Developing a chick embryo model to aid in development of personalised therapies for malignant pleural mesothelioma
Lead Research Organisation:
University of Liverpool
Department Name: Cellular and Molecular Physiology
Abstract
Mesothelioma is an aggressive and largely untreatable cancer of the lung lining, mainly caused by environmental exposure to asbestos. New treatments, or new approaches to treatment, are urgently required. We can now read detailed information about genetic changes from a small sample of a patient's cancer, which can then be used to make decisions about the most effective anti-cancer drugs to give to an individual patient as "precision medicine". Recent studies have revealed the type and frequency of genetic changes that occur in mesothelioma, which may help in predicting new treatments.
In many cancers, genetic changes switch on "oncogenes", which accelerate the speed with which cancer cells divide into two, driving tumour growth. Many cancer treatments use drugs that directly block the activity of oncogenes to prevent this uncontrolled tumour growth. However, mesothelioma is unusual, as there are no common oncogene mutations. Instead, genetic changes mostly occur in "tumour suppressor" genes, disabling proteins that would normally apply a brake to slow down dividing cells and so prevent tumour growth. This presents a difficult challenge for finding ways to treat mesothelioma, as we need to fully understand how each specific tumour suppressor mutation alters the cancerous behaviour of mesothelioma cells, in order to find an Achilles' heel that we might be able to target with drugs. Ultimately, we also need to develop the best laboratory models in which to test the drugs, before they can be given to mesothelioma patients.
Disabling mutations of the tumour suppressor BAP1 are found in more than half of all mesotheliomas. Normally, BAP1 controls the production and destruction of other proteins within the cell. Therefore, in mesothelioma without BAP1, there are potentially changes in the amounts of many different proteins that could affect cancerous behaviour. Using cells with gene-edited mutations of BAP1, we identified many of these protein changes. We found that BAP1 mutation not only affects proteins that alter the growth of cancer cells, but also proteins that control how they move, gain access to blood vessels, and spread around the body. We are currently evaluating which of these proteins make mesothelioma cells more sensitive to specific anti-cancer drugs. However, we need to test these drugs in models that can provide a good replica of human mesothelioma growth and spread.
To do this, we will develop a chick embryo model of mesothelioma, as a replacement for currently used mouse models. The chick embryo model is classified as non-protected under the Animals Scientific Procedures Act, and so is a useful technique to replace testing in animals. It has many additional advantages over mouse models, including cost effectiveness, accessibility and speed. It is an excellent model to study the growth and spread of tumour cells, as they can be easily engrafted onto the "chorioallantoic membrane". This is an accessible surface, located outside the chick embryo directly beneath the eggshell, with a good supply of blood vessels. Within a few days, a small tumour develops, which can spread across and into the membrane, potentially accessing blood vessels to spread to specific organs. Importantly, new drug treatments can be readily tested in the chick embryo model, and the tumour cells imaged over time to assess their survival and behaviour.
We will use the chick embryo model to grow mesothelioma cells, with and without BAP1 mutation, and evaluate therapeutic responses to our candidate drugs. Successful outcomes will suggest new drugs for inclusion in precision medicine trials in mesothelioma patients. During the project, we will develop the first standard operating procedures to generate and monitor mesothelioma tumours in this model. We will make these protocols, and key reagents, available to the mesothelioma research community, encouraging widespread replacement of murine models.
In many cancers, genetic changes switch on "oncogenes", which accelerate the speed with which cancer cells divide into two, driving tumour growth. Many cancer treatments use drugs that directly block the activity of oncogenes to prevent this uncontrolled tumour growth. However, mesothelioma is unusual, as there are no common oncogene mutations. Instead, genetic changes mostly occur in "tumour suppressor" genes, disabling proteins that would normally apply a brake to slow down dividing cells and so prevent tumour growth. This presents a difficult challenge for finding ways to treat mesothelioma, as we need to fully understand how each specific tumour suppressor mutation alters the cancerous behaviour of mesothelioma cells, in order to find an Achilles' heel that we might be able to target with drugs. Ultimately, we also need to develop the best laboratory models in which to test the drugs, before they can be given to mesothelioma patients.
Disabling mutations of the tumour suppressor BAP1 are found in more than half of all mesotheliomas. Normally, BAP1 controls the production and destruction of other proteins within the cell. Therefore, in mesothelioma without BAP1, there are potentially changes in the amounts of many different proteins that could affect cancerous behaviour. Using cells with gene-edited mutations of BAP1, we identified many of these protein changes. We found that BAP1 mutation not only affects proteins that alter the growth of cancer cells, but also proteins that control how they move, gain access to blood vessels, and spread around the body. We are currently evaluating which of these proteins make mesothelioma cells more sensitive to specific anti-cancer drugs. However, we need to test these drugs in models that can provide a good replica of human mesothelioma growth and spread.
To do this, we will develop a chick embryo model of mesothelioma, as a replacement for currently used mouse models. The chick embryo model is classified as non-protected under the Animals Scientific Procedures Act, and so is a useful technique to replace testing in animals. It has many additional advantages over mouse models, including cost effectiveness, accessibility and speed. It is an excellent model to study the growth and spread of tumour cells, as they can be easily engrafted onto the "chorioallantoic membrane". This is an accessible surface, located outside the chick embryo directly beneath the eggshell, with a good supply of blood vessels. Within a few days, a small tumour develops, which can spread across and into the membrane, potentially accessing blood vessels to spread to specific organs. Importantly, new drug treatments can be readily tested in the chick embryo model, and the tumour cells imaged over time to assess their survival and behaviour.
We will use the chick embryo model to grow mesothelioma cells, with and without BAP1 mutation, and evaluate therapeutic responses to our candidate drugs. Successful outcomes will suggest new drugs for inclusion in precision medicine trials in mesothelioma patients. During the project, we will develop the first standard operating procedures to generate and monitor mesothelioma tumours in this model. We will make these protocols, and key reagents, available to the mesothelioma research community, encouraging widespread replacement of murine models.
Technical Summary
Our objective is to develop a 3Rs-compliant in vivo model of malignant pleural mesothelioma (MPM) to replace/reduce use of murine models in testing new therapies. Murine MPM models include xenografts (moderate severity) and asbestos inhalation (severe). MPM is a cancer of the lung lining arising with long latency following asbestos exposure. It is aggressive, locally invasive, with common extra-thoracic dissemination. MPM patients have poor prognosis, with median survival of ~1-year and limited treatment options available. New targeted therapies are urgently required, and recent investment in MPM research is driving development of novel strategies. BAP1 is a tumour suppressor with frequent loss-of-function mutation in MPM. However, BAP1 is an essential/fitness gene, infrequently mutated in common epithelial cancers, suggesting that MPM must reprogram essential pathways to adapt. We hypothesised such reprogramming generates vulnerabilities that may be therapeutically targeted, and developed an isogenic mesothelial cell model, in which we have identified potentially drugable BAP1-dependencies affecting the survival and invasive behaviour of MPM. Pioneering work in Liverpool is establishing chick embryo models for a variety of cancers, which we plan to extend to MPM. Dual-labelled (bioluminescence/fluorescent) MPM cells will be injected intravenously, or implanted onto the chorioallantoic membrane (CAM). Intravital imaging and histology will be used to assess proliferation, angiogenesis, local invasion and metastatic dissemination of MPM. Importantly, the model will allow cost-effective in vivo testing of candidate BAP1-directed therapeutic interventions on MPM cell lines stratified by BAP1-status. Major outcomes will be (i) phenotyping MPM cell lines, (ii) in vivo screening of novel therapeutics for BAP1-mutated MPM, and (ii) proof-of-principle for a model that can be adopted to replace/reduce murine xenograft testing of other emerging therapeutic strategies for MPM.
Planned Impact
Impact will be through reduction, and in some cases replacement, of murine xenograft models of MPM, in particular for the study of new targeted/personalised therapies.
Our ongoing work is identifying candidate drugable targets in BAP1-mutated MPM. We planned to refine candidate drugs through in vitro testing, before testing differential sensitivity of MPM, stratified by BAP1-status, in murine xenografts. However, in vitro and in vivo data are not always comparable, and in vitro testing is not feasible for our top candidates, where the readouts for drug efficacy are angiogenesis or metastasis. Power calculations, using a balanced block approach to minimise numbers, suggest at least 4 mice required per group for drug testing, under procedures with moderate severity. Testing 2 experimental drugs (2 concs, plus vehicle) in xenografts of 2 BAP1+ve and 2 BAP1-ve MPM cell lines, would require 80 mice per experiment to provide statistical power in excess of 90%, with an experimentally important difference of 3-fold (s.d. 0.6). We would anticipate ultimately testing 8 drugs, requiring 360 mice.
The MPM-CAM model is classed as a partial replacement by NC3R, as the embryos are terminated at E14. This model will allow us to test e.g. 8 drugs, leading to the direct replacement of 360 mice (Aim-3). Our primary goal is to repurpose drugs, approved by FDA/EMA for use in other human diseases, to treat MPM. These do not require safety testing in mammals prior to use in man, and so the MPM-CAM model can completely replace murine xenografts. Where we identify BAP1-dependencies for which approved drugs are not available, we will initially work with tool compounds or shRNA for proof-of-concept, and establish collaborations to develop new drug molecules. In this case, the MPM-CAM will be used to pre-screen for the most effective/least toxic compounds. Although any candidate drug would ultimately need to be tested in a mammal, if e.g. only 1 of 5 test compounds were taken forward, an 80% reduction in mice would be achieved in early preclinical work.
Other research groups are investigating new drug candidates for MPM, and ultimately need to verify findings in mouse xenografts (e.g. refs 7&8). MPM is a relatively rare cancer, with around 2000 new cases per year in the UK. Despite this, within the wider scientific community over the last 5 years, around 25 studies/year report use of mouse xenografts to study MPM biology or to test new therapies. Studies typically use 5 to 10 mice per group and up to 6 treatment arms, so conservatively assuming 30 to 60 animals per published study, a minimum of 750 to 1500 mice are used annually worldwide.
Using the MPM-CAM to in vivo phenotype the MesobanK cell panel (Aims-1&2), could directly replace at least 48 mice (4x 12 lines), and enable other investigators to reduce future usage.
Murine MPM PDX models are reported with only 40% engraftment rate (ref 20). Each is typically passaged 5-times, using 5 mice per passage, totalling 26 mice per successful PDX. Given the advantages of PDXs, their use for MPM will likely increase, escalating mice numbers through establishing the PDX and allowing for engraftment failure in experiments (estimate ~3x mice compared to conventional xenografts). The MPM CAM-PDX model (Aim-4) can potentially substitute murine PDXs: replacing 50 mice to establish 10 MPM biopsies, plus ~40 mice if 2 PDXs are passaged x5.
Whilst we primarily consider the MPM-CAM as replacement for murine flank xenografts, there is some potential to replace/reduce more complex, higher severity, mouse MPM models in future (orthotopic xenografts, asbestos inhalation, or angiogenesis chamber models). Replacement estimates are based on historic data, but we anticipate recent multi-million-pound investment into MPM research (estimated >3x increase in UK), will substantially increase mouse xenografts use over the next 5 years, unless viable alternative models are developed for the MPM research community.
Our ongoing work is identifying candidate drugable targets in BAP1-mutated MPM. We planned to refine candidate drugs through in vitro testing, before testing differential sensitivity of MPM, stratified by BAP1-status, in murine xenografts. However, in vitro and in vivo data are not always comparable, and in vitro testing is not feasible for our top candidates, where the readouts for drug efficacy are angiogenesis or metastasis. Power calculations, using a balanced block approach to minimise numbers, suggest at least 4 mice required per group for drug testing, under procedures with moderate severity. Testing 2 experimental drugs (2 concs, plus vehicle) in xenografts of 2 BAP1+ve and 2 BAP1-ve MPM cell lines, would require 80 mice per experiment to provide statistical power in excess of 90%, with an experimentally important difference of 3-fold (s.d. 0.6). We would anticipate ultimately testing 8 drugs, requiring 360 mice.
The MPM-CAM model is classed as a partial replacement by NC3R, as the embryos are terminated at E14. This model will allow us to test e.g. 8 drugs, leading to the direct replacement of 360 mice (Aim-3). Our primary goal is to repurpose drugs, approved by FDA/EMA for use in other human diseases, to treat MPM. These do not require safety testing in mammals prior to use in man, and so the MPM-CAM model can completely replace murine xenografts. Where we identify BAP1-dependencies for which approved drugs are not available, we will initially work with tool compounds or shRNA for proof-of-concept, and establish collaborations to develop new drug molecules. In this case, the MPM-CAM will be used to pre-screen for the most effective/least toxic compounds. Although any candidate drug would ultimately need to be tested in a mammal, if e.g. only 1 of 5 test compounds were taken forward, an 80% reduction in mice would be achieved in early preclinical work.
Other research groups are investigating new drug candidates for MPM, and ultimately need to verify findings in mouse xenografts (e.g. refs 7&8). MPM is a relatively rare cancer, with around 2000 new cases per year in the UK. Despite this, within the wider scientific community over the last 5 years, around 25 studies/year report use of mouse xenografts to study MPM biology or to test new therapies. Studies typically use 5 to 10 mice per group and up to 6 treatment arms, so conservatively assuming 30 to 60 animals per published study, a minimum of 750 to 1500 mice are used annually worldwide.
Using the MPM-CAM to in vivo phenotype the MesobanK cell panel (Aims-1&2), could directly replace at least 48 mice (4x 12 lines), and enable other investigators to reduce future usage.
Murine MPM PDX models are reported with only 40% engraftment rate (ref 20). Each is typically passaged 5-times, using 5 mice per passage, totalling 26 mice per successful PDX. Given the advantages of PDXs, their use for MPM will likely increase, escalating mice numbers through establishing the PDX and allowing for engraftment failure in experiments (estimate ~3x mice compared to conventional xenografts). The MPM CAM-PDX model (Aim-4) can potentially substitute murine PDXs: replacing 50 mice to establish 10 MPM biopsies, plus ~40 mice if 2 PDXs are passaged x5.
Whilst we primarily consider the MPM-CAM as replacement for murine flank xenografts, there is some potential to replace/reduce more complex, higher severity, mouse MPM models in future (orthotopic xenografts, asbestos inhalation, or angiogenesis chamber models). Replacement estimates are based on historic data, but we anticipate recent multi-million-pound investment into MPM research (estimated >3x increase in UK), will substantially increase mouse xenografts use over the next 5 years, unless viable alternative models are developed for the MPM research community.