Next Steps for Immunotherapies, Cell Therapies in Oncology: Combinations, Duration of Therapy, Overcoming Resistance

Stefan Zimmermann, MD; Inge Marie Svane, MD, PhD; George Coukos, MD, PhD


February 09, 2018

Stefan Zimmermann, MD: Hello. I am Stefan Zimmermann from the immuno-oncology (IO) service at Lausanne University Hospital in Switzerland. Welcome to Medscape Oncology Insights, coming to you from the Congress of the European Society for Medical Oncology (ESMO) 2017. Today we will be discussing the rise of immunotherapies and some of the issues that are facing the field.

I'm pleased to have George Coukos, from the Ludwig Institute for Cancer Research and the department of oncology at the University of Lausanne in Switzerland, joining me in this discussion. He is also the ESMO Immuno-Oncology Congress Scientific Co-chair. Also joining me is Dr Inge Maria Svane, who directs the Center for Cancer Immune Therapy and the department of oncology at Herlev University Hospital in Denmark. She is the chair of the Scientific Committee on Immunotherapy.

This Congress has been all about immunotherapies. We have seen and will be seeing lots of great news, most of it on immune checkpoint blockade but also on novel combinations. There will be little talk about cell-based therapies, but since the two of you are experts in this field, would you go over indications and where research is heading for tumor-infiltrating lymphocytes (TILs) and adoptive T-cell transfer?

Cell-Based Therapy for Melanoma

Inge Marie Svane, MD, PhD: At the moment, we are using adoptive T-cell therapy, based on the TILs, mainly in melanoma. We have ongoing studies in Denmark but also in collaboration with a Dutch center. We are also trying to move into other indications. We are very interested in trying to develop a cell-based adoptive cell therapy for rare cancers.

Dr Zimmermann: Where do you see this fitting in the standard of care for these patients? Obviously, it's not for everyone; these will be selected patients. These are also technologies that are not widely available, so centers need to be specialized. Will this hit the mainstream at some point?

Dr Svane: That is difficult to say. You are right. First of all, production of [TIL therapy] needs facilities and know-how. It's also a very intensive treatment, of course, so the patient needs to be fit. The age limit is around 75 years. Also, [we look] for different kinds of comorbidities which could be so severe that patients could not get through treatment, especially with high-dose chemotherapy, where you are lympho-depleting the patients. It's quite intensive chemotherapy.

George Coukos, MD, PhD: Today, we are closer to envisioning TIL moving into the standard of care for melanoma. Its efficacy has been very well established in hundreds of patients, and it has been producible across centers. As Inge Marie said, the main toxicity issues relate to the administering of high-dose chemo, which requires hospitalization, and the following interleukin. The current regimen is intense. Clearly it will have to be given for failures of IO therapies in patients who have no additional therapeutic opportunities left.

What is really exciting is that TIL therapy is a single infusion. When it achieves a complete response (CR) it can be potentially curative. It does achieve a CR in approximately a quarter of the patients with current therapies.

It's a very interesting therapeutic approach that merits being fully developed. Obviously, in solid tumors we will have to explore opportunities for TIL. If we manage to demonstrate that it is potentially curative, then we will have to see the appropriate window of opportunity for each disease.

Clearly, there are now multiple lines of therapy for many solid tumors, but for some solid tumors even the front line is ineffective. If TILs manage to make an impact, it would be transformative—a game-changer for solid tumor therapy.

Dr Zimmermann: Quite a sizeable minority of patients actually achieve a complete remission that is durable, but there is certainly opportunity to increase this percentage. In what direction is the research currently going?

Dr Svane: In melanoma patients, we are studying it in combination with a BRAF inhibitor. For the rare cancer patient, it may be good to take advantage of the checkpoint inhibitors. They work in different time points in the cascade reaction in the immune system. You can probably increase infiltration of T lymphocytes by using the anti CTLA-4 antibodies before you harvest the tumor lesion. You need the first tumor lesion to harvest your TILs. Then, protect the TILs that you infuse by adding on the anti-programmed cell death 1 (anti-PD-1) antibody.

Car T Cells

Dr Zimmermann: TIL therapy is just one aspect of cell-based therapy. There has been a lot of talk now about chimeric antigen receptor (CAR) T cells, which are used in basically CD19-positive hematologic malignancies. Do you have the feeling that this approach can be validated in solid tumors also?

Dr Coukos: Having seen the power of CAR T cells in CD19 malignancies, it's clearly an incredible opportunity. Clearly, whether this can be reproduced in solid tumors is the big question. There are two important questions: Can we find a marker—a target—that is clean enough to produce therapeutic results without major toxicity? Presently there are very few such markers. For example, one marker also applicable to ovarian [cancer] is mesothelin. Some early experiments show that this could be safe enough to create a therapeutic window.

Dr Zimmermann: Very little on-target, off-tumor toxicity with this.

Dr Coukos: Exactly. The second question is, is the microenvironment of the solid tumor as permissive for CAR therapy as it has been for leukemia and lymphoma? These are the big questions in the field. A lot of people are working in this area. We know that the T-cell receptor, which is the alternate approach, has given a very strong signal for NY-ESO-1 (a very common tumor testis antigen in sarcoma patients), for example. There are also some CRs in that cohort. That is advancing from the commercial standpoint. There will be a significant amount of research in this area to see how we can best fit this technology to treat patients.

Dr Zimmermann: One hurdle is the fact that many of these antigens are actually private, or neoantigens—not talking about CARs, but for the TILs. There is certainly a desire to select the TILs that are the most avid or the most effective against these mutations. Are we close to being able to identify which antigen is actually driving this anti-tumor effect?

Dr Svane: We are really close to doing that.

Dr Coukos: We are quite close. There is now very reliable technology to identify the antigens, and it is improving constantly. We started by looking at genomic information—mutations based on exome sequencing data. Now there is a proliferation of mass-spectrometry platforms that deliver very reliable data.

Dr Zimmermann: With this you would identify the peptide?

Dr Coukos: Exactly. Identify the peptide that sits on the tumor cell directly, and use genomic data to validate that. Once you identify it, the third level of validation is to make sure that there are TILs that recognize it. If you complete these three steps, you are good to go. There are several groups worldwide dedicating significant resources, including the Ludwig Institute for Cancer Research in Lausanne. We think we are quite close to that. The future is certainly going to be in this personalization. Today it will take significant resources and a lot of time. As we know, technology always produces solutions that make things go faster and get cheaper. I think this is the future.

Dr Svane: You also need to show that they are able to actually kill the tumor cells. It was published earlier this year that not all of these new epitope-specific T cells are able to kill. Maybe a combination of T cells directed against shared antigens and private epitopes is the perfect blend.

Overcoming Heterogeneity

Dr Zimmermann: Heterogeneity is also an aspect because we now have seen with CAR T cells that invasion mechanisms sometimes limit efficacy.

Dr Coukos: The obvious approach for the heterogeneity is to chase the tumor with a clonal evolution. Clonal antigens are also coming up. In a way, you can imagine a longitudinal approach where you chase the tumor with new clones. People have shown that there are ways to evade this immune attack by just downregulating some pathways.

The tumor has the ability to mutate very fast, so there will be emerging clones that completely evade these approaches. The same happens with targeted therapy. We are facing a very complex and plastic enemy, but increasing technology will, hopefully, allow us to get a step ahead of it and eradicate it in more patients.

Dr Svane: It has been shown for the checkpoint inhibitors that there is a correlation between clinical benefit and the mutational load in the patient. Most recently, we have shown that it's actually the same for T-cell therapy, despite the fact that they have progressed on checkpoint inhibitors before[1]; there still is a correlation between the level of mutations. We need to go further into the private epitope as well.

Dr Zimmermann: Also, as medical oncologists, we want to know when to implement or how to sequence these treatments. Can we expect phase 3 trials with this cellular approach?

Dr Svane: Together with John Haanen in Amsterdam, we are running a phase 3 trial where T-cell therapy is randomized against ipilimumab in melanoma patients. This is a second-line protocol, after they have progressed on anti-PD-1. Hopefully we will answer whether we can achieve the same level of response when they are resistant to anti-PD-1.

Optimal Duration

Dr Zimmermann: These are exciting developments—absolutely. At this Congress there has been a lot of talk about immune checkpoint inhibitors, moving from later lines of therapy in advanced cancers to earlier disease settings like consolidation after chemo/radiation for locally advanced non–small cell lung cancer, or even adjuvant settings—in melanoma, for example. One question that always pops up is, what is the optimal duration for these therapies? I would be curious to have your insights. What do you do in your clinical practice?

Dr Svane: At the moment we do not know. Caroline Robert had a nice talk [at ESMO 2017] about how they decided to stop patients when they were in complete remission. They also have data from the study[2] on pembrolizumab, which lasted 2 years.

In Denmark, we decided that there are no stopping rules, and we also decide dependent on how large of a clinical response was achieved. After [achieving] complete remission, if it's verified at the next evaluation, we can stop. When you have the maximum level of partial remission and you have verified that, you can stop. The most difficult patients are the ones who achieve stable disease, because we really do not know whether they are in a kind of balance where they still need the push from the checkpoint inhibitors.

Dr Coukos: I completely agree. Now that IO drugs are emerging in the adjuvant setting, I think that's a completely open-ended question. What will we find out after these studies are completed? What will be the best practice for patients who have measurable disease and achieve a CR? I think that is the easiest decision. Currently, the discussion is, will it be 2 years, 1 year, or 6 months after having achieved a complete remission?

It depends also on how well the patient is tolerating therapy. Obviously, if there is absolutely no toxicity, you are more prone to give it a longer go. As Inge Marie said, patients who have stable disease, but still measurable disease, are the toughest calls. From biopsies that have been performed in that context, we know that in some patients it's just fibrosis left behind. In some other patients, you have small clusters of tumor cells persisting in a sea of T cells, so that would call for continuation of therapy.

I would imagine that as we become more comfortable with these approaches, we could start giving patients treatment holidays—2 or 3 months, for example—where you do nothing and then go back. These therapies leave a long memory. It takes several months for the immune system to readjust and try to evade. So we could do treatment holidays, as we do with targeted therapy, but it will take time to optimize the approach.

Dr Zimmermann: If I hear you right, you would definitely stratify your approach according to the best overall response and tolerance.

Dr Svane: You have to take into consideration that the risk for toxicity is the same all along—you accumulate risk. For example, you can suddenly develop diabetes when you have been in treatment for 1.5 years. It is important to know that it's not just [an issue] in the beginning and then you are home safe if you have not had any toxicity within the first 6 months. Fatigue is also a major problem.

Dr Coukos: Absolutely. Research needs to go toward understanding which patients are more prone to develop toxicity, where and when toxicity arises, and whether organ-targeted approaches to attenuate that toxicity can be developed.

Patient Selection

Dr Zimmermann: There is also a lot of discussion on better biomarkers, because it seems that we are still stuck with immunohistochemistry (IHC) for PD-L1 expression and there is no standardization of mutational burden, whether by gene panel testing or whole genome sequencing. Where do you see opportunities for better selection of patients and better algorithms of treatment? How can we allocate the new combinations to patients who will actually benefit according to their microenvironment?

Dr Svane: There are a lot of possibilities. The fact is, if it's not an easy test, it's difficult to implement in the clinic. It needs to be rapid and in black and white. You need to be sure that you are not telling your patients that they cannot receive this therapy if they just have a smaller chance of benefit.

We need new tests, but we also need to look more deeply into the characteristic of the individual patient—for example, [the speed at which the disease is developing]. Even though you might be having the same response rate with pembrolizumab, it might be that your patient needs a very rapid response, and therefore you are going for the combination because you know that you will have a chance of response within a couple of weeks.

Dr Coukos: For biomarkers, we have PD-L1, which has been standardized. Mutational burden is rapidly emerging from very good data. The good news is that, very likely, one can measure mutational burden just by the 400-gene panel, which would have been implemented in the clinic, also for personalizing targeted therapy. This could be a very valuable test, and at some point it will be implemented in all patients with relapsed disease or in early high-risk patients.

Beyond that, we must develop better biomarkers that can fit in a pathology laboratory. Tissue-based markers require a better understanding of the biology. If we know the biology well, biomarker development is easier. At this stage, what appears like low-hanging fruit is the presence of activated T cells in the microenvironment. This has to be standardized; there is very little effort at the moment to harmonize and standardize these and to develop a score. Circulating biomarkers are, again, an opportunity, but some additional work will be required.

Combination Therapy

Dr Zimmermann: Are you watching any new combinations closely at the moment?

Dr Coukos: One could classify patients in two large categories: responders and nonresponders. If you look at the underlying biology, one could say that responders are those who raise T-cell-inflamed tumors. There, I think the emphasis will be on how to make the immune response better. Clearly, immunosuppressive pathways that restrict T-cell function are the obvious opportunities, such as with checkpoint inhibitors or combinations of IO drugs, and eliminating the suppressers.

A very important, emerging pathway is the adenosine pathway. We are going to see some very interesting data in that space. Indoleamine 2,3-dioxygenase (IDO) inhibitors with PD-1 blockade is emerging as a powerful combination.[3] This will be very interesting for patients who have an existing response. I think targeted therapy with bona fide targeted oncology drugs will be the opportunity for patients who have no T cells.

As oncogenic pathways probably drive these immune desert tumors, as we know more about [these] pathways—PTEN loss, etc—there will be opportunities for targeted therapy to combine with immune therapy. The open question is, how does targeted therapy impact the immune system?

One pathway that I'm very excited about is [homologous recombination deficient] tumors and opportunities to interact in the DNA repair pathways. For example, PARP inhibitors, [as well as] other inhibitors, could produce profound immunomodulation in these tumors and become very interesting tools. Combinations and schedules will be extraordinarily important.

Dr Svane: A third population is patients who are refractory to the anti-PD-1 [therapies]. This group of patients is growing, and we need to look deeper into the resistance mechanisms because that is where we will actually learn, and then we will go back and start [using] the combinations. We will have to look further into that group instead of expanding our treatment for first-line. At the moment, we are stuck with anti-PD-1, and we need something for this majority of patients who either do not respond at all or who get some type of resistance to PD-1.

Dr Zimmermann: Thank you very much for a fascinating discussion. Thank you, Inge. Thank you, George. This is Stefan Zimmermann, speaking from ESMO 2017 in Madrid, Spain.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: