Evaluation of biomarkers for testicular toxicity

Non-clinical safety assessment is essential during the drug development process in the pharmaceutical industry, and involves numerous, detailed in vitro and in vivo toxicology tests (general, reproductive and genetic), and safety pharmacology studies. The testis is a common organ for adverse drug effects leading to attrition of potential compounds. It would, therefore, be useful to detect testicular toxicity as early as possible in the drug development process. Histopathology is the standard method for assessing testis toxicity, but a biomarker for ‘early warning’ detection of testicular toxicity would be far more useful in non-clinical toxicology studies. The aim of this thesis was to evaluate the feasibility of this approach. It is thought that proteins can leak from seminiferous tubules into testicular interstitial fluid following testicular damage, due to either loss of integrity of the blood-testis barrier (BTB) or germ cell damage. A potential biomarker protein could, therefore, leak out of seminiferous tubules into interstitial fluid and then into blood following toxicological insult to the testis. A suitable biomarker protein must be testis specific, abundant, and not normally be present in blood. It may also need to have a low molecular weight. To investigate if proteins do leak out of seminiferous tubules following testicular damage, three known testicular toxicants which affect different aspects of the testis were used; cadmium chloride causes disruption to the blood-testis barrier and spermatogenesis, methoxyacetic acid (MAA) specifically causes a loss of pachytene spermatocytes, and 1,3-dinitrobenzene (DNB) causes Sertoli cell vacuolation and subsequent germ cell disruption. Adult male Wistar rats were treated with various doses of these toxicants to give mild and moderate responses. Samples were collected 24 hours later. Testicular damage was investigated by immunohistochemistry for well-known germ cell markers (DAZL, VASA) and using a general antibody to seminiferous tubule proteins. The integrity of the BTB was evaluated using immunofluorescent co-localisation of occludin, ZO-1, claudin-11, N-cadherin and β-catenin, and a biotin tracer. Protein leakage was investigated using analysis of interstitial fluid samples by 1D gel electrophoresis and staining with Coomassie-based dye or Western blotting for germ cell proteins and with the general antibody to seminiferous tubule proteins. Protein leakage from seminiferous tubules into interstitial fluid was observed with high dose cadmium chloride treatment. This was coincident with a loss of integrity of the BTB. No leakage was observed with MAA treatment which caused a specific loss of pachytene spermatocytes, or DNB which caused Sertoli cell vacuolation. With both treatments the BTB did not appear to be damaged suggesting that protein leakage occurs only following loss of integrity of the BTB. This was further investigated using treatments reported to specifically disrupt the BTB, namely intra-testicular administration of glycerol or transforming growth factor-β3, with samples collected 48 hours later. The damage caused was very localised, although BTB disruption with glycerol treatment caused some protein leakage. The presence of germ cell proteins in interstitial fluid samples before and after the development of the BTB during normal development was also evaluated, although most proteins of interest were not expressed in germ cells of the immature testis before BTB formation. Finally, five potential biomarker candidate proteins (ADAM3, Calpastatin, DAZL, FABP9, VASA) were selected and investigated using samples from the testicular toxicant studies. Smaller molecular weight proteins were thought to be more likely to leak out of seminiferous tubules, however, VASA, a large molecular protein (76kDa) was shown to leak into interstitial fluid following high dose cadmium chloride treatment. However, FABP9 (low molecular weight) was found to be the most promising biomarker for loss of BTB integrity. The results suggest that a biomarker could only be detected if there is a loss of integrity of the BTB and severe disruption of spermatogenesis, thus conferring no real advantage over present histopathology-based toxicity evaluations. Therefore, an automated immunohistochemistry and image analysis method was investigated as a refined method for detection of testicular toxicity at the end of a toxicology study, and shown to have promise