Cell Type-Specific Analysis of Immune Checkpoint Signalling Networks Underpinning Cancer Immunotherapy
Lead Research Organisation:
European Bioinformatics Institute
Department Name: Petsalaki Group
Abstract
Immune checkpoints are proteins expressed on the surfaces of immune cells that suppress their activity (normally so that they don't attack us). Work on just two immune checkpoints has revolutionised cancer therapy by producing durable responses in previously untreatable diseases such as melanoma, by blocking the suppressive effects of the receptors so that the leukocytes are free to attack tumours. But there are as many as 60 immune checkpoints regulating the immune system, underscoring the extraordinary scope for medical intervention via the checkpoints, and emphasising how much work is still to be done. Remarkably, despite their enormous significance, very little is known about how the immune checkpoints work, i.e. the molecular pathways they use to switch off immune responses to tumours.
In this proposal we aim to explore how immune checkpoints differ with regard to the molecular mechanisms of their activity in three major types of immune cells (i.e. B cells and T cells, and myeloid cells), and to learn whether it will be possible to exploit these differences therapeutically. If it turns out that the immune checkpoints invoke the same pathways, it is unlikely that we will be able to make them work better collectively. However, based on what is already known about these pathways, this seems very unlikely to be the case. We will start our study at the level of a model two-cell co-culture system in vitro (i.e. in "test tubes"), where we will be able to test multiple approaches. To study what happens in the setting of authentic tumours, we will create three-dimensional (3D) cultures of actual tumours, which we can study in the course of their responses to immunotherapy.
But the main problem with understanding how the immune checkpoints work is that our current knowledge of human cellular signalling pathways is very incomplete and highly biased to well-studied ones. For example, 30-50% of the targets for the most important groups of enzymes driving signalling, called kinases and phosphatases, are completely unknown. This suggests that important pathways and processes may currently be undiscovered. Limiting our studies of immune checkpoint signalling to the known pathways would reveal only part of the jigsaw and would mean that effective new ways to treat cancer might be wholly overlooked.
To circumvent this issue, we are proposing to use a strategy that combines genetic perturbations, i.e. "gene knockouts" of all the possible kinases and phosphatases that could be involved in the signaling pathways in the immune cells under study, with measurements of signalling outcomes based on a convenient, manageable set of signaling pathway elements we can easily and accurately measure in single cells (a great leap forward). Our goal is to be able to use this small set of pathway elements to build out to the complete network. To do this we will be developing new computational pipelines in order to obtain comprehensive and accurate pictures of the whole signalling network, for each of the main sets of leykocytes involved in anti-tumour responses. Once we show that the new pipeline works, we will be able to compare and contrast how immune checkpoints vary and how different types of blockade of these proteins alters the activities of the immune cells attacking cancers.
We're very confident that our work will plug major gaps in our basic understanding of immune checkpoints which will be of considerable interest to all immunologists. But more importantly, our work could suggest important new ways to improve immune checkpoint blockade cancer immunotherapy.
In this proposal we aim to explore how immune checkpoints differ with regard to the molecular mechanisms of their activity in three major types of immune cells (i.e. B cells and T cells, and myeloid cells), and to learn whether it will be possible to exploit these differences therapeutically. If it turns out that the immune checkpoints invoke the same pathways, it is unlikely that we will be able to make them work better collectively. However, based on what is already known about these pathways, this seems very unlikely to be the case. We will start our study at the level of a model two-cell co-culture system in vitro (i.e. in "test tubes"), where we will be able to test multiple approaches. To study what happens in the setting of authentic tumours, we will create three-dimensional (3D) cultures of actual tumours, which we can study in the course of their responses to immunotherapy.
But the main problem with understanding how the immune checkpoints work is that our current knowledge of human cellular signalling pathways is very incomplete and highly biased to well-studied ones. For example, 30-50% of the targets for the most important groups of enzymes driving signalling, called kinases and phosphatases, are completely unknown. This suggests that important pathways and processes may currently be undiscovered. Limiting our studies of immune checkpoint signalling to the known pathways would reveal only part of the jigsaw and would mean that effective new ways to treat cancer might be wholly overlooked.
To circumvent this issue, we are proposing to use a strategy that combines genetic perturbations, i.e. "gene knockouts" of all the possible kinases and phosphatases that could be involved in the signaling pathways in the immune cells under study, with measurements of signalling outcomes based on a convenient, manageable set of signaling pathway elements we can easily and accurately measure in single cells (a great leap forward). Our goal is to be able to use this small set of pathway elements to build out to the complete network. To do this we will be developing new computational pipelines in order to obtain comprehensive and accurate pictures of the whole signalling network, for each of the main sets of leykocytes involved in anti-tumour responses. Once we show that the new pipeline works, we will be able to compare and contrast how immune checkpoints vary and how different types of blockade of these proteins alters the activities of the immune cells attacking cancers.
We're very confident that our work will plug major gaps in our basic understanding of immune checkpoints which will be of considerable interest to all immunologists. But more importantly, our work could suggest important new ways to improve immune checkpoint blockade cancer immunotherapy.
Technical Summary
Despite the remarkable impact of the immune checkpoints on cancer immunotherapy, very little is known about their mechanisms of inhibitory signalling in leukocytes. There are 50-60 immune checkpoints, underscoring the extraordinary scope for clinical intervention via these receptors, once they are understood. Annotated signalling pathways are unsuitable as the basis of such studies as they are static, context-agnostic, and highly biased towards well-studied processes. They are also incomplete with up to 50% of kinases/phosphatases having no known substrates. In this proposal, to guide the design of improved immunotherapy, we aim to perform a systematic and unbiased characterisation of the immune cell-type specific activating and inhibitory signalling pathways.
We will use CRISPR to knock-out all kinases and phosphatases, combined with measurements of critical phosphosites using mass cytometry (CyTOF) and computational network analysis to generate unbiased, context-specific signalling networks at the single-cell level. Based on these networks, we will characterise the crosstalk between activating signalling and four immune checkpoints, first in model cell systems and then in an autologous ex situ organoid model of the tumour microenvironment. We will also integrate CyTOF data with scRNAseq for the organoid model, to generate a model of immunotherapeutic responses in the tumour microenvironment ex situ. Starting from network signatures generated from 38 CyTOF phosphosite measurements and >17 protein ones, our computational strategy will identify extended signalling networks a) using network diffusion to expand these signatures and gain a better overview of the processes involved, and b) using our CRISPR KO/CyTOF data to include all relevant kinases/phosphatases, including the understudied ones. The results of our study will be of very considerable basic scientific interest but should also have a significant impact on the development of more effective immunotherapies.
We will use CRISPR to knock-out all kinases and phosphatases, combined with measurements of critical phosphosites using mass cytometry (CyTOF) and computational network analysis to generate unbiased, context-specific signalling networks at the single-cell level. Based on these networks, we will characterise the crosstalk between activating signalling and four immune checkpoints, first in model cell systems and then in an autologous ex situ organoid model of the tumour microenvironment. We will also integrate CyTOF data with scRNAseq for the organoid model, to generate a model of immunotherapeutic responses in the tumour microenvironment ex situ. Starting from network signatures generated from 38 CyTOF phosphosite measurements and >17 protein ones, our computational strategy will identify extended signalling networks a) using network diffusion to expand these signatures and gain a better overview of the processes involved, and b) using our CRISPR KO/CyTOF data to include all relevant kinases/phosphatases, including the understudied ones. The results of our study will be of very considerable basic scientific interest but should also have a significant impact on the development of more effective immunotherapies.
