Next-generation CAR T cells
Here, Dr John Maher responds to questions about how to overcome the challenges associated with developing chimeric antigen receptor (CAR) T-cell therapy for solid tumours and how the work as Leucid Bio is making strides towards this goal.
Could you tell me about yourself and what you do?
I am a practising clinical immunologist but spend most of my time divided between being an academic at King’s College London and working as the Chief Scientific Officer of Leucid Bio. My primary focus from the academic and commercial perspective is the development of CAR T-cell immunotherapies mainly, but not exclusively, for solid tumours. I have been working on CAR T cells since 1999 and it has been very interesting because I was there in the early days when CAR T-cell therapy was not considered to be anything other than a niche activity.
I started out in CAR T-cell research at Memorial Sloan Kettering Cancer Center, working in the laboratory of Michel Sadelain, who is a pioneer in the field. When I was there, I introduced second-generation CAR T-cell technology into the lab, benefitting from Michel’s ability to genetically engineer human T cells rather than immortalised cell lines. Second-generation CAR technology was invented here in the UK by Helene Finney and consists of a chimeric receptor in which you have both an activating unit and a co-stimulatory domain. The big difference this leads to is the fact that T cells not only kill tumour cells but can also proliferate following that encounter. I was very much taken with the potency of this approach, converting me to become a CAR T-cell scientist. Nonetheless, I was advised by a number of senior colleagues that I was wasting my time working in this particular arena. Hearteningly however, when we fast forward about 10 years later, we witnessed the tremendous impact that this technology has made in the context of treating patients with refractory blood cancers.
We spun Leucid Bio out at the end of 2014 and secured funding in 2017. That second-generation CAR T-cell technology is what we are using currently to treat head and neck cancer patients in an ongoing clinical trial.
Could you briefly outline how CAR T cells work and why they are effective?
T cells are key white blood cells of the immune system, which recognise foreign species called antigens. They are adapted to the individual in whom the T cells circulate. This is one of the reasons that organ transplantation is a problem, because if you take an organ from a donor and give it to a genetically unrelated recipient, their T cells will recognise it as a foreign invader, leading the T cells to attack it. This creates problems for new cancer immunotherapies because they must be designed with the patient in mind. A CAR is a synthetic solution to address this issue. The CAR allows the T cells to recognise tumours by virtue of a target molecule expressed on the surface of the tumour itself.
As long as the target is there, the CAR will enable the T cell to recognise that particular tumour cell. This means that the CAR does not need to be tailored to the patient, provided that their cancer expresses the target that it binds to.
The manner in which the CAR is introduced into the patient’s T cells most commonly involves the use of a disabled viral vector, which cannot replicate. It can deliver the gene encoding for the CARs into the T cells and thereafter the virus no longer propagates itself. By contrast, the T cells can replicate and all the daughters of that population of T cells will continue to express the CAR. That gives you a large pool of T cells with the potential to target tumour cells.
Why have CAR T cells seen more success in haematological cancers and is it possible to target them towards solid tumours?
In the history of cancer medicine, it is traditionally easier to treat blood cancers than solid tumours. The reason that haematologists came out of the lab and went into the clinic was because in the 1960s, they learned how to treat childhood leukaemia with chemotherapy drugs. Since then, we have seen the development of bone marrow transplantation as a very powerful, albeit intensive, treatment for patients with blood cancers who can achieve remission with traditional therapies. We have lagged behind in developing curative therapies for solid tumours but have developed therapies which can control these tumours for a period of time.
In general however, common solid tumours are very difficult to cure once they have undergone metastasis.
There are several hurdles to the development of effective CAR T-cell therapies for solid tumours. The first issue you face is the selection of a suitable target antigen that is expressed on the malignant cells. When it comes to blood cancers, the top two targets are molecules known as CD19 and BCMA. These two molecules are expressed on the malignant cell population and also on some healthy cells. You can get away with depleting those healthy cells, which are generally B cells that make antibodies, since you can treat patients with immunoglobulin replacement. This effectively circumvents the issue of having no B cells present following CAR T-cell therapy. However, this kind of trick cannot be played so easily in patients with solid tumours.
There are virtually no targets found on common solid tumours that are absolutely tumour specific or which are found on healthy, but dispensable, normal cells. Almost all the targets we look at in the context of solid tumours are found at lower levels on vital organs in the body, presenting significant risk of so-called “on-target off-tumour” toxicity. The second issue is that CAR T cells are generally administered intravenously. In the context of someone with a blood cancer, this provides direct access to the malignant cells. By contrast, solid tumours tend to originate and spread to various vital organs of the body. This presents a problem of how to get the CAR T cells to exit the bloodstream and access the sites where tumour deposits are located. This is something people are beginning to address by introducing homing receptors alongside the CAR, but it is not yet a perfect solution.
If you can manage to overcome these issues – you have a safe target and you get your T cells into the tumour – you then have a third problem: the tumour microenvironment. There are a multitude of physical, chemical and biological obstacles which tumours establish within their local microenvironment, which turn off immune cell reactions. This includes acidosis, hypoxia and nutrient deprivation. Tumour cells produce a variety of immunosuppressive small molecules, like transforming growth factor (TGF) beta, IL10, prostaglandin E2, indoleamine dioxygenase – the list goes on.
Tumours also provide a physical barrier to the delivery of any kind of drug into the tumour, including CAR T cells. You also have a host of other cells in the tumour microenvironment that are immunosuppressive, such as regulatory T-cell macrophages, myeloid-derived suppressor cells, cancer-associated fibroblasts and mesenchymal stromal cells. All of these factors operate in the tumour microenvironment, which do not let T cells activate and present challenges for T-cell orchestrated immune responses.
How can we overcome some of those challenges?
If I had an answer to all of those problems, I would be a famous and wealthy man! However, if you break it down, starting with target selection, we now have very high‑end technologies available to us built around OMICs, which can help us to investigate new targets (eg, endogenous retroviral antigens) and target combinations in solid tumours.
Another tumour-associated target that we are particularly interested in are the NKG2D ligands. This is a family of eight ligands, upregulated by stress and recognised by cells such as natural killer (NK) cells which express the NKG2D receptor. A cell which expresses an NKG2D ligand is, by definition, a stressed cell, so the immune system sees it as a threat. Owing to stress arising from DNA damage and the abnormal nature of the tumour microenvironment, over 80 percent of human cancers express NKG2D ligands. These ligands constitute the target of Leucid Bio’s lead clinical CAR T-cell asset (called LEU-011) which we hope to evaluate in a clinical trial that Leucid Bio is looking to initiate next year. We have great hopes for this approach based on the pre-clinical data that we have generated.
Furthermore, the co-expression of a homing receptor can enable some degree of semi-selective trafficking of T cells into tumours. In the past, we have used CXCR2, a chemokine receptor, for this purpose. It has seven different ligands – CXCL1, 2, 3, 5, 6, 7 and 8, although CXCL8 is better known as IL8. Those ligands in particular tend to be produced by stress cells to flag the immune system, to draw in cells like neutrophils. Many solid tumours overproduce CXCR2 ligands, particularly IL8. In pre-clinical models, we have found that you can improve the trafficking of CAR T cells into human tumours if you co-express CXCR2. Other groups have shown similar findings with different homing receptors so there is great scope for tailoring the approach here.
There are also immune checkpoints which are potentially important in the tumour microenvironment, for example, PD-L1, LAG-3, TIM-3 and CTLA‑4. Consequently, we can envision the treatment of patients with a CAR T-cell product in combination with an immune checkpoint inhibitor which is tailored for that particular tumour type. TGF beta is a well-recognised immunosuppressive factor within the tumour microenvironment and there are traps which can be used to deplete TGF beta to counteract this. Alternatively, a switch receptor may be expressed in the CAR T cell which binds TGF beta and delivers a positive signal to the T cells as a result.
There are also inhibitors of other potential checkpoints such as prostaglandin E2, IDO inhibitors and arginase which offer further opportunities to develop alternative hybrid approaches.
Another solid-tumour focused approach we were involved in recently was a collaboration with James Arnold at King’s College London in which we designed hypoxia‑sensing CARs. The CAR recognises a potential dangerous target such as ErbB dimers, but since it only moves to the cell surface under conditions of hypoxia, it operates preferentially within the tumour to enhance safety.
Can you tell me about some of the work at Leucid Bio?
We are very interested in NKG2D ligands as a target family for CAR T-cell immunotherapy, building on the pioneering work of Charles Sentman. The rationale here is that these are natural targets for immune surveillance, so our hope is that by potentitating this system, we may see impact against solid tumours.
The other major thing we are doing at Leucid is redesigning CARs. Traditional cars are linear fusion receptors, where you have a targeting moiety, a spacer domain, a transmembrane domain, a signalling domain – which often has two or three components to deliver activating and co-stimulatory signals to the T cells. However, we have found that there is only so far you can push that prototype or model before you start to lose functionality. This is a particular problem with the signalling domain, where if you have three units in the endo domain, by definition, two of them are away from their natural location in proximity to the plasma membrane. To address this, we are building out rather than building up – we call this a laterally configured CAR.
An example of this principle we published last year is called a parallel CAR. This consists of a traditional second‑generation CAR (eg, containing CD28) paired with a chimeric co-stimulatory receptor in which you have a different co‑stimulatory domain, such as 4-1BB. We found that this platform consistently outperforms traditional linear CARs. We are also working on an adapter CAR system, whereby the targeting moiety is separated from the signalling machinery. In nature, many receptors that operate in the immune system have this adapter-type structure, prompting us to take a leaf out of nature’s page by building CARs which exploit this configuration. LEU-011, our lead asset, is an adapter CAR targeting NKG2D ligands.
The final point is that we are interested in is off-the-shelf allogeneic CAR T-cell solutions. Last year we published a paper describing a method to expand gamma-delta T cells to enhance their intrinsic antitumour activity. What we found with these T cells was that if they are expanded in the presence of TGF beta – a prototypic immunosuppressive cytokine – you paradoxically achieve a better yield of gamma-delta T cells with enhanced intrinsic antitumour activity. Gamma-delta T cells are interesting in this regard because, in contrast to traditional alpha beta T cells, they do not recognise antigen in a human leukocyte antigen (HLA)-dependent manner. As a result, they do not cause graft‑versus-host disease, unlike allogeneic alpha-beta T cells. This makes it difficult to use alpha beta T cells as an off-the-shelf therapy, unless you can neutralise the expression of the T-cell receptor (TCR) using genome editing. Gamma‑delta T cells are not hindered in this manner and therefore provide a ready to use off-the-shelf cellular chassis for CAR T-cell immunotherapy. For an off-the-shelf approach, you could therefore put a CAR into a gamma‑delta and if you expand the cells in TGF beta, you will also boost their intrinsic antitumour activity. That is something we are quite interested in developing at Leucid as well.
Dr John Maher is the scientific founder and Chief Scientific Officer of Leucid Bio. He is also a clinical immunologist who leads the “CAR Mechanics” research group within King’s College London. He played a key role in the early development of second-generation (CD28) CAR technology while a visiting fellow at Memorial Sloan Kettering Cancer Center, an approach that has achieved clinical impact in haematological malignancies. His research group is focused on the development of adoptive immunotherapy using CAR engineered and gamma-delta T cells, with a primary emphasis on solid tumour types. In addition, he is a consultant immunologist at Eastbourne Hospital.