Growth differentiation factor-15 as a prognostic biomarker in ovarian cancer

Methods GDF-15 concentration was measured by immunoradiometric assay in plasma samples from patients with invasive ovarian cancer ( n = 125), borderline ovarian tumor (BOT, n = 43), and benign ovarian tumor ( n = 144), from healthy women ( n = 40), as well as in effusion samples ( n = 44) from women with advanced ovarian cancer. Sections of ovarian carcinoma ( n = 20), BOT ( n = 9), and cystadenoma ( n = 7) were immunostained for GDF-15. Results Median plasma GDF-15 concentration was elevated in ovarian cancer as compared to healthy controls and women with benign ovarian tumors or BOT ( p < 0.001). GDF-15 plasma concentration correlated inversely with survival time and was an independent predictor of survival, after correction for FIGO stage and age ( p = 0.01). GDF-15 protein was cytoplasmatically expressed in serous tumor cells and detectable in high concentrations in effusion samples. Conclusion GDF-15 emerges as a new potential biomarker in ovarian cancer. Keywords Biomarker GDF-15 Gynecological cancer Ovarian cancer Survival Introduction Ovarian cancer has the highest mortality rate of all gynecologic cancers and is the fifth leading cause for female cancer-related death. Circulating CA 125 is currently the biomarker commonly used, but its value is limited, especially in the early stages of disease, where up to 50% of women have concentrations within normal range [1] . Because of the current poor prognosis for women with ovarian cancer, there is a need for identification of new biomarkers improving our understanding, diagnosis, and follow-up of ovarian cancer. Growth differentiation factor-15 (GDF-15) is a distant member of the transforming growth factor (TGF)-β superfamily, also named macrophage-inhibitory cytokine-1 (MIC-1), and was originally identified in an activated macrophage cell line [2] . The major biological role of GDF-15 is uncertain, but GDF-15 is suggested to regulate a wide variety of physiologic processes, including growth inhibition, induction of apoptosis, cell detachment, and tumor invasiveness [3] . In healthy individuals, GDF-15 is weakly expressed in most tissues, with the notable exception of the placenta [4] . Accordingly, very high circulating concentrations of GDF-15 can be measured in pregnant women in their second and third trimesters [5,6] . While the function of GDF-15 during pregnancy remains unknown, low serum concentrations of GDF-15 have been found to indicate an increased risk of miscarriages [7] . We have recently demonstrated augmented circulating concentrations of GDF-15 in pregnancies complicated by preeclampsia or diabetes mellitus, conditions associated with an augmented risk of cardiovascular disease [6] . Along that line, elevated circulating concentrations of GDF-15 have been associated with an increased risk of future cardiovascular events in apparently healthy elderly women [8] . In contrast to its low expression under baseline conditions, GDF-15 is strongly increased in pathological situations associated with inflammation, acute tissue injury, and malignancy [9–12] and is induced rapidly by cytokines [2,4,13] . GDF-15 overexpression has been found in prostate, thyroid, pancreatic, and colonic cancers as well as in malignant melanomas [3,11,14–18] . As GDF-15 has a potential clinical value in other malignancies, the aim of our study was to explore the role of GDF-15 as a potential biomarker in ovarian cancer, relating plasma concentrations to clinicopathologic parameters and survival time. We also wanted to explore ovarian tumor and effusion expression of GDF-15. Materials and methods Study subjects Preoperative fasting EDTA plasma samples were collected from women with primary ovarian cancer ( n = 125), borderline ovarian tumor (BOT, n = 43), and benign ovarian tumors ( n = 144) operated on at Oslo University Hospital, Ulleval, during 2003–2009. No patients had a prior diagnosis of cancer or had received chemotherapy or surgery for the present disease. Fasting EDTA plasma samples were also collected from 20 premenopausal and 20 postmenopausal healthy women, working at the same hospital. Also, 42 effusion samples (38 ascites and 4 pleural fluids) from 38 women with advanced ovarian cancer from the same study cohort were available. Plasma was collected and stored as previously described [6,19] , as were ascites and pleural fluids [20] . Clinical information was collected from hospital records as well as by patient interviews. Follow-up information was obtained from hospital records and from physicians in private practice. Clinicopathologic data for the total study groups are presented in Table 1 . The study was approved by the Regional Committee of Medical and Health Research Ethics (REK) in South-Eastern Norway, and all participating women provided signed consent. GDF-15 immunoassay GDF-15 in plasma and effusions was measured by an immunoradiometric sandwich assay using a polyclonal, affinity chromatography–purified goat anti-human GDF-15 IgG antibody (R&D Systems, Minneapolis, MN). All analyses were performed in duplicate (and blinded for diagnosis) in the laboratory where the assay was developed [21] . The assay has a detection limit of 20 ng/L, an intra-assay imprecision ≤ 10.6% and an interassay imprecision ≤ 12.2% [21] . The assay is not appreciably influenced by the anticoagulant matrix, and GDF-15 concentrations could therefore be compared between EDTA plasma and effusions [21] . GDF-15 immunohistochemistry Representative tumor sections from 20 women with serous ovarian carcinomas (10 with low and 10 with high plasma GDF-15), 9 women with serous BOT, and 7 women with serous cystadenomas were formalin-fixed, paraffin-embedded, and stained, using a primary rabbit-anti-GDF15 polyclonal IgG antibody (Atlas Antibodies AB, Stockholm, Sweden, 1:300 dilution). PT Link in combination with Target Retrieval Solution (50× Tris–EDTA buffer, pH = 9.0) (both from Dako, Glostrup, Denmark), was used as a pretreatment. Staining was performed using the EnVision FLEX + system (Dako). Immunoreactivity was scored by two authors (A.J.B. and B.D., the latter is a gynecopathologist). Staining extent was scored on a 0 to 4 scale, corresponding to staining of 0%, 1–5%, 6–25%, 26–75%, and 76–100% of cells, respectively. Standard laboratory analyses High-sensitivity C-reactive protein (CRP) and CA 125 levels were analyzed at the Department of Clinical Chemistry, Oslo University Hospital, Ulleval, and have been presented previously for some of the patients included in the present study [19] . Statistical analysis Statistical analysis was performed using SPSS (version 17.0, Chicago, IL). A P value of <.05 was considered statistically significant. The clinical characteristics are presented as median values and range, or percentage of patients, and the GDF-15 results are presented as medians and 95% CI of the medians. Nonparametric Mann–Whitney U test was used when comparing continuous data. Spearman's correlation was used to calculate correlation coefficients. χ 2 Test was used for categorical data. Univariate survival analyses were performed for patients with ovarian cancer and plasma analyses of GDF-15, using the Kaplan–Meier method and log-rank test, grouping low versus high concentration (dichotomizing GDF-15 concentration at the median 1242 ng/L, thus comparing groups with GDF-15 < 1242 vs. ≥ 1242 ng/L). Multivariate survival analyses were performed with Cox regression. A receiver operating characteristic (ROC) curve was constructed for GDF-15 and CA 125. Results Clinical characteristics are presented in Table 1 . As expected, preoperative risk-of-malignancy index (RMI) [22,23] , calculated as the product of an ultrasound adnexal tumor score grading (1–3), menopausal state grading (1–3), and CA 125 serum concentration (U/mL), differed significantly as expected between the three patient groups with ovarian tumors ( Table 1 , p < 0.001). Fig. 1 A shows the individual GDF-15 plasma concentrations for all patient groups and healthy controls. GDF-15 plasma concentrations in benign ovarian tumors and BOT Plasma GDF-15 concentration correlated positively with age for both the benign ovarian tumor and BOT groups (Spearman's correlation coefficients: 0.56 and 0.63, p < 0.001). BMI or parity was not associated with plasma GDF-15 concentrations in either groups (data not shown). Survival up to 6 years postoperatively was followed in the BOT group. Most women recruited were diagnosed with early-stage disease (FIGO stage I–II in 41 out of 43 patients), and only one cancer-related death occurred, while another patient had a clinical relapse. GDF-15 concentrations in ovarian cancer Median plasma GDF-15 concentration was elevated in ovarian cancer as compared to healthy controls and women with benign ovarian tumors or BOT ( Table 1 , all p < 0.001). Median GDF-15 increased with FIGO stage in ovarian cancer (996, 1024, 1298, and 1876 ng/L for stages I–IV, respectively, Spearman's correlation coefficient: 0.26, p = 0.01). Preoperative RMI correlated positively with preoperative plasma GDF-15 in the ovarian carcinoma group (Spearman's correlation coefficient: 0.25, p = 0.005). Median GDF-15 did not differ between poorly vs. well-differentiated or moderately differentiated tumors or between groups of women with normal or elevated BMI or between women with parity < 2 and multiparous women (≥ 2) (data not shown). Plasma GDF-15 concentrations correlated positively with age (Spearman's rho: 0.41, p < 0.001) and postmenopausal women tended to have a higher preoperative median GDF-15 concentration as compared to pre-and perimenopausal women (median value 1283 vs 850 ng/L, p = 0.05). In our ovarian cancer population, median CA 125 was 614 U/mL (stage I–II: 144 U/mL vs. stage III–IV: 906 U/mL, p < 0.001). Table 1 shows median CA 125 concentrations for all FIGO groups. There was a weak positive correlation between CA 125 and GDF-15 (Spearman's correlation coefficient: 0.25, p = 0.005). Also, CRP correlated positively with GDF-15 (Spearman's correlation coefficient: 0.57, p < 0.001) as did CRP and CA 125 (Spearman's correlation coefficient: 0.41, p < 0.001). Fig. 1 B presents ROC curves for GDF-15 and CA 125, calculating sensitivity and specificity for the biomarkers' ability to identify ovarian cancer for all possible cutoffs, as compared to women operated on for BOT or benign ovarian tumors. For GDF-15, the AUC (area under the curve) is 0.79 ( p < 0.001). For example, the sensitivity for preoperatively diagnosing ovarian cancer in our study of women with an identified adnexal tumor, with a GDF-15 threshold of 736 ng/L, is 0.8, and the specificity is 0.58. As shown in Fig. 1 B, CA 125 performed better as a preoperative discriminator in our patient population, with an AUC of 0.90 ( p < 0.001). Survival up to 6 years postoperatively was followed in the ovarian cancer group. At follow-up, 57 women were dead of ovarian cancer, 40 were disease-free, 25 were alive with disease, and 3 died from other causes. For the group of 57 dead women, there was a significantly lower median survival time with high preoperative GDF-15 (15 months) as compared to low GDF-15 (28 months), p = 0.003 (comparing < 1242 vs. ≥ 1242 ng/L). Preoperative GDF-15 plasma concentration negatively correlated with total survival time among women that had died of ovarian cancer (Spearman's correlation coefficient: −0.56, p < 0.001). As illustrated in Fig. 1 C, survival for the whole ovarian cancer group was better in the group with lowest GDF-15 concentration (43% 5-year survival) as compared to the group with high GDF-15 concentration preoperatively (27% 5-year survival), p < 0.001 (log-rank test). When adjusted for age (dichotomized at the median) and FIGO stage, high GDF-15 remained an independent predictor of short overall survival (Cox regression: RR 2.1, 95% CI 1.2–3.7, p = 0.014). This was also the case if age was entered as a continuous variable (Cox regression: RR 2.0, 95% CI 1.1–3.5, p = 0.025). Results remained unaltered upon excluding the 2 patients with carcinosarcoma and 1 patient with granulosa cell tumor from the survival analyses. Preoperative plasma GDF-15 was also negatively correlated with disease-free survival time (Spearman's correlation coefficient: −0.29, p = 0.001). GDF-15 concentrations in the effusions from 38 women with advanced ovarian carcinoma were high, with median values of 5117 ng/L (minimum–maximum 1547–20,828 ng/L) in ascites ( n = 38) and 5724 ng/L (2048–7721 ng/L) in pleura effusion ( n = 4). There was a close correlation between plasma and ascites GDF-15 concentrations (Spearman's correlation coefficient 0.65, p < 0.001). Median ascites and pleura GDF-15 concentrations were on average 3-fold higher than median plasma GDF-15 in these 38 women with advanced ovarian carcinoma. GDF-15 expression in ovarian serous tumors None of the investigated serous ovarian tumors (cystadenomas, BOT, and adenocarcinoma) showed GDF-15 expression in the endothelial or stromal compartment, whereas tumor cells in all three categories displayed cytoplasmic GDF-15 staining ( Figs. 2 A–F ). Nuclear expression of GDF-15 was not observed. Fig. 2 G shows the extent of cytoplasmic expression of GDF-15 for the three tumor groups, with no significant difference between the tumor groups ( p = 0 .7). In addition to the GDF-15 staining of tumor cells, peritumoral macrophages expressed cytoplasmatic GDF-15 in BOT and carcinomas. Also, one serous ovarian carcinoma showed extracellular GDF-15 staining in a glandular lumen, possibly representing secreted GDF-15. Among women with ovarian carcinoma, nine had tumors investigated with GDF-15 immunohistochemistry in addition to ascites GDF-15 measurement. None of the five women with lowest ascites concentration of GDF-15 had staining in > 5% of tumor cells, whereas two of the four women with highest ascites GDF-15 had a GDF-15 staining in 26-75% tumor cells. Discussion Our study demonstrates that the preoperative plasma concentration of GDF-15 is elevated in women with ovarian cancer as compared to healthy pre- and postmenopausal women and women with benign adnexal tumors or BOT. A major finding of the study was that preoperative GDF-15 predicted survival time for women with ovarian cancer, independent of stage and age. Elevated GDF-15 also predicts cardiovascular deaths in patients with established cardiovascular disease [24–26] as well as predicts an augmented risk of future cardiovascular events in apparently healthy elderly women [8] . However, noncancer deaths were rare ( n = 3) in our ovarian cancer population, and they were censored in the survival analyses. Therefore, GDF-15 emerges as a potentially useful prediction biomarker in ovarian carcinoma. The large overlap in preoperative plasma GDF-15 concentrations between women postoperatively diagnosed with benign ovarian tumor, BOT, and ovarian cancer ( Fig. 1 A) reduces the usefulness of GDF-15 in preoperative ovarian cancer diagnosis. We found, however, that plasma GDF-15 correlated well with a clinically used marker of ovarian cancer, the risk of malignancy index (RMI), previously shown to distinguish benign from malignant adnexal masses [22,23] , by combining ultrasound adnexal tumor score and menopausal state grading with CA 125 serum concentration. The plasma GDF-15 concentrations of the healthy Norwegian postmenopausal women in our study (median 684 ng/L) corresponded to previous findings in elderly healthy Swedish women (median 762 ng/L) [21] , using the same immunoradiometric assay. Notably, none of the healthy pre- and postmenopausal women in our study had a plasma concentration above the median of the ovarian cancer group (1242 ng/L). Hitherto, more than 30 serum tumor markers have been evaluated in combination with CA 125 in tumor marker panels for detecting ovarian cancer, increasing sensitivity by 5 to 10 percentage points over that of CA 125 alone, but with a decline in specificity [ 1 ]. Our study was not designed to evaluate GDF-15 as a screening marker in unselected female populations, as our patient population included women referred for primary surgery due to a diagnosis of ovarian tumor. Serum GDF-15 has previously been shown to be a potential biomarker in the diagnosis of pancreatic [14] , prostatic [3] , and colorectal carcinoma [17] . A GDF-15/PSA (prostate specific antigen) score improves the specificity for detecting prostatic cancer as compared to PSA alone [3] . We are not aware of previous studies exploring the role of GDF-15 as a biomarker in women with gynecological cancer. One previous study demonstrated GDF-15 protein expression by Western blotting in ovarian cancer cell lines. This study also showed that a nonsteroidal anti-inflammatory drug (NSAID) arrested ovarian cancer cell growth in vitro, via induced expression of GDF-15, which acted as a proapoptotic factor, suggesting GDF-15 as a potential mediator of the chemopreventive effect of NSAID in ovarian cancer [27] . We showed by immunohistochemistry that all investigated serous ovarian tumors (benign tumors, BOT, and adenocarcinomas) express GDF-15 protein. We did not detect protein expression of GDF-15 in endothelial cells. GDF-15 has previously been shown to be expressed in other tumor types, including pancreatic adenocarcinomas with cytoplasmic labeling of malignant cells [14] , similar to the findings in our study. Since our immunohistochemistry study was restricted to serous tumors, with a limited number of patients in each diagnostic group, further studies are needed to conclude whether ovarian tumor GDF-15 protein expression by immunohistochemistry correlates with circulating GDF-15 (either plasma or effusion concentrations). Release of GDF-15 from the ovarian tumor to the circulation may not necessarily correlate with tumor protein expression as evaluated by immunohistochemistry and might vary with tissue- and tumor-specific properties, including variations in tumor vascularization and oxidative stress. There may be several mechanisms leading to the elevated plasma GDF-15 concentration demonstrated in our study in women with ovarian cancer. GDF-15 is induced by inflammatory cytokines such as interleukin-1 and tumor necrosis factor-α and is produced by activated macrophages [2] . We found GDF-15 expression in peritumoral macrophages in BOT and ovarian carcinoma. Cancer and inflammation are clearly related [28,29] , and our finding of elevated plasma GDF-15 in ovarian cancer may represent the effect of a more pronounced inflammatory response as compared to the situation in healthy subjects or in women with BOT or benign ovarian tumor. We have previously reported higher preoperative median serum CRP in ovarian cancer as compared to the benign ovarian tumor and BOT groups [19] . This was also the case for the present larger study groups (13.3 vs. 1.3 and 2.0 mg/L, respectively, p < 0.001). As reported in the results section, CRP correlated positively with GDF-15 in the ovarian cancer group, whereas this was not the case for the benign ovarian tumor and BOT groups (data not shown). GDF-15 is strongly induced by the tumor suppressor gene p53 and other antitumorigenic agents, such as NSAID and peroxisome proliferator-activated receptor (PPAR)-γ agonists, suggesting that GDF-15 may be a downstream target of signaling pathways regulating cell cycle arrest and apoptosis [30–33] . GDF-15 has been proposed to mediate anticancer effects, for example, by promoting apoptosis [14,33] . GDF-15 has shown in vitro and in vivo effects on tumor growth and/or apoptosis [33] . GDF-15 is a target gene for p53 and p53 induction of GDF-15 has been demonstrated in vivo and in vitro [14] . In vitro and in vivo studies have shown that GDF-15 production and secretion are related to p53 pathway activation and also that GDF-15 is involved in p53-independent cellular functions [32–34] . On the other hand, GDF-15 has been suggested to facilitate cancer cell detachment, migration, and metastasis, as loss of adhesion of tumor cells has been demonstrated in vitro [11] . It has also been suggested that locally immobilized GDF-15, viewed as an available pool of the bioactive cytokine, may possibly inhibit the dissemination process, whereas decreased availability of GDF-15 may lead to compromised local disease control and subsequent dissemination [11,33,35] . Therefore, the net biological effects of augmented GDF-15 tumor expression, as well as of circulating GDF-15 concentration, on tumor growth and metastasis need to be further explored, and may also vary between cancer forms. In conclusion, preoperative plasma GDF-15 is elevated in women with ovarian cancer and ovarian serous tumors express GDF-15 protein. As preoperative GDF-15 predicted survival time for women with ovarian cancer, independent of age and stage, GDF-15 emerges as a potentially useful prognostic biomarker in ovarian carcinoma. Further longitudinal studies are warranted to assess the value of GDF-15 as a clinical tool in monitoring treatment response and disease progression during follow-up of women diagnosed with ovarian cancer. Also, GDF-15 could be explored as a biomarker for optimizing individual postoperative treatment strategies in ovarian cancer. Conflict of interest statement Drs. Kempf and Wollert have filed a patent and have a contract with Roche Diagnostics to develop a GDF-15 assay for cardiovascular applications. No products/companies are discussed in the article in which any author has an interest. Therefore, there are no potential conflicts of interest. Acknowledgments We are grateful for help in patient recruitment and biobank assistance from Lise Levy and Elin Ødegaard, both Oslo University Hospital. References [1] D.L. Clarke-Pearson Clinical practice. Screening for ovarian cancer N Engl J Med 361 9-7-2009 170 177 [2] M.R. Bootcov A.R. Bauskin S.M. Valenzuela A.G. Moore M. Bansal X.Y. He MIC-1, a novel macrophage inhibitory cytokine, is a divergent member of the TGF-beta superfamily Proc Natl Acad Sci U S A 94 14-10-1997 11514 11519 [3] D.A. Brown C. Stephan R.L. Ward M. Law M. Hunter A.R. Bauskin Measurement of serum levels of macrophage inhibitory cytokine 1 combined with prostate-specific antigen improves prostate cancer diagnosis Clin Cancer Res 12 1-1-2006 89 96 [4] L.N. Lawton M.F. Bonaldo P.C. Jelenc L. Qiu S.A. Baumes R.A. Marcelino Identification of a novel member of the TGF-beta superfamily highly expressed in human placenta Gene 203 5-12-1997 17 26 [5] A.G. Moore D.A. Brown W.D. Fairlie A.R. Bauskin P.K. Brown M.L. Munier The transforming growth factor-ss superfamily cytokine macrophage inhibitory cytokine-1 is present in high concentrations in the serum of pregnant women J Clin Endocrinol Metab 85 2000 4781 4788 [6] M. Sugulle R. Dechend F. Herse M.S. Weedon-Fekjaer G.M. Johnsen K.B. Brosnihan Circulating and placental growth-differentiation factor 15 in preeclampsia and in pregnancy complicated by diabetes mellitus Hypertension 54 2009 106 112 [7] S. Tong B. Marjono D.A. Brown S. Mulvey S.N. Breit U. Manuelpillai Serum concentrations of macrophage inhibitory cytokine 1 (MIC 1) as a predictor of miscarriage Lancet 363 10-1-2004 129 130 [8] D.A. Brown S.N. Breit J. Buring W.D. Fairlie A.R. Bauskin T. Liu Concentration in plasma of macrophage inhibitory cytokine-1 and risk of cardiovascular events in women: a nested case–control study Lancet 359 22-6-2002 2159 2163 [9] J.B. Welsh L.M. Sapinoso S.G. Kern D.A. Brown T. Liu A.R. Bauskin Large-scale delineation of secreted protein biomarkers overexpressed in cancer tissue and serum Proc Natl Acad Sci U S A 100 18-3-2003 3410 3415 [10] W.D. Fairlie A.G. Moore A.R. Bauskin P.K. Russell H.P. Zhang S.N. Breit MIC-1 is a novel TGF-beta superfamily cytokine associated with macrophage activation J Leukoc Biol 65 1999 2 5 [11] T. Liu A.R. Bauskin J. Zaunders D.A. Brown S. Pankhurst P.J. Russell Macrophage inhibitory cytokine 1 reduces cell adhesion and induces apoptosis in prostate cancer cells Cancer Res 63 15-8-2003 5034 5040 [12] T.A. Zimmers X. Jin E.C. Hsiao S.A. McGrath A.F. Esquela L.G. Koniaris Growth differentiation factor-15/macrophage inhibitory cytokine-1 induction after kidney and lung injury Shock 23 2005 543 548 [13] V.M. Paralkar A.L. Vail W.A. Grasser T.A. Brown H. Xu S. Vukicevic Cloning and characterization of a novel member of the transforming growth factor-beta/bone morphogenetic protein family J Biol Chem 273 29-5-1998 13760 13767 [14] J. Koopmann P. Buckhaults D.A. Brown M.L. Zahurak N. Sato N. Fukushima Serum macrophage inhibitory cytokine 1 as a marker of pancreatic and other periampullary cancers Clin Cancer Res 10 1-4-2004 2386 2392 [15] T. Nakamura A. Scorilas C. Stephan G.M. Yousef G. Kristiansen K. Jung Quantitative analysis of macrophage inhibitory cytokine-1 (MIC-1) gene expression in human prostatic tissues Br J Cancer 88 7-4-2003 1101 1104 [16] N.J. de Wit J. Rijntjes J.H. Diepstra T.H. Van Kuppevelt U.H. Weidle D.J. Ruiter Analysis of differential gene expression in human melanocytic tumour lesions by custom made oligonucleotide arrays Br J Cancer 92 20-6-2005 2249 2261 [17] D.A. Brown R.L. Ward P. Buckhaults T. Liu K.E. Romans N.J. Hawkins MIC-1 serum level and genotype: associations with progress and prognosis of colorectal carcinoma Clin Cancer Res 9 2003 2642 2650 [18] F. Weber L. Shen M.A. Aldred C.D. Morrison A. Frilling M. Saji Genetic classification of benign and malignant thyroid follicular neoplasia based on a three-gene combination J Clin Endocrinol Metab 90 2005 2512 2521 [19] E. Odegaard B. Davidson B.V. Elgaaen M.K. Fagerhol V. Engh M. Onsrud Circulating calprotectin in ovarian carcinomas and borderline tumors of the ovary Am J Obstet Gynecol 198 2008 418.e1 418.e7 [20] E. Odegaard B. Davidson V. Engh M. Onsrud A.C. Staff Assessment of endoglin and calprotectin as potential biomarkers in ovarian carcinoma and borderline tumors of the ovary Am J Obstet Gynecol 199 2008 533.e1 533.e8 [21] T. Kempf R. Horn-Wichmann G. Brabant T. Peter T. Allhoff G. Klein Circulating concentrations of growth-differentiation factor 15 in apparently healthy elderly individuals and patients with chronic heart failure as assessed by a new immunoradiometric sandwich assay Clin Chem 53 2007 284 291 [22] I. Jacobs D. Oram J. Fairbanks J. Turner C. Frost J.G. Grudzinskas A risk of malignancy index incorporating CA 125, ultrasound and menopausal status for the accurate preoperative diagnosis of ovarian cancer Br J Obstet Gynaecol 97 1990 922 929 [23] S. Tingulstad B. Hagen F.E. Skjeldestad M. Onsrud T. Kiserud T. Halvorsen Evaluation of a risk of malignancy index based on serum CA125, ultrasound findings and menopausal status in the pre-operative diagnosis of pelvic masses Br J Obstet Gynaecol 103 1996 826 831 [24] K.C. Wollert T. Kempf T. Peter S. Olofsson S. James N. Johnston Prognostic value of growth-differentiation factor-15 in patients with non-ST-elevation acute coronary syndrome Circulation 115 27-2-2007 962 971 [25] T. Kempf H.S. von T. Peter T. Allhoff M. Cicoira W. Doehner Prognostic utility of growth differentiation factor-15 in patients with chronic heart failure J Am Coll Cardiol 50 11-9-2007 1054 1060 [26] T. Kempf J.M. Sinning A. Quint C. Bickel C. Sinning P.S. Wild Growth-differentiation factor-15 for risk stratification in patients with stable and unstable coronary heart disease: results from the AtheroGene study Circ Cardiovasc Genet 2 2009 286 292 [27] J.S. Kim S.J. Baek T. Sali T.E. Eling The conventional nonsteroidal anti-inflammatory drug sulindac sulfide arrests ovarian cancer cell growth via the expression of NAG-1/MIC-1/GDF-15 Mol Cancer Ther 4 2005 487 493 [28] A.J. Schetter N.H. Heegaard C.C. Harris Inflammation and cancer: interweaving microRNA, free radical, cytokine and p53 pathways Carcinogenesis 31 2010 37 49 [29] W. Shan J. Liu Inflammation: a hidden path to breaking the spell of ovarian cancer Cell Cycle 8 1-10-2009 3107 3111 [30] M.K. Agarwal K. Hastak M.W. Jackson S.N. Breit G.R. Stark M.L. Agarwal Macrophage inhibitory cytokine 1 mediates a p53-dependent protective arrest in S phase in response to starvation for DNA precursors Proc Natl Acad Sci U S A 103 31-10-2006 16278 16283 [31] M. Tan Y. Wang K. Guan Y. Sun PTGF-beta, a type beta transforming growth factor (TGF-beta) superfamily member, is a p53 target gene that inhibits tumor cell growth via TGF-beta signaling pathway Proc Natl Acad Sci U S A 97 4-1-2000 109 114 [32] P.X. Li J. Wong A. Ayed D. Ngo A.M. Brade C. Arrowsmith Placental transforming growth factor-beta is a downstream mediator of the growth arrest and apoptotic response of tumor cells to DNA damage and p53 overexpression J Biol Chem 275 30-6-2000 20127 20135 [33] A.R. Bauskin D.A. Brown T. Kuffner H. Johnen X.W. Luo M. Hunter Role of macrophage inhibitory cytokine-1 in tumorigenesis and diagnosis of cancer Cancer Res 66 15-5-2006 4983 4986 [34] M. Albertoni P.H. Shaw M. Nozaki S. Godard M. Tenan M.F. Hamou Anoxia induces macrophage inhibitory cytokine-1 (MIC-1) in glioblastoma cells independently of p53 and HIF-1 Oncogene 21 20-6-2002 4212 4219 [35] A.R. Bauskin D.A. Brown S. Junankar K.K. Rasiah S. Eggleton M. Hunter The propeptide mediates formation of stromal stores of PROMIC-1: role in determining prostate cancer outcome Cancer Res 65 15-3-2005 2330 2336