Developing new paradigms for overcoming drug resistance in cancer using novel humanised mouse models

Although there have been significant advances in the development of new treatments for cancer, where the drugs are now targeted at the genetic changes which drive tumour growth, major challenges still exist in translating this into long-term patient survival. In addition, drug side-effects particularly of new combination treatments, continue to be a major problem. The reason why the often dramatic patient responses, have not resulted in sustained long-term survival is because the tumours rapidly become resistant to the high concentrations of these very potent drugs.

Whereas there is currently an enormous emphasis on cancer genomics, tumour cell heterogeneity, microenvironment, etc., there are essentially no studies on drug pharmacology i.e. how to use the new drugs optimally. Defining the pharmacology of emerging targeted anti-cancer drugs, rather than using dosing regimens based on MTD, has to be a fundamental component of developing new effective treatments for cancer, particularly for drug combinations. Understanding how best to use targeted agents can make the difference between success and failure of potentially promising new treatment regimens.

Historical evidence, based particularly on the treatment of infectious diseases, has taught us that drug resistance can be overcome by either using combinations of drugs concomitantly, or by altering the drug dosing regimen. A major challenge for the improved treatment of cancer is therefore which combination of drugs should be used and what is the dosing schedule? Based on the enormous number of emerging targeted anti-tumour agents, it is impossible to address this challenge by clinical trial alone, and new pre-clinical models which are more predictive of patient responses are urgently needed.

An important difference between animal models and humans is how drugs are handled by the body. This results in different rates of drug exposure and elimination and different drug metabolites. To address this problem, we have made a number of genetically altered mouse models in which drugs are metabolised in a manner similar to that observed in humans. It is the aim of this project to establish the potential of this model system to predict drug responses in man and to use this model to test whether drugs are still effective, alone or in combination, at doses which will be tolerated by patients. Our ultimate aim is to develop a computer-based model which will be used for the informed design of combination clinical trials which delay the onset of drug resistance. If successful, the application of the model systems developed in this programme could be of enormous benefit for the improved treatment of cancer, both in terms of patient survival and reduction in drug side-effects. The model validated in this programme could be of great for the development of new drugs, not only for the treatment of cancer but also for the treatment of other human diseases.

Whereas there is currently an enormous emphasis on cancer genomics, tumour cell heterogeneity, microenvironment, etc, there are essentially no studies on drug pharmacology i.e. how to use the new drugs optimally. Defining the pharmacology of emerging targeted anti-cancer drugs, rather than using dosing regimens based on MTD, has to be a fundamental component of developing new effective cancer treatments, particularly for drug combinations which are being developed to overcome drug resistance. Understanding how best to use targeted agents can make the difference between success and failure of new treatment regimens.
Almost all new targeted anti-cancer agents are substrates for the cytochrome P450 system, and the transcription factors (CAR, PXR) which regulate their expression. These enzymes play a major role in circulating & tumour drug concentrations, and the generation of pharmacologically active and/or toxic metabolites. Major species differences in this enzyme system can seriously confound extrapolation of data from animals to man. To circumvent this, we have created a mouse where the major human P450s and CAR & PXR, have replaced their murine counterparts so that Phase I drug metabolism is essentially that seen in man.
We will use these models to investigate the pharmacological drivers of drug resistance. We will use melanoma as a model since dramatic patient responses are observed with the current standard of care (dabrafenib and trametinib), and the emergence of drug resistance is a rapid and major problem. We will use the humanised models to understand PK, PD and efficacy relationships and establish the correlation between drug exposure, target inhibition, pathways of cell death and drug efficacy. These data will be used to develop a computational model which allows the relationship between drug exposure and efficacy of drug combinations at different drug exposures to be predicted and therefore facilitate the informed design of combination clinical trials.

This programme will impact on a number of medically important areas. These include defining new paradigms for the treatment of cancer and in drug development and use.

Despite recent advances in cancer treatment using targeted anti-tumour agents, the prognosis for advanced disease is poor and toxicity and drug resistance remain as major clinical problems. There is thus an increasing focus on using drug combinations for cancer treatment. Since the standard of care for most drugs is close to maximum tolerated dose (MTD), and the design of combination clinical trials remains essentially empirical, there is an urgent need for better model systems to provide an evidence-based approach for the prioritisation of drug combination treatments. Continued development of novel targeted anti-cancer agents without recourse to pharmacological principles not only exposes patients to unnecessary toxicities, but encourages the development of drug resistance thus impairing patient well-being and reducing the potential long-term benefits of new drug treatments. Our work will validate and utilise novel experimental approaches to develop new paradigms to overcome drug resistance and facilitate the informed design of combination clinical trials. The number of current and emerging targeted anti-cancer drugs exceeds by orders of magnitude the ability to test drug combinations by clinical trial (Al-Lazikani et al., 2012 Nat.Biotechnol, 30,679).

This programme will create novel model systems and in silico models for the informed design of clinical trials using drug combinations aimed at overcoming drug resistance. Although we are using melanoma as a model system, the results can be applied to the treatment of other cancers. This would represent an enormous step forward, generating an experimentally-based system for the prediction of patient outcome and moving away from the current empiricism and use of MTD, i.e. reducing patient morbidity associated with cancer therapy.

We will also validate a unique mouse model humanised for pathways of drug disposition which would have wide application in drug development for optimising drug efficacy and predicting drug-drug interactions.

Primary beneficiaries of the research will be cancer patients receiving treatment with new, molecularly targeted agents, including drug combinations. The in silico model developed will benefit clinicians and clinical trial designers, while the NHS will also be a major beneficiary. These studies are potentially of enormous benefit to the pharmaceutical industry, allowing them to design clinical trials in an informed manner and reduce drug attrition, and to predict drug combinations which have the highest probability of success.

Cancer patients will benefit by receiving optimised treatment regimens, including drug combinations, at doses and schedules that reflect pharmacological principles, rather than MTD. This will have the advantage of maintaining efficacy in the face of reduced toxicity, and may also significantly reduce (or potentially eliminate) the onset of drug resistance, thus prolonging patient survival & improving quality of life.
Clinicians will have access to a computer model that allows them to design an optimised therapeutic regimen for patients. This model will also be of use to those designing clinical trials, particularly involving drug combinations.

NHS benefits will be economic - reduced drug bills and patient costs - and logistical - fewer, shorter hospital admissions due to less toxic drug regimens.

Our humanised mouse model will be of significant interest to the Pharmaceutical industry for use in their drug development pipeline, since most drugs fail during the early stages of human trials on safety & toxicity criteria which may be attributed to species differences with pre-clinical testing. Use of the humanised model thus has significant potential to reduce late-stage withdrawal of with associated savings in time and costs.