Abstract

Signet-ring cell carcinoma (SRCC) has specific epidemiology and oncogenesis in gastric cancer, however, with no systematical investigation for prognostic genomic features. Here we report a systematic investigation conducted in 1868 Chinese gastric cancer patients indicating that signet-ring cells content was related to multiple clinical characteristics and treatment outcomes. We thus perform whole-genome sequencing on 32 pairs of SRC samples, and identify frequent CLDN18-ARHGAP26/6 fusion (25%). With 797 additional patients for validation, prevalence of CLDN18-ARHGAP26/6 fusion is noticed to be associated with signet-ring cell content, age at diagnosis, female/male ratio, and TNM stage. Importantly, patients with CLDN18-ARHGAP26/6 fusion have worse survival outcomes, and get no benefit from oxaliplatin/fluoropyrimidines-based chemotherapy, which is consistent with the fact of chemo-drug resistance acquired in CLDN18-ARHGAP26 introduced cell lines. Overall, this study provides insights into the clinical and genomic features of SRCC, and highlights the importance of frequent CLDN18-ARHGAP26/6 fusions in chemotherapy response for SRCC.

Fig. 1

Survival outcomes in gastric cancer patients with different signet-ring cell frequency (2006–2012). Survival curves of patients among the non-SRCC group (N = 837), con-SRCC group (N = 522), and SRCC group (N = 345) were illustrated in all patients (a), and stages III/IV (b). Impact of chemotherapy introduction on survival was also illustrated separately in terms of SRCC content in all patients (c), and stage III/IV (d)

Fig. 2

Landscape of key genetic alterations in HSRCC of gastric cancer. The patient samples are shown on the x-axis. Information of mutation rate, alcohol history, smoke history, tumor location, stage, patient age, and sex are shown on the top of y-axis, followed by the key genetic alterations including significant mutated genes. Frequency of each alteration was illustrated on the right of the mutation heat plot

Fig. 3

Somatic copy number variations and structure variation in HSRCC. a Somatic structure variations of all patients were combined and illustrated with CIRCOS plot. Translocations between CLDN18 and ARHGAP26/6 were highlighted in red line. Recurrent mutated genes (SNV/INDEL only) were indicated in the outlier of rim and the SMGs were labeled in red (including ARHGAPs and ARHGEFs). Cytoband was illustrated in the inner ring, followed by illustration of copy number alteration (orange represent gain and green represent loss). Structure variations were shown inside of the CIRCOS plot, red lines represent the recurrent CLDN18-ARHGAP26/6 fusions, green and black lines represent inter-chromosomal and intra-chromosomal translocations. b Illustration of breakpoint of CLDN18 and ARHGAP26/6 in DNA level (upper arrows) and RNA level (lower arrows). Fusions of CLDN18 with exon 10 of ARHGAP26, exon 12, and ARHGAP6 were indicated with green, purple and orange upper arrows in DNA, respectively. The junctions of CLDN18 and ARHGAPs in RNA level were indicated with red, yellow, and blue lower arrows or dashed lines in the gene map demonstration and Sanger sequencing graphs, respectively

Fig. 4

Clinical characteristics of patient with CLDN18-ARHGAP26/6. 829 Patients (combining 32 patients for whole-genome sequencing and 797 patients for validations) with distinct clinical features in terms of age, gender, and signet-ring cell content were illustrated in a–c, respectively. P value was estimated by using logistic regression, and the center values represent median age; WT, wildtype. Each box includes data between the 25th and 75th percentiles, with the horizontal line indicating the median. Whiskers indicate the maximal and minimal observations that are within 1.5 times the length of the box

Fig. 5

Impact of CLDN18-ARHGAP fusion on chemotherapy treatment outcomes.Survival curves with/without chemotherapy treatment in patients without CLDN18-ARHGAP fusion (a) or fusion burden (b) at stage III and IV. P value was estimated by using Cox model

Fig. 6

Impact of CLDN18-ARHGAP26 overexpression in cell line. Cell proliferation (a), cell migration (b), and drug response to oxaliplatin/5-fluorouracil (c) was estimated in CLDN18-ARHGAP26 stably expressed BGC-823 cells and its matched control. Each experiment has been replicated for three times, and data is presented as mean ± SD