A proteomic signature of ovarian cancer tumor fluid identified by highthroughput and verified by targeted proteomics

Abstract

Tumor fluid samples have emerged as a rich source for the identification of ovarian cancer in the context of proteomics studies. To uncover differences among benign and malignant ovarian samples, we performed a quantitative proteomic study consisting of albumin immunodepletion, isotope labeling with acrylamide and in-depth proteomic profiling by LC-MS/MS in a pool of 10 samples of each histological type. 1135 proteins were identified, corresponding to 505 gene products. 223 proteins presented associated quantification and the comparative analysis of histological types revealed 75 differentially abundant proteins. Based on this, we developed a panel for targeted proteomic analysis using the multiple reaction monitoring (MRM) method for validation of 51 proteins in individual samples of high-grade serous ovarian tumor fluids (malignant) and benign serous cystadenoma tumor fluids. This analysis showed concordant results in terms of average amounts of proteins, and APOE, SERPINF2, SERPING1, ADAM17, CD44 and OVGP1 were statistically significant between benign and malignant group. The results observed in the MRM for APOE were confirmed by western blotting, where APOE was more abundant in malignant samples. This molecular signature can contribute to improve tumor stratification and shall be investigated in combination with current biomarkers in larger cohorts to improve ovarian cancer diagnosis.

Biological significance

Despite advances in cancer research, ovarian cancer has a high mortality and remains a major challenge due to a number of particularities of the disease, especially late diagnosis caused by vague clinical symptoms, the cellular and molecular heterogeneity of tumors, and the lack of effective treatment. Thus, efforts are directed to better understand this neoplasia, its origin, development and, particularly the identification and validation of biomarkers for early detection of the disease in asymptomatic stage. In the present work, we confirmed by MRM method in individual ovarian tumor fluid samples the regulation of 27 proteins out of 33 identified in a highthroughput study. We speculate that the presence and/or differential abundance observed in tumor fluid is a cooperation primarily of high rates of secretion of such tumor proteins to extra tumor environment that will at the end accumulate in plasma, and also the accumulation of acute-phase proteins throughout the entire body. On top of that, consideration of physiological influences in the interpretation of expression observed, including age, menopause status, route-of-elimination kinetics and metabolism of the tumor marker, coexisting disease, hormonal imbalances, life-style influences (smoking, alcoholism, obesity), among others, are mandatory to enable the selection of good protein tumor marker candidates for extensive validation.

Introduction

Ovarian cancer (OC) is the most fatal gynecologic malignancy with 140,000 deaths per year, the sixth most common cancer worldwide among women, and is the leading cause of mortality among gynecological malignancies in developed countries [1], [2]. Due to vague clinical symptoms, it is difficult to detect ovarian cancer at an early stage, in which the 5-year survival rate is approximately 92%. Therefore, most patients will be diagnosed with advanced stage disease with wide peritoneal metastasis (approximately 75% stages III or IV), in which the 5-year survival rate is only 30% [3], [4], [5], [6].

Moreover, OC is a heterogeneous disease at the cellular and molecular levels, and comprises a variety of tumors with diverse histopathological features and biological behavior [7], [8]. Since ovarian cancer cells with various histological types may express tumor markers differently, it is important to use multiple tumor markers to detect all ovarian cancers [9]. On the other hand, and because of such heterogeneity, the study of each histologic type separately, may present the most significant advance in understanding the disease [10]. Tumor markers, including a diversity of substances like cell surface antigens, cytoplasmic proteins, enzymes, hormones, oncofetal antigens, receptors, oncogenes and their products, are assuming a growing role in different aspects of cancer care, whether for diagnosis, screening, prediction or treatment monitoring [11], [12].

The discovery of such tumor markers depends on suitable biological samples and an appropriate methodology. Different types of biological samples can be used in cancer research, each one having advantages and disadvantages. Alternatively, our group and others have also explored the use of intra-tumor fluids, which provides information about the protein expressed by the tissue, enabling the detection of low abundance secreted proteins (compared with human plasma) [10], [13], [14], [15]. Therefore, the use of OC intra-tumor fluids as a sample for tumor marker studies allows the detection of panels of potential targets that will fall into the circulation and could truly represent candidate biomarkers that can be validated in other types of samples, such human plasma.

The increasing interest in applying proteomics methods to study OC have been contributing to assist the understanding the pathogenesis of cancer, elucidating the mechanism of drug resistance and in the development of biomarkers for ovarian cancer detection [10], [14], [16], [17], [18], [19], [20]. Analysis of protein sets, also denoted as molecular signatures, is a promising approach to adequately grasp biological complexity and to efficiently contribute for the diagnose complex diseases, monitor physiological changes and successfully develop new drugs [21]. To contribute to the elucidation of new sets of OC tumor markers, we combined highthroughput and targeted proteomic analysis to discover and verify panels of differentially abundant proteins in ovarian tumor fluid samples (Fig. 1).