DRUG REACTIONS

20

MidJIs for p~mbtg toxicigr: aRPlicatioDS in drug disc~veJ:y and development -Rosie Stather-

Unacceptable levels of toxicity can lead to the abandonment of drugs at any stage of the developmental process, whether it be during drug discovery, preclinical testing or clinical trials. Investigational toxicological methods using models and hypotheses can often resolve problems and enable chemical templates or therapeutic discovery programmes to be salvaged. An example of this is the use of human hepatocytes to predict the potential toxicity of a drog in humans, or to elucidate the mechanism of toxicity once it has been observed. The integration of investigative toxicology in the drug discovery and developmental process was discussed at the 7th International Congress of Toxicology [Seattle, US; July 1995]. In the majority of cases, the identification and testing of a potential drug candidate leads to toxicological issues being raised. There are 2 steps that can be taken at this stage: the development of the molecule can be abandoned; or studies can be undertaken in an attempt to discover what is causing the toxicity and to determine if it can be prevented. According to Dr R Ulrich from Upjohn, Kalamazoo, US, mechanistic toxicology allows researchers to determine the relevance of toxicity seen in animal models to that seen in humans. It can also provide information regarding the mechanism behind the toxicity. There are 2 possible mechanisms of toxicity that can be investigated: the biological mechanism (the effect of the compound on the cell or organism), and the chemical mechanism (discovering which moiety of the molecule is responsible for causing the toxicity).

Investigating the use of hepatocyte models The first step in any investigative problem-solving exercise is to develop animal or cell models that accurately reflect the situation in humans. The liver is the target organ for many toxicological investigations, since it is the organ most often affected by toxicity. By using hepatocytes from humans and animals, Dr Ulrich and colleagues have explored several different drug-induced hepatic disorders that became apparent at various stages of drug discovery and development. These disorders include: • quinoxaline-induced metabolic inhibition • phospholipid storage disorders induced by amphiphilic amines and antibacterial agents • peroxisome proliferative disorders induced by lipid-lowering agents.

Human cell models Models that use human hepatocytes are of particular interest, as the use of human hepatocytes can help 'bridge the gap' between toxicity seen in animals and that seen in humans [see figure]. Human hepatocytes were used as an in vitro model for the toxicological testing of a nonbenzodiazepine anxiolytic compound, panadiplon. This compound had not shown any signs of toxicity in preclinical studies in rats and monkeys. However, when panadiplon was tested in volunteers, there was evidence of hepatotoxicity. Human hepatocytes were then used to investigate the reasons for panadiplon's hepatotoxicity in humans. Dutch Belted rabbits were also found to experience hepatotoxicity with panadiplon, and therefore these animals were used as the in vivo model for investigating panadiplon-associated toxicity.

15 JuI1995INPHARMA'"

As well as helping to elucidate the mechanisms behind observed toxicities, human cell culture models contribute to the drug discovery effort by permitting the examination of small quantities of analogues to help establish structure/activity relationships. Advantages of using human hepatocytes as a model for investigating toxicity Toxicity observed in animal model

Toxicity observed

in humans

......... I

~ Studies in human hepatocyte.

Help 'bridge the gap' between animal and human responses

Identify potential toxicity in human.

I

Help in elucidating mechanisms of observed toxicity

Eliminate problem, compound or subgroup of molecule

Profiling the toxicity ofthioxanthones Investigational toxicologists from Sanofi, US, have conducted studies using hepatocytes and bone marrow stem cells to successfully profile the hepatotoxicity and myelotoxicity of discovery-stage thioxanthone antitumour agents. Dr AI Roberts from Sanofi explained that the data gained from in vitro hepatotoxicity studies, and in vivo and ex vivo myelotoxicity studies, were integrated into the database of information used by their drug development team to select suitable thioxanthones for advancement. The thioxanthones are a series of cytotoxic antineoplastic agents that have shown marked antitumour activity against a range of murine solid tumours. The original compound in the series, hycanthone (first developed in the 1970s to treat schistosomiasis) was advanced into phase II trials as an anti tumour agent. However, the phase IT trials with this agent were stopped because of a lack of efficacy and evidence of hepatotoxicity.

Accwnulating data to guide development In the 1980s, Sanofi began to test a new series of thioxanthones. Criteria for the development of these second-generation thioxanthones included greater efficacy than hycanthone, a lack of hepatotoxicity, and an acceptable level of myelotoxicity. A lead compound, WIN-33377 [Sanofi-Winthrop], was selected from this new series to begin preclinical testing. WIN-33377 is now in phase I trials for the treatment of solid tumours, and phase II studies are expected to begin later this year. 0156-270319510995-o0020I$01,Oao Adls International Limited 1995. All rights reserved

21

DRUG REACTIONS The role of investigational toxicologists in the evolution of the second-generation thioxanthones was to provide the drug development teams with toxicity data to guide the developmental process. An in vitro hepatotoxicity model that uses primary cultures of murine hepatocytes was developed. This in vitro model was validated using data from in vivo studies in mice - the in vitro test successfully identified the hepatotoxic status of 8/9 test compounds. Therefore, the in vitro model could be used to eliminate thioxanthone compounds that were hepatotoxic and unsuitable for preclinical testing.

Myelotoxidty not problematic with WIN-33377 Myelotoxicity was found to be a dose-limiting toxicity for WIN-33377 in animal studies. In vivo studies in mice have shown that thioxanthones can reduce WBC counts as early as 1 day after drug administration. Furthermore, ex vivo assays of bone marrow stem cells showed that stem cell counts were also affected by day 1 after administration. However, other studies using human bone marrow indicated that this tissue was relatively insensitive to myelotoxicity. Thus, with respect to the thioxanthones, myelotoxicity data was used to 'rank' compounds in terms of their myelotoxic effects, whereas hepatotoxicity data was used to indicate a 'go' or a 'no go' situation for a particular compound.

In animal studies, both pravastatin and lovastatin have been shown to induce skeletal myopathy. Using cultures of rat myocytes, Dr Flint and colleagues were able to show that lovastatin was IOO-fold more toxic than pravastatin. In addition, by conducting experiments where cholesterol biosynthesis pathways were modified, Dr Flint's group demonstrated that reduced cholesterol synthesis by lovastatin and pravastatin was not the cause of the myotoxicity seen with these agents. 800314536

~ Ediloritd comment: There have been case reports of myopathy and rhabdomyolysis occurring in patients treated with the HMG eoA reductase inhibitors pravastatin, lovastatin or simvastatinfor hyperlipidaemia.

Figuring out the mechanisms of drug-induced toxicity Dr Oliver Flint and colleagues from Bristol-Myers Squibb in Syracuse and Princeton, US, have used in vitro techniques to investigate 2 specific examples of drug-induced toxicity: • nucleoside analogue-induced cardiotoxicity • HMG CoA reductase inhibitor-induced myotoxicity. In preclinical studies, oral or IV fluorodideoxyadenosine (FddA) induced cardiotoxicity when administered at pharmacologically relevant doses. In contrast, extremely high doses of fluorodidanosine (fluorodideoxyinosine) were required to produce similar effects. Furthermore, coformycin, an adenosine deaminase inhibitor that prevents the conversion of adenosine to inosine, increased the toxicity of fluorodideoxyadenosine. In addition, fluorodideoxyadenosine reduced the beat rate of rat embryo cardiomyocytes. This effect was abrogated by the adenosine Al receptor antagonist l,3-dipropyl-8-cyc1opentylxanthine (DPCPX), but reduction of cardiomyocyte protein synthesis was not ameliorated by DPCPX. These fmdings suggest that factors other than adenosine receptor binding are involved in fluorodideoxyadenosine-induced cardiotoxicity. 'Thus, the in vitro results mimic the observed in vivo toxicity, indicating that the marked cardiotoxicity seen with FddA [fluorodideoxyadenosine J is a class effect limited to adenosine and its analogues' , concluded Dr Flint.

0156·270319510995-000211$01.00° Adlslntematlonal Limited 1995. All rights reserved

INPHARMA'"15 Ju11995