Evaluation of drug effects

Drug toxicity

The study of the toxic effects of compounds on human organs is obviously complicated by the impossibility of experimental test. This has fuelled the development of toxicogenomics resources that describe the effects of toxic compounds in system models that can be in vitro (including human cell lines) or in vivo (including rat and mouse whole organs). A problem remains in evaluating how these models recapitulate toxicity in humans, and which aspects of toxicity they reproduce. We propose an approach to evaluate toxicogenomics models with the literature on human toxicity. The main idea is that a given model system (cell type or animal organ) supports the description of a particular aspect of human toxicity (e.g. causing hair loss) when a set of compounds, already associated with that toxicity in the literature, changes the expression of genes with a particular function in the tested model system (e.g. cell-cycle related genes) [1]. We applied this approach as a proof of concept integrating gene expression profiles for human hepatocytes, rat hepatocytes, rat liver and rat kidney, treated with 33 compounds toxic for liver and kidney (from the Open TC-GATEs database) and biomedical literature annotations (from the PubMed database) to evaluate the suitability of these toxicogenomics models to model different aspects of human toxicity.

In a different approach using the same data as above, we defined a protocol to use collectively the toxicogenomics data of the effects of drugs, concentrations and time points to derive one ranked list of genes affected per model system (rat kidney, rat liver, rat hepatocytes or human hepatocytes) [2]. The protocol is optimized by comparison to the lists of genes mentioned in the literature for kidney or liver toxicity. Some poorly studied genes, like the PCNA clamp associated factor (PCLAF), and genes regulated by the master regulator of the antioxidant response NFE2L2, such as NQO1 and SRXN1, appear as consistent toxicity responders. Interestingly, there are differences between the ranked genes in the models that reflect the aspects of organ toxicity that can be best recapitulated by each of these models.

Another approach to test drug toxicity are organ-on-chips. We contributed to a project to generate a liver-kidney-on-chip using a multifluidic chip to model the effects of drug metabolites. In this application we used two connected chambers, one with hepatic cells (chamber 1, HepG2) and the other with kidney cells (chamber 2, Hek293) [3]. The HepG2 cells expressed proteins of the CYP450 family (involved in liver drug metabolism) and were able to metabolize the drug aflatoxin, inducing toxicity in the Hek293 cells, when the flow was directed from the hepatic cell chamber to the kidney cell chamber. Reversing the flow significantly reduced the toxicity in the kidney cells. Using this chip and cell lines, we screened a library of 200 compounds to find compounds becoming toxic to the cells in chamber 2 (Hek293) after being metabolized by the cells in chamber 1 (HepG2). [di-(1,3-diethylbenzylimidazol-2-ylidene)]gold(I) iodide (MC3), an N-heterocyclic carbene (NHC) gold complex, tested positive [4]. NHC gold complexes are aryl hydrocarbon receptor (AHR) ligands. Further characterization of the effect of MC3 indicated that it upregulated CYP1A1 (and not other members of the CYP450 family) in a process dependant on the activation of the AHR. Expression profiling indicated the activation of TGFß1. In vivo immunosuppressing effects were demonstrated by treating scurfy mice (a mouse model of autoimmune disease), which displayed higher levels of activated AHR. These results explain the mechanisms of gold compounds that have been used as immunosuppressants.

Anti-cancer drug optimization

Most classical anticancer drugs are indiscriminately toxic towards both cancerous and normal dividing cells due to the interference with the fundamental cellular processes of DNA metabolism or mitosis. They are therefore gradually being replaced by drugs targeting molecules and processes more specific to cancer cells. However, targeted drugs often provide marginal clinical benefits compared to classical anticancer drugs. We hypothesized that it should be possible to improve the therapies that use strong but toxic classic anti-cancer drugs by finding targets and drugs with synergistic effects. We applied this approach to find that the efficacy of the classical drug etoposide, a topoisomerase II poison, can be improved by considering treatment-driven gene expression changes [5]. We came to this conclusion by treating AML-derived cell lines with etoposide, which resulted in numerous gene expression changes. Some of these changes mediated the etoposide-driven cell killing, as revealed by mimicking them using drugs targeting proteins encoded by the affected transcripts. Some of the identified targets (BCL2, HDAC, mTOR, and PARP) are undergoing clinical trials in AML and other cancer entities. Crucially, targeting these drivers of etoposide cytotoxicity allowed to lower the etoposide concentration without compromising the AML cell killing efficacy. Altogether, treatment-driven gene expression changes can be utilized to improve the efficacy of classical cancer drugs and their pre-treatment prediction can be of immense value in guided cancer therapy.



[1] Taškova, K., J.F. Fontaine, R. Mrowka and M.A. Andrade-Navarro. 2018. Evaluation of in vivo and in vitro models of toxicity by comparison of toxicogenomics data with the literature. Methods. 132, 57-65.

[2] Taškova, K., J.F. Fontaine, R. Mrowka and M.A. Andrade-Navarro. 2019. Literature optimized integration of gene expression for organ-specific evaluation of toxicogenomics datasets. PLoS One.  14, e0210467.

[3] Theobald, J., A. Ghanem, P. Wallisch, A. Banaeiyan, M.A. Andrade-Navarro, K. Taskova, M. Haltmeier, A. Kurtz, H. Becker, S. Reuter, R. Mrowka, X. Cheng and S. Wölfl. 2017. Liver-kidney-on-chip to study toxicity of drug metabolites. ACS Biomaterials Science & Engineering. 4, 78-89.

[4] Cheng, X., I.L. Shytaj, S. Haberle, R. Gama-Brambila, J. Theobald, S. Ghaffory, J. Wölker, K. Taškova, A.S. Bauer, J.D. Hoheisel, N. Tsopoulidis, O.T. Fackler, A. Savarino, M.A. Andrade-Navarro, I. Ott, M. Lusic, E. Hadaschik and S. Wölfl. 2020. NHC-gold compounds mediate immune suppression through induction of AHR-TGFβ1 signalling in vitro and in scurfy mice. Communications Biology. 3, 10.

[5] More, P., U. Goedtel-Armbrust, V. Shah, M. Mathaes, T. Kindler, M.A. Andrade-Navarro and L. Wojnowski. 2019. Drivers of topoisomerase II poisoning mimic and complement cytotoxicity in AML cells. Oncotarget. 10, 5298-5312.