In vitro analysis of drug‐induced protein mislocalization and accumulation

Abstract

Unexpected protein inclusions in rat proximal tubule epithelial cells are a reoccurring phenomenon in non‐clinical safety studies that might interfere with the drug development outcome, even when relevance to humans has not been assessed. Propiverine, a frequently‐prescribed medication for diverse bladder control diseases, gained only limited marketing authorization due to large cytosolic and intranuclear protein inclusions in rat kidney. Although it was shown that these protein inclusions mainly consist of the peroxisomal enzyme D‐amino acid oxidase (DAAO), the role and relevance of DAAO accumulation in humans remain unknown. This thesis aimed to provide a better understanding of the principle mechanism(s) of propiverine‐induced DAAO mislocalization and accumulation. For that purpose, a new in vitro model of EYFP‐DAAO expressing human and rat renal cell lines was established, thoroughly characterized and successfully applied to analyze the protein interaction network of DAAO, manipulate subcellular trafficking, and monitor protein dynamics in vitro. Although nuclear localization of a peroxisomal protein appears to be exceptional, manipulation of diverse active nuclear transport mechanisms did not change DAAO localization per se. In contrast, interference with the peroxisomal trafficking, i.e. deletion of the peroxisomal targeting signal type 1 (PTS1) or PEX5 knock‐down, resulted in passive diffusion of EYFP‐DAAO into the nucleus. The latter not only contradicts the often‐cited nuclear diffusion limit of 40 kDa, but also supports the hypothesis that propiverine and/or its metabolite(s) may interfere with peroxisomal transport and/or import, thereby provoking nuclear (mis)localization. Subsequently, synthesizing a biotin‐modified propiverine derivative for a Co‐IP coupled LC‐MS/MS analysis allowed for the identification of propiverine‐specific interactors from rat kidney homogenate. Besides binding to chaperones, that are DAAO interaction partners and involved in peroxisomal trafficking, propiverine was found to primarily interact with DAAO and with other PTS1 proteins. Indeed, several PTS1 proteins, the PTS1 receptor PEX5, and the peroxisome biogenesis factor PEX3 were found to mislocalize and/or accumulate to varying degrees after propiverine treatment in vivo. However, neither PTS2 nor peroxisomal membrane proteins (PMPs) were affected by propiverine exposure in rats. Although a direct interference with DAAO might abrogate peroxisomal trafficking, evidence was provided that the nuclear presence of DAAO is rather an inherent characteristic of peroxisomal proteins. The latter was supported by the fact that DAAO was found to be mainly degraded via the nuclear proteasome and upon proteasomal inhibition accumulates primarily within the cell’s nuclei. This and the finding that propiverine directly interacts with the protein degradation machinery, e.g. several proteases and the regulatory subunits of the 26S proteasome, might eventually provoke the formation of nuclear DAAO accumulations. Moreover, in contrast to previous findings, a two‐fold increase in peroxisome abundance was demonstrated in propiverine‐treated rats. An increased peroxisome biogenesis due to accumulating PEX3 might compensate for import‐deficient peroxisomes, thus preventing acute toxicity in preclinical trials. In conclusion, this work shows that propiverine provokes a more general and broader dysregulation of peroxisome homeostasis and peroxisomal protein trafficking than initially thought. Propiverine most likely features a dual effect on the localization and the degradation of PTS1 proteins in rat kidney, i.e. hindrance of peroxisomal import by direct interaction with DAAO (and other PTS1 proteins) and diminished protein homeostasis due to interference with the protein degradation machinery and various chaperones. As neither propiverine nor its main hepatic metabolite changed DAAO localization in vitro, we conclude that organ‐ and/or species‐specific metabolites should be investigated in future studies by using in vitro metabolism studies in conjunction with the here established cell models.