Immunocompetent in vivo CRISPR screening to identify key transcription factors which enhance persistence and efficacy of CAR-T cells in cancer
Lead Research Organisation:
University College London
Department Name: Haematology
Abstract
Chimeric antigen receptor T cells (CAR-Ts) is an exciting treatment where a patient's own immune cells are engineered in the lab to recognise tumour. These cells are then given back to patients by drip where they live and grow in the body, and can find and kill cancer cells. CAR-Ts have been particularly effective in children with acute lymphoblastic leukaemia (ALL). Most ALL patients respond to CAR-Ts but less than half remain cancer free at 2 years. Return of cancer cells (relapse) is often seen when the CAR-Ts are no longer detectable (ie they have died out) so are not able to control tumour. If we could make CAR-Ts last longer ('persist'), we could improve the number of patients who do not relapse and remain cancer-free.
There isn't one single factor which makes CAR-Ts persist in a treated patient. A CAR-T cell will be more likely to persist if it is 'fit' itself: for example, it is able to kill tumour efficiently and to replicate itself, forming an 'army' to attack the cancer. However, the tumour cells can fight back, and try to survive by releasing chemicals which incapacitate CAR-T or other immune cells that might attack. Thus CAR-Ts also need to be able to survive near tumour cells.
Small proteins called transcription factors (TFs) exist naturally in cells and are used as 'control panels' to rapidly turn on lots of switches at the same time. They are used by the cell to rapidly change cell function or behaviour for a variety of reasons. TFs are critically important in deciding how immune cells, such as CAR-T cells, behave. Other scientists have described a small number of TFs that make CAR-Ts kill cancer more effectively or survive longer in patients, but these are unlikely to be the best options, as they were discovered in highly artificial experiments which don't replicate human cancer well.
We have designed a series of experiments to uncover which TFs can be used to make CAR-T cells that persist longer. 1) We will assess many TFs at the same time 2) TFs can be increased or decreased and scientists have mostly described TFs that improve CAR-T function if they are reduced or deleted. This isn't useful for CAR-T cell therapy as it is difficult and expensive to reduce levels of proteins in cells used to treat patients. With the technology we have now, it is easier to increase a protein in a CAR-T cell. So we will look for TFs that make a CAR-T work better/ last longer if the TF is increased 3) We will test the CAR-Ts and TFs in an animal model that replicates the 'inhospitable' environment of tumour, so we can make sure the CAR-Ts work better in these conditions and therefore probably work better in patients too.
In this project, we will for the first time, systematically increase the expression of one of >800 possible TFs in CAR-T cells. A mix of cells (each with increased expression of only 1 TF) will be injected into a mouse with leukaemia. The particular cells with TFs that are able to make CAR-Ts persist will last longer than usual - and these can then be analysed and identified in the laboratory. To check our results, we can then introduce those identified TFs into a new batch of CAR-Ts and then see if they make longer lasting CAR-Ts in other mice, and then in human cells.
The expectation is that this research can create better and longer lasting CAR-T cells for patients with ALL in the first instance, and we aim that this work leads directly to a clinical trial in ALL. However, tumours are different so we expect that the TFs that will make a CAR-T cell work better in ALL may not be the same as those which work well in other types of cancer. We are developing a number of animal models in the lab to assess CAR-T cells against different cancer types. Once we have shown that these series of experiments are useful for creating longer lasting CAR-T cells in ALL, we will replicate these experiments for different cancers, with the aim of developing the next generation of more effective CAR-T cells.
There isn't one single factor which makes CAR-Ts persist in a treated patient. A CAR-T cell will be more likely to persist if it is 'fit' itself: for example, it is able to kill tumour efficiently and to replicate itself, forming an 'army' to attack the cancer. However, the tumour cells can fight back, and try to survive by releasing chemicals which incapacitate CAR-T or other immune cells that might attack. Thus CAR-Ts also need to be able to survive near tumour cells.
Small proteins called transcription factors (TFs) exist naturally in cells and are used as 'control panels' to rapidly turn on lots of switches at the same time. They are used by the cell to rapidly change cell function or behaviour for a variety of reasons. TFs are critically important in deciding how immune cells, such as CAR-T cells, behave. Other scientists have described a small number of TFs that make CAR-Ts kill cancer more effectively or survive longer in patients, but these are unlikely to be the best options, as they were discovered in highly artificial experiments which don't replicate human cancer well.
We have designed a series of experiments to uncover which TFs can be used to make CAR-T cells that persist longer. 1) We will assess many TFs at the same time 2) TFs can be increased or decreased and scientists have mostly described TFs that improve CAR-T function if they are reduced or deleted. This isn't useful for CAR-T cell therapy as it is difficult and expensive to reduce levels of proteins in cells used to treat patients. With the technology we have now, it is easier to increase a protein in a CAR-T cell. So we will look for TFs that make a CAR-T work better/ last longer if the TF is increased 3) We will test the CAR-Ts and TFs in an animal model that replicates the 'inhospitable' environment of tumour, so we can make sure the CAR-Ts work better in these conditions and therefore probably work better in patients too.
In this project, we will for the first time, systematically increase the expression of one of >800 possible TFs in CAR-T cells. A mix of cells (each with increased expression of only 1 TF) will be injected into a mouse with leukaemia. The particular cells with TFs that are able to make CAR-Ts persist will last longer than usual - and these can then be analysed and identified in the laboratory. To check our results, we can then introduce those identified TFs into a new batch of CAR-Ts and then see if they make longer lasting CAR-Ts in other mice, and then in human cells.
The expectation is that this research can create better and longer lasting CAR-T cells for patients with ALL in the first instance, and we aim that this work leads directly to a clinical trial in ALL. However, tumours are different so we expect that the TFs that will make a CAR-T cell work better in ALL may not be the same as those which work well in other types of cancer. We are developing a number of animal models in the lab to assess CAR-T cells against different cancer types. Once we have shown that these series of experiments are useful for creating longer lasting CAR-T cells in ALL, we will replicate these experiments for different cancers, with the aim of developing the next generation of more effective CAR-T cells.
Technical Summary
A lack of persistence of chimeric antigen receptors T cells (CARs) in patients with acute lymphoblastic leukemia (ALL) remains a barrier to durable remissions. A few transcription factors (TFs) have been described to influence CAR persistence, identified by targeting known immune modulating TFs, or in patients with persisting TF-disrupted CAR clones.
CRISPR screens can be used to search for TFs that improve CAR persistence, but screens to date had limited relevance as: 1)Genome-wide screens have limited power and typically implicate negative only regulatory factors. Gene editing to reduce (vs engineering to increase) expression is challenging for CAR manufacture. 2) Assessment in vitro or in immunodeficient animal models poorly replicate the suppressive factors which impact CAR function in patients.
We intend to use a CRISPR activation screen of a library of TFs in murine (mu-) CARs in syngeneic murine tumour models, with the aim of uncovering TFs that will augment CAR persistence in patients.
AIM1 Optimise an activating CRISPR screen targeting transcription factors in CARs. We will use a CRISPR/dCas9-SPH activation system with a plasmid library targeting >800 muTFs and co-expressing anti-CD19 CAR. TF-activated muCARs will be injected into a syngeneic ALL model (E2a-PBX).
AIM2 Identify key TFs associated with enhanced CAR persistence in vivo. MuCAR treated mice will be bled weekly and persistent TF-activated CARs FACS sorted and sequenced.
AIM3 TFs will be modulated individually in muCARs to validate their functional role in persistence in vivo.
AIM4 Assess effect of TF modulation in human (hu-)CARs. Hu sequences of promising TFs will be co-transduced with hu anti-CD19 CARs into huT cells prior to functional validation in vitro and in vivo(cell lines, patient derived xenografts).
We next aim to apply this targeted, activating screen in other relevant tumour models to reveal key TFs able to improve clinical responses following CAR-T in cancer.
CRISPR screens can be used to search for TFs that improve CAR persistence, but screens to date had limited relevance as: 1)Genome-wide screens have limited power and typically implicate negative only regulatory factors. Gene editing to reduce (vs engineering to increase) expression is challenging for CAR manufacture. 2) Assessment in vitro or in immunodeficient animal models poorly replicate the suppressive factors which impact CAR function in patients.
We intend to use a CRISPR activation screen of a library of TFs in murine (mu-) CARs in syngeneic murine tumour models, with the aim of uncovering TFs that will augment CAR persistence in patients.
AIM1 Optimise an activating CRISPR screen targeting transcription factors in CARs. We will use a CRISPR/dCas9-SPH activation system with a plasmid library targeting >800 muTFs and co-expressing anti-CD19 CAR. TF-activated muCARs will be injected into a syngeneic ALL model (E2a-PBX).
AIM2 Identify key TFs associated with enhanced CAR persistence in vivo. MuCAR treated mice will be bled weekly and persistent TF-activated CARs FACS sorted and sequenced.
AIM3 TFs will be modulated individually in muCARs to validate their functional role in persistence in vivo.
AIM4 Assess effect of TF modulation in human (hu-)CARs. Hu sequences of promising TFs will be co-transduced with hu anti-CD19 CARs into huT cells prior to functional validation in vitro and in vivo(cell lines, patient derived xenografts).
We next aim to apply this targeted, activating screen in other relevant tumour models to reveal key TFs able to improve clinical responses following CAR-T in cancer.
