Serine/arginine-rich splicing factors (SRSFs) and HNRNPA1 have oncogenic properties. However, their proteomic expressions and practical priority in gastric cancer (GC) and colorectal cancer (CRC) are mostly unknown. To apply SFs in clinics, effective marker selection and characterization of properties in the target organ are essential.
We concurrently analyzed SRSF1, 3, and 5-7, and HNRNPA1, together with the conventional tumor marker carcinoembryonic antigen (CEA), in stomach and colorectal tissue samples (n = 420) using semiquantitative immunoblot, subcellular fractionation, and quantitative real-time polymerase chain reaction methods.
In the semiquantitative immunoblot analysis, HNRNPA1 and SRSF7 levels were significantly higher in GC than in gastric normal mucosa, and SRSF7 levels were higher in intestinal-type compared with diffuse-type of gastric adenocarcinoma. Of the SFs, only HNRNPA1 presented greater than 50 % upregulation (cancer/normal mucosa > 2-fold) incidences and CEA-comparable, acceptable (>70 %) detection accuracy (74 %) for GC. All SF protein levels were significantly higher in CRC than in colorectal normal mucosa, and HNRNPA1 levels were higher in low-stage CRC compared with high-stage CRC. Among the SFs, HNRNPA1 and SRSF3 presented the two highest upregulation incidences (88 % and 74 %, respectively) and detection accuracy (90 % and 84 %, respectively) for CRC. The detection accuracy of HNRNPA1 was comparable to that of CEA in low (≤ II)-stage CRC but was inferior to that of CEA in high (>II)-stage CRC. Extranuclear distributions of HNRNPA1 and SRSF6 (cytosol/microsome) differed from those of other SRSFs (membrane/organelle) in both cancers. In an analysis of the six SF mRNAs, all mRNAs presented unacceptable detection accuracies (≤70 %) in both cancers, and all mRNAs except SRSF6 were disproportionate to the corresponding protein levels in GC.
Our results provide a comprehensive insight into the six SF expression profiles in GC and indicate that, among the SFs, HNRNPA1, but not HNRNPA1 mRNA, is the most effective, novel GC marker. Regardless of the good to excellent detection accuracy of SRSF3 and HNRNPA1 in CRC, the SFs have lower practical priority than CEA, especially for high-stage CRC detection.