Evaluation of serum amyloid A as a biomarker for gastric cancer.

BACKGROUND: Serum amyloid A (SAA) is a useful biomarker for gastric cancer in an animal model. We investigated the potential of SAA as a biomarker for gastric cancer in humans. METHODS: Serum levels of SAA from 96 gastric cancer patients were measured before and after curative gastrectomy; 32 patients with gastric ulcers and 52 healthy subjects were the control groups. The immunohistochemical study was performed to evaluate the protein expression over gastric cancer tissue slides. RESULTS: The mean SAA concentration was higher in gastric cancer patients (88.54 +/- 50.44 mg/l) than in healthy subjects (3.36 +/- 2.29 mg/l) and gastric ulcer patients (10.48 +/- 8.97 mg/l) (P < .05). The SAA concentration was associated with tumor stage (P = .0244) and location (P = .0016) but not with Lauren's histological type (P = .839). In the multivariate analysis, SAA level was correlated with tumor location (P < .0001) and lymph node status (P < .05). During follow-up, the mean SAA concentration increased significantly in 24 patients with tumor recurrence (P < .05) but did not change in 77 patients without recurrence. In the survival analysis, patients with SAA levels > 97 mg/l had a nearly fourfold increase in risk of death. Immunoreactivity was most prominent in blood vessel regions but not within cancer cells. CONCLUSIONS: These data not only demonstrated SAA was useful in predicting survival of patients with gastric cancer, but they also showed that SAA was a valuable tool for postoperative follow-up.

Identification of tumor-associated plasma biomarkers using proteomic techniques: from mouse to human.

In an effort to identify tumor-associated proteins from plasma of tumor-bearing mice that may be used as diagnostic biomarkers, we developed a strategy that combines a tumor xenotransplantation model in nude mice with comparative proteomic technology. Five human cancer cell lines (SC-M1, HONE-1, CC-M1, OECM1, GBM 8401) derived from stomach, nasopharyngeal, colon, oral and brain cancers were subcutaneously inoculated into nude mice and compared to control nude mice injected with phosphate-buffered saline. One month later, plasma from mice inoculated with cancer cells was collected for proteomic analysis using two-dimensional gel electrophoresis (2-DE) and mass spectrometry (MS). Comparison of plasma 2-DE maps from tumor-bearing mice with those produced from control mice revealed the overexpression of several mouse acute phase proteins (APPs) such as haptoglobin. Another APP, serum amyloid A (SAA), was found only in mice bearing tumors induced by the stomach cancer cell line SC-M1, which has not previously been demonstrated in xenotransplatation experiment. Furthermore, by using immunohistochemistry, SAA and haptoglobin were found to originate from the mouse hosts and not from the human cancer cell line donors. The protein alterations were further confirmed on patients with stomach cancers where up-regulated levels of SAA were also observed. These results indicate that APPs may be used as nonspecific tumor-associated serum markers. SAA in particular may serve as a potential marker for detecting stomach cancer. Taken together, the combination of the xenotransplatation model in nude mice and proteomics analysis provided a valuable impact for clinical applications in cancer diagnostics. In addition, our findings demonstrate that a panel of APPs might serve as screening biomarkers for early cancer detection.