Fig 1.
Summary of the algorithm used for the three-step process to identify surrogate genes of CRC recurrence via LVI and PNI with sequential outcome.
TPM, transcripts per kilobase million.
Table 1.
Six selected genes associated with lymphovascular invasion or perineural invasion.
Fig 2.
Proliferative activities of six molecules in overexpressing (left) and underexpressing (right) cells. SiNC, negative control siRNA. *p < 0.01–0.05; **p < 0.001.
Fig 3.
Invasive activities of six molecules in overexpressing (left) and underexpressing (right) cells. SiNC, negative control siRNA. *p < 0.01–0.05; **p < 0.001.
Fig 4.
Invasive activities of GSN- and OAS2-overexpressing (left) and -underexpressing (right) cells measured by gelatine zymography. SiNC, negative control siRNA. *p < 0.01–0.05; **p < 0.001.
Fig 5.
GSN immunoreactivity with Nm23-H1 was examined by immunoblotting/ immunoprecipitation (A) and indirect immunofluorescence (B). An anti-GSN monoclonal antibody was used to pull down the endogenous and overexpressed GSN. As GSN is exclusively observed in the cytoplasm, the GSN in this study indicates isoform 2.
Fig 6.
A. Immunoblotting of epithelial-mesenchymal transition (EMT) and pathway-related molecules in GSN- and OAS2-overexpressing (left columns) and -underexpressing (right columns) cells. B. Immunoblotting of autophagy-related molecules in GSN- and OAS2-overexpressing cells. SiNC, negative control siRNA.
Fig 7.
Recurrence-free survival (RFS) was compared between GSN and OAS2 mRNA-overexpressing and–underexpressing tumors in the CIT cohorts (GSE39582, n = 566).