Elevated tumor and serum levels of the hypoxia-associated protein osteopontin are associated with prognosis for soft tissue sarcoma patients

Background Osteopontin (OPN) overexpression is correlated with a poor prognosis for tumor patients. However, only a few studies investigated the prognostic impact of expression of OPN in soft tissue sarcomas (STS) yet. Methods This study is based on tumor and serum samples from 93 adult STS patients. We investigated OPN protein levels in serum (n = 86) and tumor tissue (n = 80) by ELISA and OPN mRNA levels in tumor tissue (n = 68) by quantitative real-time PCR. Results No correlation was found between OPN levels in serum and tumor tissue. Moreover, an elevated OPN protein level in the serum was significantly associated with clinical parameters such as higher stage (p = 0.004), higher grade (p = 0.003), subtype (p = 0.002) and larger tumor size (p = 0.03). OPN protein levels in the tumor tissue were associated with higher stage (p = 0.06), higher grade (p = 0.003), subtype (p = 0.07) and an increased rate of relapse (p = 0.02). In addition, using a Cox's proportional hazards regression model, we found that an elevated OPN protein level in the serum and tumor tissue extracts is a significant negative prognostic factor for patients with STS. The relative risks of tumor-related death were 2.2 (p < 0.05) and 3.7 (p = 0.01), respectively. Conclusion Our data suggest OPN protein in serum as well as in tumor tissue extracts is an important prognostic factor for soft tissue sarcoma patients.

Background OPN, a phosphorylated glycoprotein, is detected in body fluids as well as in tumor tissues. Initially, it was suggested that OPN is expressed by macrophages infiltrating the tumor rather than by tumor cells themselves [1]. However, elevated serum or plasma levels of OPN have been detected in a variety of human cancers, which has been correlated with tumor progression and metastasis [2][3][4]. Moreover, tumor OPN protein over expression is also linked to a worse prognosis in different tumor entities such as breast cancer [5,6], lung cancer [7], cervical cancer [8,9], prostate cancer [10] and head and neck cancer [11][12][13]. Furthermore, previous studies suggest that high tumor OPN mRNA expression also correlates with an unfavorable prognosis [14][15][16][17][18].
In support of its prognostic value, direct or indirect inhibition of the OPN signaling pathway results in reduced malignant tumor potential with effects on cell survival, migration, invasion, tumor growth and metastasis (reviewed in [4,19,20]). However, only a few studies have investigated the role of osteopontin in tumor progression of sarcoma patients. For example, in primary sarcomas of the pulmonary artery, OPN protein could be detected in tumor cells and in macrophages and it is potentially involved in tumor progression and metastasis [21]. A further immunohistochemical study of STS showed that OPN in tumor tissue is associated with tumor stage, grade and overall survival of STS patients at 5-years [22]. However, another study using immunohistochemical evaluation of OPN expression did not provide predictive information on the outcome of osteosarcoma patients [23]. We therefore analyzed OPN mRNA and protein levels of tumor samples and OPN serum levels of 93 histological verified adult STS patients. Moreover, to characterize the role of OPN in soft tissue sarcoma, we correlated OPN levels with clinical parameters and prognosis.

Patients and tumor material
Tumor tissue and serum samples of 93 adult patients with histologically verified primary STS were analyzed (Table 1). In detail, we have analyzed OPN protein levels in serum (86 patients) and tumor tissue (80 patients) and OPN mRNA level in tumor tissue (68 patients). Patients and tissue samples have been partially described previously [24][25][26]. The study was carried out in compliance with the Helsinki Declaration and it was approved by the Ethics Committee of the Medical Faculty of the University Halle. All patients gave written informed consent (Institute of Pathology, University of Halle, Germany and Department of Surgery 1, University of Leipzig, Germany). Patients' mean age was 56 years (19 to 87 years). The median follow-up time was 44 months (2 to 146 months). The overall survival according to stage was 86% for stage I (n = 15), 59% for stage II (n = 39), 37% for stage III (n = 19) and 0% for stage IVa (n = 9). Tumor tissue was collected following surgical resection and snap-frozen in liquid nitrogen. Tumors were classified according to the van Unnik grading system [27].
Reverse transcription-polymerase chain reaction and realtime quantitative PCR Total RNA was isolated using the RNeasy Mini Kit according to the manufacturer's instructions (Qiagen, Hilden, Germany). Subsequently, 1000 ng of RNA was reverse transcribed with random hexamer primers and Superscript II reverse transcriptase (Invitrogen, Karlsruhe, Germany). The cDNA synthesis was performed as described previously [28]. The mRNA levels of OPN and the housekeeping gene, hypoxanthin phosphoribosyl-transferase (HPRT), were quantified using Quantitect SYBRGreen PCR kit (Qiagen, Hilden, Germany) in a Rotorgene 3000 thermal cycler (LTF Labortechnik, Wasserburg, Germany). The OPN and HPRT transcripts were amplified with primer pairs of

OPN-ELISA measurements
Tumor tissue extracts were prepared from frozen tissues using a standard extraction protocol. Briefly, 20 slices of 10 μm thickness were cut. The total protein was extracted by solubilizing the tissues in 1 ml Giordano buffer using sonication on ice. The cell lysates were clarified by centrifugation at 15000 rpm for 15 min. The protein content was determined using the Bradford assay. Fresh blood samples collected before surgical treatment were coagulated for 20 min at room temperature and centrifuged at 3000 rpm for 10 min (Sarstedt monovette, Nümbrecht, Germany). The serum and tumor tissue extracts were aliquoted and stored at -80°C until analysis. Aliquots of each sample were analyzed in duplicate using a commercial ELISA system for OPN (IBL International, Hamburg, Germany) according to the manufacturer's instructions [3].

Statistical methods
All statistical analyses were performed with the SPSS software package 16.0 for Windows (SPSS Inc., Chicago, IL). The differences in the numerical data between two groups were evaluated using the Pearson's test or Fisher's exact test. For survival analyses the overall survival of STS patients was used as follow-up end point. The survival curves were generated using Kaplan-Meier analysis and the log rank test was used to test for differences. For univariate and multivariate analyses (adjusted to tumor stage) the Cox's proportional hazard regression model was used to calculate the hazard ratio in the analysis of survival. A p-value of less than 0.05 was considered to be statistically significant.

Osteopontin levels in serum and tumor tissue
In patients with soft tissue sarcoma (STS) the median protein concentration of OPN in the serum and tumor tissue extracts was 704 ng/ml (184-2660 ng/ml) and 51.9 ng/mg (3.2-868 ng/mg), respectively (Fig. 1). The median OPN/HPRT transcript level ratio was 12.8 molecules OPN/pg HPRT (with a range of 0-110 molecules OPN/pg HPRT, Fig. 1). In particular, there was no correlation between OPN mRNA and OPN protein levels of tumors (n = 66; p = 0.4), serum and tumor OPN protein levels (n = 73; p = 0.9) or tumor OPN mRNA and serum OPN protein levels (n = 62; p = 0.7).

Osteopontin levels and survival analyses
In subgroups split at the median serum or tumor OPN protein level, there was no significant difference in overall survival. However, when specimens were divided into four groups according to their OPN protein levels, the serum and tumor OPN protein levels did associate with different clinical parameters and overall survival. The group with the lowest tumor (≤ 34.34 ng/mg) or serum (≤ 516.54 ng/ml) OPN protein levels, i. e., the lowest quartile, had a favorable outcome compared to the other three groups (Table 3). Kaplan-Meier analyses showed a significant decrease of overall survival for the three groups with high OPN protein levels compared to the group with the lowest tumor (p = 0.007; log-rank test) and serum (p = 0.04; log-rank test) OPN protein levels. The mean survival time of the patient group with the lowest OPN protein level was 103 months (serum) and 68 months (tumor) whereas patients of the three remaining groups (high OPN protein levels) survived on average 60 months and 58 months (for serum and tumor tissue OPN protein levels, respectively). In addition, patients with both high serum and high tumor OPN protein levels had the worst prognosis (data not shown). Using a univariate Cox regression hazard model for STS patients with high OPN protein level in the serum or in tumor tissue extracts, the risk of tumorrelated death was a 2.2 (95% confidence interval (CI) 1.01-5.1; p < 0.05) and 3.7 (95% confidence interval (CI) 1.3-10.4; p = 0.01), respectively (Fig. 2, Table 3). However, in a multivariate Cox's regression hazard model adjusted to tumor stage OPN protein level was not significantly correlated with prognosis.

Bivariate analysis of osteopontin levels with clinical parameters
Bivariate analysis of OPN expression did show an association of serum and tumor OPN protein levels with  (Table 3).

Discussion
OPN overexpression is linked to an unfavorable prognosis in a variety of human cancers. However, only a few studies to date have investigated the prognostic impact of osteopontin in sarcoma patients. In the present study, we analyzed OPN mRNA and protein levels in tumors and the OPN protein levels from the serum of 93 soft tissue sarcoma patients and correlated OPN levels with clinical parameters and prognosis. In our study no correlation was found between OPN levels in serum and tumor tissue. It is well known that extracellular OPN is involved to tumor progression with signaling to proliferation, migration and survival. Also intracellular OPN is involved in cell migration [29]. Additionally, there is also evidence that suggests that intracellular OPN has other cellular functions, particularly in relation to cell proliferation [30]. But regulation between intra-and extracellular OPN forms remains unknown. Nordsmark et al. (2007) assume a relation between plasma concentration of OPN and intratumoral OPN production [31]. However, in agreement with our Our data suggest that increased serum and tumor OPN protein levels are significantly associated with clinical parameters, such as tumor stage, tumor grade, subtype of tumor, tumor size and the rate of relapse (Table  3). In the Kaplan-Meier analysis, a significantly decreased overall survival was observed for patients with strong OPN protein expression in their serum (p = 0.04; Table 3) and tumor tissue (p = 0.007; Table 3). Furthermore, the univariate Cox's regression hazard model demonstrated that STS patients with a high serum and tumor OPN protein level have an increased risk of tumor-related death of 2.2 (p < 0.05) and 3.7 (p = 0.01; Fig. 2), respectively (Table 3). This is in agreement with a previously reported immunohistochemical study of 33 soft tissue sarcoma patients that showed an increased OPN level significantly correlated with higher tumor stage, grade and overall survival [22]. Furthermore, primary sarcomas of the pulmonary artery display an abundant immunohistochemical OPN protein staining of tumor cells and extracellular matrix. Therefore, OPN protein may play a substantial role in tumor progression of vessel sarcomas [21]. However, an immunohistochemical OPN study did not find any correlation between OPN protein expression and outcome of osteosarcoma patients [23]. In this regard, Luo et al. were able to show that different osteosarcoma cell lines had significantly lower OPN levels than mature osteoblasts [32]. This may indicate that OPN has no significance for osteosarcoma tumor growth. However, antisense oligodeoxynucleotides against human OPN reduced the tumorigenicity of xenotransplanted osteosarcoma tumors in nude mice [33]. Additionally, our analysis shows that siRNA-induced osteopontin inhibition results in reduced clonogenic survival and migration of soft tissue sarcoma and breast cancer cell lines (Hahnel et al., manuscript submitted). Our analysis of tumor OPN mRNA expression of 68 soft tissue sarcoma patients did not show any association with clinical parameters or prognosis. In contrast, another study of 41 osteosarcoma patients did show that a high OPN mRNA level is correlative with overall survival, event-free survival and relapse-free survival [34]. Moreover, a study of 15 adult soft tissue sarcoma patients showed a significant increase of OPN mRNA levels compared to normal tissue [22]. In addition, OPN splice variants may be involved in tumor progression. In particular, osteopontin-c has been postulated to be a significant factor for glioma [35], breast cancer [17,36] and hepatocellular carcinoma [37]. Clearly, more data are needed to characterize the role of osteopontin mRNA in sarcomas.

Conclusion
In the present study, we analyzed OPN mRNA and protein levels in tumors and the OPN protein levels from the serum of 93 soft tissue sarcoma patients. Our data suggest that increased serum and tumor OPN protein levels are significantly associated with clinical parameters, such as tumor stage, tumor grade, subtype of tumor, tumor size and the rate of relapse. In addition, using a univariate Cox's proportional hazards regression model, we found that an elevated OPN protein level in the serum and tumor tissue extracts is a significant negative prognostic factor for patients with STS. Our data suggest OPN protein in serum as well as in tumor tissue extracts is an important prognostic factor for soft tissue sarcoma patients.