Refine
Has Fulltext
- yes (2)
Is part of the Bibliography
- yes (2)
Document Type
- Doctoral Thesis (2)
Language
- English (2)
Keywords
- pharmacokinetics (2) (remove)
Institute
- Institut für Pharmakologie und Toxikologie (2) (remove)
Kinetic assessment by in vitro approaches - A contribution to reduce animals in toxicity testing
(2015)
The adoption of directives and regulations by the EU requires the development of alternative testing strategies as opposed to animal testing for risk assessment of xenobiotics. Additionally, high attrition rates of drugs late in the discovery phase demand improvement of current test batteries applied in the preclinical phase within the pharmaceutical area. These issues were taken up by the EU founded 7th Framework Program “Predict-IV”; with the overall goal to improve the predictability of safety of an investigational product, after repeated exposure, by integration of “omics” technologies applied on well established in vitro approaches. Three major target organs for drug-induced toxicity were in focus: liver, kidney and central nervous system. To relate obtained dynamic data with the in vivo situation, kinetics of the test compounds have to be evaluated and extrapolated by physiologically based pharmacokinetic modeling.
This thesis assessed in vitro kinetics of the selected test compounds (cyclosporine A, adefovir dipivoxil and cisplatinum) regarding their reliability and relevance to respective in vivo pharmacokinetics. Cells were exposed daily or every other day to the test compounds at two concentration levels (toxic and non-toxic) for up to 14 days. Concentrations of the test compounds or their major biotransformation products were determined by LC-MS/MS or ICP-MS in vehicle, media, cells and plastic adsorption samples generated at five different time-points on the first and the last treatment day.
Cyclosporine A bioaccumulation was evident in primary rat hepatocytes (PRH) at the high concentration, while efficient biotransformation mediated by CYP3A4 and CYP3A5 was determined in primary human hepatocytes (PHH) and HepaRG cells. The lower biotransformation in PRH is in accordance with observation made in vivo with the rat being a poor model for CYP3A biotransformation. Further, inter-assay variability was noticed in PHH caused by biological variability in CYP3A4 and CYP3A5 activity in human donors. The inter-assay variability observed for PRH and HepaRG cells was a result of differences between vehicles regarding their cyclosporine A content. Cyclosporine A biotransformation was more prominent in HepaRG cells due to stable and high CYP3A4 and CYP3A5 activity. In addition, in vitro clearances were calculated and scaled to in vivo. All scaled in vitro clearances were overestimated (PRH: 10-fold, PHH: 2-fold, HepaRG cells: 2-fold). These results should be proven by physiologically-based pharmacokinetic modeling and additional experiments, in order to verify that these overestimations are constant for each system and subsequently can be diminished by implementation of further scaling factors.
Brain cell cultures, primary neuronal culture of mouse cortex cells and primary aggregating rat brain cells, revealed fast achieved steady state levels of cyclosporine A. This indicates a chemical distribution of cyclosporine A between the aqueous and organic phases and only minor involvement of biological processes such as active transport and biotransformation. Hence, cyclosporine A uptake into cells is presumably transport mediated, supported by findings of transporter experiments performed on a parallel artificial membrane and Caco-2 cells. Plastic adsorption of cyclosporine A was significant, but different for each model, and should be considered by physiologically based pharmacokinetic modeling.
Kinetics of adefovir dipivoxil highlights the limits of in vitro approaches. Active transporters are required for adefovir uptake, but were not functional in RPTECT/TERT1. Therefore, adefovir uptake was limited to passive diffusion of adefovir dipivoxil, which itself degrades time-dependently under culture conditions.
Cisplatinum kinetics, studied in RPTEC/TERT1 cells, indicated intracellular enrichment of platinum, while significant bioaccumulation was not noted. This could be due to cisplatinum not reaching steady state levels within 14 days repeated exposure. As shown in vivo, active transport occurred from the basolateral to apical side, but with lower velocity. Hence, obtained data need to be modeled to estimate cellular processes, which can be scaled and compared to in vivo.
Repeated daily exposure to two different drug concentrations makes it possible to account for bioaccumulation at toxic concentrations or biotransformation/extrusion at non-toxic concentrations. Potential errors leading to misinterpretation of data were reduced by analyses of the vehicles as the applied drug concentrations do not necessarily correspond to the nominal concentrations. Finally, analyses of separate compartments (medium, cells, plastic) give insights into a compound’s distribution, reduce misprediction of cellular processes, e.g. biotransformation, and help to interpret kinetic data. On the other hand, the limits of in vitro approaches have also been pointed out. For correct extrapolation to in vivo, it is essential that the studied in vitro system exhibits the functionality of proteins, which play a key role in the specific drug induced toxicity. Considering the benefits and limitations, it is worth to validate this long-term treatment experimental set-up and expand it on co-culture systems and on organs-on-chips with regard to alternative toxicity testing strategies for repeated dose toxicity studies.
This study should contribute to the important field of pharmacogenetics by: firstly, establishing an easy and safe phenotyping method that combines the activity determination of all three previously mentioned CYPs (CYP2D6, CYP2C9, and CYP2C19) into one phenotyping cocktail and secondly, improving the knowledge about the predictive power of the genotype for the measured phenotype. It was indeed possible to develop a save, easy-to-use, fast and simultaneous phenotyping procedure for the important genetic polymorphic enzymes CYP2D6 and CYP2C9. To accomplish that, interaction studies with the chosen probe drugs dextromethorphan (DEX, CYP2D6), flurbiprofen (FLB, CYP2C9) and omeprazole (OME, CYP2C19) were conducted. It could be proven that DEX and FLB can be administered in combination, whereas OME alters the phenotyping results of CYP2C9. This is a new finding as in 2004 a phenotyping cocktail was published that used FLB and OME in combination. However, to our knowledge, no interaction tests were carried in that study. The new phenotyping procedure is not only verified by prior probe drug interaction studies, it also has other advantages over phenotyping cocktails found in literature. Firstly, save probe drugs are used in very small doses. This is possible due to the new sensitive LC-MS/MS methods that were evaluated. Secondly, the new phenotyping procedure is very fast and on-invasive. Urine has to be collected only for 2 h and the results also suggest that the time consuming glucuronide cleavage of the CYP2D6 dependent metabolite dextrorphan, usually carried out before CYP2D6 phenotyping, may be unnecessary. Most importantly, however, new insights into the phenotype prediction from genotype for CYP2C9 and CYP2D6 could be gained within this study. Nearly 300 phenotyped Caucasian subjects were also genotyped for the most important known variant alleles for CYP2D6, CYP2C9 and CYP2C19 using several established and newly developed genoptyping methods. Therefore, a direct correlation between phenotype and genotype could be conducted for CYP2D6 and CYP2C9. Employing linear modeling, it was possible to assign activity coefficients to each of the detected CYP2D6 and CYP2C9 alleles, thereby estimating their contribution to the resulting enzyme activity. This might facilitate the prediction of the CYP2D6 and CYP2C9 metabolic status of a subject knowing only its respective genotypes. Especially the new CYP2D6 genotype phenotype correlation model might allow for more precise phenotype prediction for the included variant alleles than was possible until now. Taken together, this study substantially contributes to the important research field of pharmacogenetics by (i) developing a save and easy-to-use phenotyping combination for CYP2D6 and CYP2C9, and (ii) by establishing activity coefficients for each of the detected CYP2D6 and CYP2C9 alleles, thereby allowing for a more precise prediction of the phenotype from genotype.