STAT3 as a prognostic marker in human gastric cancer

Gastric cancer is the fourth most common cancer worldwide with approximately 930 000 newly diagnosed cases annually.1 Compared with other cancers, morbidity rates are quite high (∼700 000/year), making gastric cancer the second most common cause of cancer death worldwide after lung cancer.1 Following expansion of the primary tumor mass, cancer cells commonly invade through the muscularis mucosa and extravasate into blood vessels, thus promoting metastasis via lymphatic nodes and vessels.2 Disseminating metastasis occurs in 80–90% of individuals with intestinal type gastric cancer who have a 5-year survival rate of 75% in those diagnosed in early stages, and less than 30% of those diagnosed in late stages.3,4 Compared with other cancers, a substantial proportion of gastric cancer patients who have undergone curative surgery develop recurrent disease. The use of histopathological features, such as depth of primary tumor and lymphonodal involvement as indicators of prognosis, has improved survival. However, in recent years new prognostic molecular markers have been sought, as it is anticipated that measurement of molecular markers would more accurately indicate long-term outcome. Molecular markers are classified as molecules (most commonly proteins) the expression of which is altered in cancer, and which are functionally related to the cancer phenotype. Thus, the presence or absence of molecular markers can be used as indicators in evaluating cancer patient prognosis. Although several molecular markers have been proposed as prognostic factors in gastric cancer patients, such as epidermal growth factor receptor (EGFR), vascular endothelial growth factor (VEGF), hypoxia-inducible factor 1-α (HIF-1-α) and p53,5–7 none have been accepted into standard clinical practice; histopathological scoring therefore remains the standard protocol. Ideally, gastric molecular markers should accurately correlate with the well characterized tumor–node–metastasis (TNM) protocol for calculating prognosis of gastric cancer patients; that is, the degree of tumor depth (T1–4), presence or absence of metastatic sites (M0–1) and the degree of lymphatic invasion (N1–4) by the primary gastric tumor (Table 1).8 The STAT proteins, of which there are seven members, are a family of transcription factors which regulate expression of genes involved in both normal and pathological cellular processes. Accumulating evidence has demonstrated that various STAT members, when chronically activated (by protein phosphorylation) or expressed (increased synthesis), can act as mediators of oncogenesis in a range of human malignancies.9 This strong association with particularly epithelial cancers, and the extent of altered expression, suggests that certain family members might be useful prognostic markers or even therapeutic targets for cancer, because they are the focus of pathological signaling. Of the STAT family members, STAT3, is most commonly activated in human cancer.9 Numerous in vitro studies have shown that STAT3 promotes proliferation and invasion of gastric cancer cells,10,11 and upregulates expression of potent angiogenic factors.11 Additionally, it has been shown that STAT3 desensitizes cancer cells to anti-proliferative and apoptotic stimuli.12,13 As a transcription factor, STAT3 promotes cancer cell growth by upregulating the transcription of many well characterized genes including the matrix metalloproteinases (MMP), cyclins, VEGF and Src.11–14 Furthermore, STAT3 has been shown to suppress expression of Fas,15 thus reducing activity of apoptotic pathways (Table 1). In this issue of the Journal, Kim et al. extend the association of STAT3 as a prognostic marker in gastric cancer, by demonstrating that STAT3 protein expression is associated with poor survival outcome in a moderate sized Korean cohort, after retrospective analysis of 5-year survival data.16 The authors set out to determine whether or not STAT3 and STAT5 protein expression, as measured semiquantitatively by immunohistochemistry, can be used as a surrogate for TNM staging. They found that gastric adenocarcinomas were more likely to be positive for STAT3 protein than adjacent uninvolved stomach, and that TNM staging scores in STAT3-positive tumors were significantly higher than for tumors that were negative for STAT3. Moreover, the prevalence of STAT3 expression correlated with TNM progression, although small numbers in later stages means that these data need to be interpreted cautiously. Finally, despite the relatively low numbers of patients for whom complete data could be recorded, there was a significantly reduced 5-year survival rate in patients with STAT3-positive tumors (40% survival) compared with those whose tumors were STAT3 negative (80%). Together, these data suggest that STAT3 expression may be both a useful staging marker and prognostic indicator, if validated by more extensive prospective studies in larger cohorts. This is further underscored by several previous reports which have also shown increased STAT3 or activated STAT3 expression in gastric adenocarcinomas.10,17–19 An important issue here is the careful specification of the protein being measured, and also the use of complementary and more quantitative methods than immunohistochemistry, such as western blotting. Because STAT3 is activated after phosphorylation by JAK kinases, and the expression of total STAT3 protein changes relatively little with oncogenic progression, the best and most sensitive measure of STAT3 is the ratio of activated to total STAT3 (western blotting) or phosphorylated STAT3 (immunohistochemistry). This conclusion is underscored by recent studies10 which, unlike the study by Kim et al. in this issue, have shown that activated gastric STAT3 expression is not only increased by approximately fivefold in Helicobacter pylori-associated gastritis, but is also associated with CagA expression. Animal studies in which STAT3 can be manipulated genetically in mice prone to gastric tumor growth strongly support a role for activated STAT3 in both tumorigenesis and the chronic inflammation accompanying it.20 This suggests that activated STAT3 is associated with early precancerous pathology in the stomach, and that therapeutic intervention by limiting STAT3 transcription, may contribute to preventing oncogenic progression. In summary, current evidence, such as the work of Kim et al. reviewed here, supports an important role for chronically activated STAT3 as an oncogenic transcription factor in gastric cancer. Not only may quantitation of STAT3 activity become a sensitive molecular prognostic marker for disease outcome, but development of new methods designed to limit STAT3 activity may be useful therapeutically.