olved in initiating lung cancer metastasis, thus making fibronectin an attractive candidate biomarker and therapeutic target. Another promising candidate identified in our study was Afamin, which was found to be up-regulated in the nonprogressing group compared with BPH group, but was downregulated in the progressing and metastatic disease. Afamin is a PubMed ID:http://www.ncbi.nlm.nih.gov/pubmed/22180813 member of the albumin gene family expressed by the liver and kidneys, and has been shown to be a specific binding protein for vitamin E. Interestingly, our data is consistent with a previous study which showed a decrease in afamin levels in ovarian cancer where it has been proposed as a candidate biomarker. Thus, the potential of afamin as a candidate prostate cancer biomarker requires further study. eEF1A1, is a member of the elongation factor AZ-505 site proteins that normally functions to mediate the selection and binding of the aminoacyl-tRNA to the ribosome during protein synthesis, and ensures translational accuracy. eEF1A1 was of particular interest to us as we had previously shown its levels to be increased in higher metastatic variant prostate cancer cells, and another study had reported that down-regulation of eEF1A1 by RNA interference in Du145 cells led to an inhibition of cell proliferation, invasion and migration. Interestingly a truncated form of eEF1A1 known as prostate tumour inducing gene 1 has been shown to be expressed in prostate carcinoma patient-derived blood samples and proposed to be a sensitive biomarker for prostate cancer. Additionally, PTI-1 was shown to be expressed in prostate cancer tissues but not in BPH or normal prostate tissues. Our finding of increased eEF1A1 expression in osteoblasts in the vicinity of metastatic prostate cells, is in line with previous reports indicating a cross-talk between prostate cancer cells and osteoblasts during bone metastasis. Prostate cancer bone metastasis are typically osteoblastic, and it is well known that tumour cells can stimulate osteoblasts to proliferate and differentiate. Thus, eEF1A1 over-expression seen in osteoblasts may occur as a response to the presence of tumour cells. The increased expression of eEF1A1 in the serum of patients with cancer compared with BPH is an important finding as is suggests that the search for novel cancer biomarkers should encompass not only those factors secreted directly by the cancer cells but also `surrogate markers’ produced indirectly as a reaction to the presence of tumour cells. During our analysis we identified and quantified 122 unique proteins. This 
is in line with a previous iTRAQ study analysing serum samples with similar MS instrumentation. Although depletion strategies can improve sensitivity for detection of less abundant proteins, this approach itself is associated with limitations as the higher abundant proteins are known to act as a `sponge’ by complexing with lower molecular weight proteins. For instance, a targeted analysis of alpha-2-macroglobulin binding partners indicated that this highly abundant plasma glycoprotein can bind various cytokines, growth factors and heat shock proteins. Consistent with this, we were unable to detect PSA during our iTRAQ analysis despite it being known to be present at relatively high concentrations in the cancer and metastatic patient groups. PSA is known to complex with alpha1-antitrypsin and alpha-2-macroglobulin, both of which were targets for removal by the IgY-14 affinity column. Thus, it is important that future biomarker studies analyz
