Clinical proteomics in breast cancer

Abstract

Breast cancer imposes a significant healthcare burden on women worldwide. Early detection is of paramount importance in reducing mortality, yet the diagnosis of breast cancer is hampered by a lack of adequate detection methods. In addition, better breast cancer prognostication may improve selection of patients eligible for adjuvant therapy. Hence, new markers for diagnosis and prognosis are warranted to improve breast cancer care. As proteins can provide the link between gene sequence and cellular physiology, and are readily accessible in biological matrices such as blood, they hold promise as potential biomarkers for cancer. Recent advances in mass spectrometry have enabled fast and simple analysis of the protein content of (clinical) samples. We applied surface-enhanced laser desorption ionisation-time of flight mass spectrometry (SELDI-TOF MS) for the detection of potential biomarkers for breast cancer (BC). In a retrospective study, serum protein patterns comprising, among others, C3a des-Arginine anaphylatoxin, were found that could discriminate patients with breast cancer from healthy persons. However, since sensitivity and specificity were 60-75%, these profiles are presently not suitable for use in general diagnostic screening. To investigate the validity and robustness of reported serum protein profiles for BC, an independent patient population of BC and controls was analysed in our laboratory. None of the (partly) recovered markers was validated, exemplifying analytical (e.g., reproducibility) and statistical (e.g., data overfitting) problems. Furthermore, we found that pre-analytical variables such as storage duration can severely alter serum protein profiles, rendering strict sample handling protocols and assessment of potential confounding by pre-analytical variables of pivotal importance in proteomic investigations. In a retrospective follow-up study, we found the serum haptoglobin phenotype to be a strong and independent predictor of recurrence free survival. These results were, however, not confirmed following validation, emphasising the importance of validation in precluding chance results and erroneous conclusions. A subset of the validation set was subsequently fractionated, to facilitate detection of the low abundant serum proteome. A pattern of three proteins (one of which was identified as an inter-alpha-trypsin inhibitor heavy chain 4 fragment) was found to contain significant prognostic value. For clinical application, the validity of these candidate markers needs to be assessed by (quantitative) analyses of independent study populations. In prospectively collected breast tissue, we detected several proteins (two of which were identified as albumin fragments) which were gradually changing in abundance between control, benign breast disease and BC, or between the different stages of BC. Provided their validation, these proteins can provide insight into the pathophysiological mechanisms associated with, or underlying breast cancer development and aid in diagnosis. In conclusion, overseeing the results of SELDI-TOF MS breast cancer studies up to now, this platform holds promise as a screening tool for biomarker discovery. Yet, its successful application requires further consideration of key issues, e.g., sample handling procedures, enhancement of the dynamic range, reproducibility, and external validation of results. Provided that these studies are performed with adequate statistical power and analytical rigour, they could eventually fulfil the great promise that protein biomarkers have for improving breast cancer care.