Oral squamous cell carcinoma (OSCC), a frequently occurring cancer in the head and neck region, is characterized by an aggressive growth pattern, local invasiveness, and spread to cervical lymph nodes. Patient outcome depends on the conventional prognostic factors used in clinical practice. Advances in surgical and nonsurgical treatments have led to increased local tumor control in recent years. However, overall survival rates have not improved because of the prevalence of locoregional tumor recurrence and distant metastasis. Although there is general agreement that tumor infiltration of the resection margin is one of the most relevant predictive factors for the development of a local recurrent carcinoma, the presence of tumor-free margins does not guarantee against recurrence because carcinoma can develop following discontinued expansion of tumor cells in the vicinity [1, 2]. To improve survival periods of these patients, molecular and histological markers must be identified to target tumors with a high likelihood of metastatic spread. To date, no reliable or clinically applicable marker of tumor aggressiveness has been identified for OSCC. Different markers/marker complexes have been identified as active in tumor suppression or antitumor defense and display potential as prognostic factors. Furthermore, molecular biology investigations of resection margins have shown that detection of mutant p53 genes is linked with increased incidence of recurrent tumors. The p53 molecule is a 53-kD polypeptide. It acts as a transcription factor that controls the cell cycle by either arresting cells in the G1 phase through activation of the p21 gene or triggering apoptosis by activating genes. Another more recent approach to oral carcinogenesis focuses on the escape of malignant cells from apoptotic signals. Extensive research has been carried out on p53 in this respect, and there is broad evidence for its role in the manifestation of oral carcinoma. However, published data indicates that p53 alone is not particularly valuable in predicting prognosis. Additional markers of apoptosis such as Fas, Fas ligand (FasL), and Bax, as well as anti-apoptotic molecules such as bcl2/BAG-1, are reported to be relevant to prognosis in a smaller number of publications. All of these have shown a significant correlation with prognosis, but it is difficult to draw conclusions on the prognostic validity of these markers on the basis of the present data [3–5].

Another approach in predicting prognosis in cancer is expression of heat shock proteins (HSPs). HSPs are found in all organisms and all cell types. They are the most phylogenetically conserved proteins known with respect to both structure and function [6]. Usually, HSPs are expressed at low levels, and under normal physiological conditions, many members of the HSP family are involved in protein synthesis. When a cell is stressed, oligomeric complexes disassemble and polypeptides unfold. Under these conditions, the role of HSPs is to reverse such changes and, if refolding becomes impossible, to potentially speed up the removal of such denatured proteins. Expression of HSPs is induced even under nonstress conditions, including those of the cell cycle, development, and differentiation [7, 8]. During carcinogenesis, HSPs have been reported to alter their expression levels, showing either an increase or a decrease [9, 10]). HSP70 expression in colorectal carcinoma and breast carcinoma has been significantly correlated with low differentiation and poor prognosis [11, 12], whereas in renal cell carcinoma it has been reported to be associated with good prognosis [13]. Five main families of HSPs are known: low molecular weight, HSP65, HSP70, HSP90, and HSP100. Over the last few years, HSPs have also been shown to play a role in the antigenicity of tumors. Expression of HSPs on the surface of tumor cells, instead of their normal intracellular location, suggests they play a role in inducing an immune response against cancer. In a chemically induced mouse sarcoma, Ullrich et al. identified a tumor-specific transplantation antigen that appears to be an HSP, which is expressed on the cell surface and induces protective immunity [14]. Furthermore, a protein related to the HSP70 family has been shown to be immunogenic in oncogene-transformed rat fibroblasts [15], and HSP70 derived from MethA sarcoma (but not from normal tissue) has been demonstrated to be immunogenic, not in itself but in association with tumor peptides. The prognostic significance of HSP70 in esophageal squamous cell carcinoma has been reported [16]. The aim of the present study was to investigate the expression of highly stress-inducible HSP70 in OSCC and its use in predicting prognosis.

Many studies in cell biology have addressed the roles of HSPs as molecular chaperones, including protein folding and unfolding, translocation, and prevention of inappropriate protein aggregation [7, 19]. These findings show an association between HSPs and cell proliferation and the prevention of apoptosis [20]. While HSP expression has been recognized as a factor of prognostic value in certain tumors, the data are limited and the results often contradictory. For example, inducible HSP72 has been shown to be a negative prognostic factor for disease-free survival (DFS) in patients with lymph node-negative breast cancer [21], whereas other authors have shown that HSP72 positively correlates with estrogen receptors and inversely with expression of mutant p53 [22]. Santarosa et al. demonstrated prognostic implications of expression of HSP70 in patients suffering from renal cancer [23]. Although some evidence indicates that HSPs are involved in various aspects of cell transformation and immune response against cancer [24], their biological role and its implications for the clinical course in cancer patients are not clear. In OSCC, the number of positive nodes, macroscopic extracapsular spread, and tumor infiltration of the resection margins have been described as significant prognostic factors [1, 2]. Studies concerning HSP70 as a prognostic factor in esophageal carcinoma suggest that reduction of HSP70 expression is significantly correlated with poor prognosis [16, 25]. Although several studies have been performed to elucidate the relationship between HSPs and tumors in various organs, to our knowledge, there are comparatively few reports related to OSCC. Sugarman et al. reported that HSP70 expression is not a definitive marker of oral malignancy or malignant potential [26]. In terms of prognostic significance, Ito et al. examined 24 specimens of patients suffering from OSCC. Although HSP immunohistochemistry revealed changes in HSP expression during tumorigenesis of squamous epithelium of the tongue, there was no correlation between HSP staining and survival period, stage, lymph node metastasis, histological grade, or p53 immunostaining [27]. These results are in line with those of Gandour-Edwards et al, who found that HSPs were expressed in normal upper respiratory tract squamous mucosa, and their role in carcinoma thus remains unclear. None of the markers (p53, HSP27, or HSP70) demonstrated prognostic significance for 5-year survival. We confirm the previously identified association of cervical lymph node metastases with decreased survival [5].

The results of the previously published data are not in line with the findings of the present study. In this study, we demonstrated that for patients suffering from T2-tumors, positive expression of HSP70 results in a significantly lower mortality risk compared with patients with negative expression. The main difference regarding the study design may be that, in contrast to previous studies, our patient population was divided according to tumor stages in the different T-stadium of the tumor. The purpose of this division was to analyze the prognostic significance of each T-stadium separately. Considering the entire patient group, our findings would suggest the same result: HSP70 has no prognostic implications for our patients. Also in the seperated T3- and T4-tumors, HSP 70 did not show any prognostic significance. Only after focusing on patients suffering from T2-tumors could a significant difference be detected. The survival of patients suffering from T2 tumors with positive HSP70 expression was 8 times higher than that for patients with negative HSP70 expression.

The results of our study suggests that expression of HSP70 affects survival only in the early stage of the disease and that HSP70 membrane expression, in particular, is a target for natural killer cells [28]. Therefore in T2-tumors, the increased level of HSP70 may result in an extended tumor control by the natural killer cells. By implication, T1-T2 tumors of OSCC with low expression of Hsp70 could require more radical treatment.