Each year, more than 530,000 women develop cervical carcinoma (CC), and 270,000 die from this disease globally[1], which makes CC the third most common and the fifth most lethal cancer worldwide. Based on both clinical and histopathological variables, uterine cervical diseases are categorized into three groups: non neoplastic, intraepithelial neoplasia and carcinoma. Carcinoma is further classified into two main subtypes: cervical adenocarcinomas and squamous cell carcinoma (SCC). The survival rate for patients with CC, similar to many other cancers, is also related to early disease detection. The patient survival rate depends significantly on the clinical (FIGO: International Federation of Gynecology and Obstetrics) stage and is as high as 90% (good) when the disease is diagnosed early and as low as 15% (poor) when the diagnosis is delayed[2–4]. Therefore, early diagnosis and treatment could significantly increase the survival rates of CC patients and their quality of life. As oncogenic human papillomavirus (HPV) can be detected in nearly all CC patients, the role of HPV in the progression of CC has been determined. Over the course of years, only a small proportion of women with a chronic HPV infection will develop cancer, which may underline the potential diverse patterns of HPV infection. Therefore, although HPV DNA testing appears to be sensitive, it lacks specificity[5–7], and a recent report indicated that some types of CC, such as minimal deviation adenocarcinoma, were not associated with high-risk HPV[8]. Thus, in addition to HPV DNA testing, the discovery of novel diagnostic and therapeutic markers against this malignancy should be prioritized.

Serum amyloid A (SAA) is a positive acute-phase protein with four different isotypes in humans. SAA is generated primarily by the liver in response to trauma, infection, inflammation, and even neoplastic stimuli[9]. SAA1 and SAA2 each consist of 122 amino acids, including signal peptide sequences, and share more than 90% amino acid identity[10]. In addition, SAA3 is a pseudogene, and SAA4 is constitutively expressed at a constant level and is thus known as cSAA. The acute-phase SAAs (aSAA), SAA1 and SAA2, increase in concentration approximately several hundred-fold in response to inflammatory stimuli[11, 12]. However, recent studies have revealed other major sources of SAA outside of the liver: specifically, cancer tissues of the esophagus[13], lung[14], pancreas[15], ovary[16], and uterine endometrium[5]. Moreover, recent data suggest that A-SAA, by activating the transcriptional factor nuclear factor kappa-B (NFκB)[17–21] and inducing the expression of matrix metalloproteinases (MMPs)[22–24], could considerably influence carcinogenesis. NFκB is a transcriptional factor for many functionally heterogeneous genes, notably the cytokine genes and anti-apoptotic genes. The activation of these genes could suppress apoptosis[25]. Additionally, a number of mechanisms have shown that MPPs could stimulate angiogenesis and activate cell migration.

To the best of our current knowledge, the localized expression of SAA in CC has not yet been reported. In this study, using RT-PCR, immunohistochemistry (IHC), and ELISA, we aimed to examine SAA expression and secretion in human CC.

In this study, using immunohistochemistry and real-time PCR analyses, we demonstrated the local over-expression of SAA at the protein and mRNA levels in human uterine cervical carcinomas, which may suggest that the tumor cells are the predominant source of SAA in cancer patients. In addition, we evaluated SAA levels in 138 serum samples derived from patients with NNL diseases of the cervix, CIN (precursor of cancer), and uterine cervical carcinomas and found that the circulating levels of SAA gradually increased with the carcinogenesis and progression of cervical cancer. Our findings suggest that there is a critical role for SAA in cervical carcinogenesis and may provide a promising target for future clinical therapy and postoperative surveillance.

Our findings indicate that the elevated SAA levels in local tissues and sera were due to its expression and secretion from CC cells. First, we reported a high SAA gene expression level in CC cases. By real-time PCR, the SAA1 and SAA4 mRNA expression levels were significantly higher in CC tissues than in NNL cervical tissues, with 27.7- and 22.8-fold higher copy numbers, respectively. Additionally, cytoplasmic positivity for SAA was detected by IHC in all CC-derived tissues, which verified that the presence of SAA mRNA is accompanied by SAA protein synthesis. Furthermore, in addition to the high levels of SAA gene expression, we detected high levels of circulating SAA. Additionally, in the limited number of patients undergoing chemotherapy, the serum levels of SAA were low, which could indirectly confirm that CC cells could secrete and express SAA.

Moreover, we also found that SAA expression was associated with CC progression. Notably, although there were no significant differences between the serum levels of the patients harboring NNL gynecological diseases and CIN and patients with stage I and II CC, the mean SAA serum levels of these groups were gradually increased with disease progression (Figure 4b; Table 4), with the highest serum levels in the stage II group. Cocco[5] and Urieli-Shoval[16] recently reported on SAA expression in two major gynecological cancers, endometrial endometrioid carcinoma and ovarian epithelial tumor, respectively. Using in situ hybridization and IHC, they demonstrated local and differential expression of SAA in human endometrial carcinoma and ovarian epithelial tumors compared with normal endometrial and ovarian epithelial tissues. Furthermore, they demonstrated that there was progressively higher SAA positivity through the different stages of dysplasia to overt carcinoma. Although the biological importance of SAA in cancer patients is not well understood, these findings in human endometrial endometrioid carcinoma and ovarian epithelial tumors, combined with our results in uterine cervical carcinoma, appear to suggest a novel role for SAA autocrine production in gynecological tumorigenesis and progression.

Several studies have proposed that SAA protein functions are relevant to tumor cell invasion and metastasis. First, SAA may influence carcinogenesis. As mentioned above, SAA could activate the transcriptional factor NFκB, which could suppress apoptosis. Given that inflammation has recently been proposed to be associated with tumorigenesis[27, 28], SAA may play a role in local inflammation in the microenvironment of the malignant tissue by inducing the production of pro-inflammatory cytokines; tumor necrosis factor-α (TNFα); interleukin-1b (IL-1); and the chemokines CCL1, CCL3, and CCL4[29–31]. SAA1 also promotes the up-regulation of genes involved in phagocytosis, anti-apoptosis, and tissue remodeling[26]. Second, as it does for human leukocytes[32, 33], SAA may induce the adhesion of tumor cells and may enhance tumor cell survival and proliferation, as shown in rheumatoid arthritis[34–36]. Third, SAA may enhance tumor cell invasion and metastatic spread by becoming directly involved in enhancing the activity of matrix degrading enzymes (MMP/TIMP-1) and by increasing TNFα production in and through its association with collagen cleavage in vivo[22, 35, 36]. Finally, SAA-derived peptides may inhibit tumor cell attachment to extracellular matrix proteins, as shown for T-lymphocytes and platelets[37, 38]. Whether any of these possibilities have relevance in the development of human CC requires further investigation.

To the best of our knowledge, this is the first systematic study describing the expression of SAA in human CC tissues. We have demonstrated that SAA mRNA and protein expression levels were high in cervical tumors compared with NNL cervical tissues. In addition, high concentrations of SAA were detected in the sera of CC patients. Given that an HPV screening assay is not in itself sufficient for identifying cervical cancer, some other methods, such as the novel biomarker Lam-5, which is detected in cervical adenocarcinoma[39], and the hTERC amplification test[40], have been reported. Although our data are preliminary and exploratory and require further validation in future studies, they support the hypothesis that SAA may be involved in cervical tumorigenesis and may be a candidate tissue and serum biomarker and perhaps a therapeutic target. Furthermore, combining the new assays described above may constitute a more useful method for monitoring cervical cancer.

However, although our data on the SAA serum levels reveal an obviously increasing trend in patients with NNL cervical diseases, CIN, and stage I and II cervical carcinomas, the NNL and CIN groups and the stage I and II carcinoma groups lack statistical significance. Therefore, larger studies that include more patients would be necessary to confirm this hypothesis that SAA levels increase with increasing disease progression. Additionally, which up- or downstream signaling pathways could be useful for CC treatment remains unknown. Future studies that investigate the specific mechanism of SAA secretion in CC cells may help clarify this issue.