The microenvironment in human ovarian carcinoma - characterization through proteomic analysis of tissue interstitial fluid

Abstract

The interstitial fluid of tissues and that of tumors in particular, represents the secretome and is thus a valuable source of tissue-specific proteins. The differential expression of proteins that are secreted in tumor tissue can reflect specific characteristics of the tumor biology. To determine changes in the tumor secretome, access to fluid that reliably reflects the local microenvironment is essential, and can enable us to identify proteins that can be used to monitor or treat disease. Substantial efforts have gone into the search for tumor-specific proteins, mainly by using proteomic strategies, but it remains a challenge to go beyond lists of candidate proteins to clinical relevant biomarkers. The ability to translate proteomic findings to the clinic is hampered by the low capacity in verification of suggested biomarker candidates. To be able to characterize changes in the tissue microenvironment of human tumors, access to the tumor interstitial fluid (TIF) is necessary. We investigated whether the fluid isolated from ovarian carcinoma tissue by centrifugation was representative for interstitial fluid. Determination of creatinine and Na+ in isolated fluid and plasma showed that the ex vivo admixture from the intracellular compartment is limited, and we concluded that the isolated fluid was representative for undiluted TIF. Although the tumor microenvironment has received increased attention in the recent years, studies on TIF from human tissue samples are limited. We summarized the recent progress in the use of TIF for proteomic analysis. When comparing data from studies using TIF as a substrate there were substantial differences between proteomes of the same cancer, as well as between proteomes from different cancer types. These differences may be due to the choice of isolation technique as well as proteomic strategy. Major challenges reside in the lack of available techniques for validating the origin of the isolated fluid and hence of proteins in the sample. The assumption that IF has a higher concentration of proteins produced in the tumor compared to plasma is widespread. By assessing the known biomarker CA-125 in TIF and corresponding plasma we could demonstrate that this assumption is true. This underlines the advantage of using TIF as a substrate for proteomic analysis. Furthermore, the gradient between TIF and plasma was dependent on stage, and the concentration of CA-125 in TIF may harbor additional information of relevance for differential diagnosis or prognosis for ovarian cancer. Using IF as a substrate we have analyzed the proteomes of both healthy and malignant gynecological tissues. By extensive fractionation before mass spectrometry analysis, we were able to detect low abundance proteins that are up-regulated in the tumor microenvironment. Furthermore, variation in protein concentration in individual samples was assessed, and we found substantial heterogeneity between individual tumors. Validation by targeted MS and antibody-based techniques gave similar results, and label-free quantification in unfractionated individual samples indicated that the actin-related protein WD repeat-containing protein 1 can have a central role in tumor progression. The extensive Orbitrap proteomes produced from individual samples of both healthy and malignant gynecological tissues can harbor a number of other tumor-specific proteins. The exclusion of proteins from parts of the tumor microenvironment has been studied in animal models, but has not earlier been determined in human tumors. We quantified abundant plasma proteins by mass spectrometric analysis of unfractionated and undiluted interstitial fluid. Fifteen proteins with varying molecular weight and pI were used as probes to determine the relation of available distribution volume with molecular weight and charge in healthy and malignant gynecological tissues. We found that the fractional available distribution volume of albumin was significantly increased in ovarian carcinomas compared with healthy ovarian tissue. Furthermore, the available distribution volume of large plasma proteins was dependent on molecular weight in healthy ovarian and endometrial carcinoma tissues, but not in ovarian carcinoma tissue. Subsequently, we quantified the composition of extracellular matrix components, showing a high concentration of collagen, and low concentration of hyaluronan in healthy compared with malignant gynecological tissues. In conclusion, we have demonstrated the advantage of using TIF for biomarker discovery in humans, as well as to characterize the tumor interstitium by measuring the available distribution volume of proteins. Proteomic analysis of TIF can have the potential to improve differential diagnosis, prognosis and choice of treatment, and reveal possible targets for therapeutic intervention for ovarian and endometrial cancer.