Natural Killer Cells Could Bypass Limitations of CAR T Cells in CLL, Expert Says

Katy Rezvani, MD

Katy Rezvani, MD

Although chimeric antigen receptor (CAR) T-cell therapy can offer profound benefits for patients with select hematologic malignancies, chronic lymphocytic leukemia (CLL) has proven to be a difficult disease to treat with this therapy. Now, investigators of an ongoing trial are hoping that CAR natural killer (NK) cells may provide a way to sidestep some of the limitations with CAR Ts.

Because CAR T-cell therapy is specific to each patient, it can be both timely and expensive. CAR T cells have also been associated with the development of graft-versus-host disease (GVHD), which can ultimately make T cells ineffective.

A new CAR therapy, CAR NK cells, is currently being investigated in a phase I/II dose escalation trial (NCT02727803), according to lead investigator Katy Rezvani, MD, PhD, who gave a presentation on CAR NK cells during the 2018 SOHO Annual Meeting.

Because CAR NK cells are not patient-specific or associated with GVHD, this treatment shows promise in overcoming the limitations with CAR T cells. In patients with CLL, this therapy can also potentially be combined with immunomodulatory agents such as lenalidomide (Revlimid) to increase antitumor activity, Rezvani said.

In an interview withTargeted Oncology,Rezvani, professor and director of translational research in the Department of Stem Cell Transplantation, Division of Cancer Medicine at The University of Texas MD Anderson Cancer Center, discussed the potential she sees with CAR NK cells. She also highlighted some of the challenges with both CAR T-cell therapy in patients with CLL, as well as challenges to overcome with CAR NK cells in this patient population.

TARGETED ONCOLOGY:At this year’s annual SOHO meeting, you discussed the application of CAR NK cells in CLL. Can you explain the rationale behind using CAR NK cells in this patient population?

Rezvani:I’m sure your readers are aware of CAR T cells and the shifting paradigm around acute lymphocytic leukemia (ALL), lymphoma, and now, for CLL. The problem we have with CAR T cells is that they are patient-specific, so you need to be able to collect lymphocytes or T cells from the patient so that you can generate CAR T cells for that particular patient. That makes the whole process logistically very hard in some ways and cumbersome because the patient will have to go on a leukapheresis machine and have a sufficient number of T cells collected, which are sent to the center’s facility or lab where the cells are expanded. When the cells are ready, they are then shipped back.

This whole process can take anywhere from 2 to 5 weeks, and sometimes the process fails. For instance, if a patient has had a lot of chemotherapy, they might not be able to collect a sufficient number of cells for the expansion transaction. Occasionally, infection [occurs] during the manufacturing process so that the product is not safe to infuse to the patient.

All of this makes the product very expensive, because of the fact that it is patient-specific. If we have a product that can be made from a source other than the patient that can be given to patients safely without causing GVHD, then that could potentially overcome the limitations that I mentioned. You can make that product, an allogeneic product, manufacture it, and basically freeze it, save it, and give it to the patient as an off-the-shelf like you would with a drug.

Allogeneic T cells have a problem in that they cause GVHD. We know that from transplantation, even if you HLA  match. The GVHD is driven by the T-cell receptors on the surface of T cells recognizing the alloantigens in the patient. They are not safe, unless you knock out and delete the T-cell receptor gene. That’s actually what this company Cellectis is doing, and other people are also exploring that strategy of knocking out the T-cell receptor. Obviously, this does add another layer of complication to the whole manufacturing process because now you have to do 2 steps. One is to put the CAR in, and the next is to knock out the T-cell receptor. If your product has some contaminating T cells that still express the T-cell receptor, then you still run the risk of those T cells growing and causing GVHD.

We chose to use NK cells because NK cells are a subset of lymphocytes. Actually, on a per-cell basis, NK cells are the most cytotoxic immune effectors that we have. They’re very interesting cells because they are poised to recognize a normal cell. On the surface, they have this variety of receptors that are either activating or inhibitory, so they have multiple mechanisms that are very inherit to them that allows them to recognize tumor cells.

The other very interesting property of NK cells is that they don’t cause GVHD. We know that because I would say hundreds of patients, if not thousands, have been treated with allogeneic NK cells or with NK cell lines that have been mismatched to the patient, and GVHD has rarely been reported. There was only 1 study that showed NK cells that were infused to the patient in the setting of matched unrelated during transplant and without immunosuppression, it was associated with GVHD. However, you might argue that patients have had a transplant anyway, so they were destined to develop GVHD. In any other situation where the NK cells were infused into patients after chemotherapy, no one has seen GVHD.

That’s why we decided to use NK cells as a source of immune effectors for CAR engineering. Because of the fact that they already have very potent activity against cancer cells through their receptors. As I mentioned, you can actually pick your donor based on KIR ligand mismatch and really harness the properties of NK cells. They don’t cause GVHD; you can have an allogeneic product. Once you put the CAR in the NK cells, you give them the properties of recognizing the tumors through the CAR on the surface as long as the tumor cells express the antigen of interest.

TARGETED ONCOLOGY:Are there any details you can share on this trial?

Rezvani:Our current trial is open, and it’s enrolling patients with CLL, non-Hodgkin lymphoma (NHL), and ALL. It’s open to all 3 groups. This is, basically, off-the-shelf cord blood derived CAR NK cells that are infused to patients with CLL, ALL, and NHL after the patient receives chemotherapy. We use a very similar regimen to the CAR T-cell therapy, so patients get fludarabine and cyclophosphamide, and then we infuse the CAR NK cells that we generate in our facility.

We are analyzing the data, but it is not yet published. We are hoping that we can submit the full analysis of the data in the next few months.

TARGETED ONCOLOGY:What sort of challenges do you foresee with CAR NK cells in CLL?

Rezvani:CLL has been a tough disease to treat with CAR T cells, in that the responses have been lower than what has been reported for CAR T with ALL and CAR T for NHL. The question is going to be for the responses we see with CAR NK cells, are there going to be long-term remissions? That’s why we need a longer follow-up on the patients we treated. Should we develop strategies of combining the targeted therapies that are currently being used for CLL with CAR NK cells? I think once we have more data on the efficacy of CAR NK-cell therapy, and also on the longevity of response, we will have a better idea of this.

There are certain drugs that are used for CLL, like lenalidomide, for instance, that is immunomodulatory. To me, it makes a lot of sense to combine CAR NK cells with lenalidomide, which will have both activity against CLL but also support the expansion of the antitumor activity of CAR NK cells. I think we will obviously need a larger number of patients that are being treated with CLL that receive the treatment.

One reason that it is thought that CAR T cells may not be as effective in CLL as in ALL and NHL is that the T cells in patients with CLL are dysfunctional. That’s being shown in different settings where investigators have shown that the T cells in CLL patients don’t respond to antigens as effectively, they don’t form immunological synapses, etc. Maybe the problem isn’t necessarily that CLL is not targetable, but in CAR therapy, the problem is that immune effectors that are being used to target CLL are not optimal, in that they are somewhat dysfunctional.