Thomas Kipps, MD, PhD, discusses the details of the integrated analysis from the iwCLL meeting, his ongoing research with ROR1 as a target in chronic lymphocytic leukemia, and pivotal combination studies being conducted with ibrutinib.
Thomas Kipps, MD, PhD
High-risk patients with chronic lymphocytic leukemia (CLL) who have an 11q deletion—said to be an adverse predictor of poor outcomes—can have durable responses when treated with the BTK inhibitor ibrutinib (Imbruvica), according to pooled analyses of results from the phase III HELIOS, RESONATE, and RESONATE-2 studies presented at the 17th International Workshop on Chronic Lymphocytic Leukemia (iwCLL) Biennial Meeting.
The findings suggest that risk factors typically associated with poor clinical outcomes may be less relevant with ibrutinib treatment, explains lead study author Thomas Kipps, MD, PhD, professor of medicine, Evelyn and Edwin Tasch Chair in Cancer Research, and deputy director of research operations at the UC San Diego Moores Cancer Center, who presented the findings at the meeting. In an interview with OncLive, Kipps explained the details of the integrated analysis from the iwCLL meeting, his ongoing research with ROR1 as a target in CLL, and pivotal combination studies being conducted with ibrutinib.Kipps: We were involved in these clinical trials, but I was the senior author of the trial that we put together of patients being randomized to treatment with ibrutinib versus chlorambucil. These were patients older than age 65 and patients who had some medical comorbidities. Particularly, in Europe, chlorambucil is not a bad choice by European standards, although it is less frequently used in the United States. We had the trial where patients were randomized to receive ibrutinib versus chlorambucil as part of the RESONATE-2 study, and what we observed in that was that patients treated with ibrutinib fared better than patients treated with chlorambucil. There was a significant improvement not only in progression-free survival, but also in overall survival despite the fact there was some crossover. There seemed to be a distinct advantage for patients treated with ibrutinib.
Based on that study, the FDA has approved the use of ibrutinib for patients as initial therapy, independent of age or medical comorbidities, so it was a significant study. What we noted in some of our patients treated on ibrutinib was that we had patients with deletion 11q who did fairly well. It was striking because, historically, patients who have CLL who have deletion 11q tend not to fare as well with chemotherapy or chemoimmunotherapy.
In fact, although patients may achieve a good response to therapy, they seem to have a short period of time before they relapse. Therefore, it has been called an adverse prognostic marker. This marker had been pointed out in the paper by [Harmut] Döhner and colleagues in The New England Journal of Medicine in 2000, and it showed that patients whose CLL cells had deletion 11q fared second to worst to only the patients who have deletion 17p—in terms of outcomes and survival.
Now, we had cytogenetic and follow-up data on patients treated in the era of chemoimmunotherapy, and we looked at survival rates by virtue of the cytogenetic features. What we found was that the survival of patients with 17p had improved quite substantially, but the survival of patients with 11q had not improved as much as the patients with 17p. It was still an adverse risk marker compared with patients who had other cytogenetic lesions or no detectable cytogenetic lesions.
We published on that, and you can look at those data in comparison to Döhner’s study. It seemed that even with the advent of an antibody added to chemotherapy, in the era of chemoimmunotherapy, these patients were still having a relatively adverse prognosis. The observation that patients with 11q were doing better in the RESONATE-2 study was noteworthy.
We then got the data together for other studies that were done with ibrutinib, mainly the RESONATE study—the original one—which compared patients treated with ibrutinib versus patients treated with ofatumumab (Arzerra). We also had the data from the HELIOS study, in which patients were treated with bendamustine and rituximab (Rituxan; BR) and placebo versus BR and ibrutinib. We wanted to assimilate all of the data for patients treated on the ibrutinib arm and to compare how patients did with these different cytogenetic lesions. The reason we remarked on that was because, in the early phase II studies done of the preregistration studies of ibrutinib, enrolling patients who had multiple rounds of prior therapy, it did appear that the patients with 11q minus were faring worse than the other patients—except for those with 17p.
When we did this, we found that the analysis revealed that patients with 11q, when we assimilated the data from all of these groups, actually did better than patients who had no detectable cytogenetic abnormalities. Now, the P value is approaching significance, so if anything, we can say, with some assurance, that the patients are not faring worse. We need to follow these data along.
I must say that, with the HELIOS study, patients did not receive single-agent ibrutinib, but also got chemoimmunotherapy, mainly BR. As a consequence of that, it is not a totally clean study in terms of only patients receiving ibrutinib. But the numbers were useful to have by assimilating this other study so that we could look at the other prognostic groups.
One of the take-home messages of this study is that, yes, ibrutinib does quite well. The comparator arm treatment showed that patients with 11q deletion were not faring as well but [that] with ibrutinib, patients were faring better. It gets into this issue of whether 11q is as much of an adverse prognostic marker, particularly in the era of newer targeted therapies, so that is something noteworthy.ROR1 is an orphan receptor, and it was found in the 1990s. You look for genes that encode proteins that obstruct assemblance of what we call tyrosine kinases (TKs). These are enzymes, and the purpose is to put a phosphate group onto a protein and that changes the biology of the protein. It is a way that the body uses to signal. For example, [Bruton’s tyrosine kinase] is a TK, and these TKs are really researched. We know that ibrutinib, which inhibits TK, has therapeutic utility. In any case, a lot of research was done for all different TKs in the body to identify the kinases that our genes encode.
ROR1 was found along with another receptor, ROR2, and they were orphans because they were not like the other kinases. Nobody knew what they did or what their ligands were or what purpose they had. We had been working on that now for several years because we did an earlier study where we wanted to get patients to develop immune responses against their leukemia, so we engineered the leukemia cells to make them better vaccines. We gave patients their own leukemia cells, and some patients actually did pretty well after that. We noted that some of these patients made antibodies that reacted against their leukemia. Therefore, when we figured out the antibodies, we actually looked to see if those same antibodies reacted against normal cells, and they did not.
We wanted to go after that. What we observed was that no antibodies reacted against ROR1, which we found to be on CLL cells but not on normal lymphocytes. Unlike CD20, which is on normal lymphocytes as well as CLL cells, this was only on CLL cells. The work done by embryologists and developmental biologists [has] found that ROR1 comes primarily from embryo genesis, so we use it in early development and we turn it off in adult life. It's interesting that leukemia—in adult life—turns this protein on somehow. We are trying to figure that out, as well.
We are asking the question whether this is playing some role in the biology of leukemia. On one hand, it might represent a useful means to target CLL, but it is not expressed on normal tissue. If you have an antibody or even a CAR T cell that reacts against it, you might be able to spare the normal cells, which is not what we do when we give drugs such as rituximab or obinutuzumab (Gazyva). We do get loss of normal B cells. Independent of whether it’s a good target or not, we have been trying to work out how this actually works.
We have been developing mouse model systems to look at the role of ROR1 and how that might influence the development of leukemia or the progression of leukemia. To make a long story short, we can engineer leukemia with or without this protein and we found that, in mouse models of leukemia, if they have the protein, then they have much more aggressive disease. The animals die much more quickly. Those seem to be playing a role in causing the disease to be more aggressive.
We recently completed a large study with the CLL Research Consortium, [where] we looked at 1500 patients, and we examined the level of ROR1 expression. The level of ROR1 expression was on almost all patients. However, some patients had very small levels and some had higher ones. In a small number of patients, we found barely detectable levels of ROR1, so it’s a continuum. We did an analysis of the so-called ROR1-negative cells of patients and the ROR1-positive cells, and we found differences in gene expression by that measure that mirrored what we saw in the mouse leukemia that had ROR1 versus not having ROR1.
But when we looked at the outcomes of these patients, those with high levels of ROR1 actually progressed more rapidly and they required therapy more rapidly. They had a shorter survival than patients with low ROR1 levels, which is quite striking. We think that in the mouse models, but also in patients [with] CLL, that ROR1 may be playing a role to drive progression.
Now, we have developed some agents to attack ROR1. We have a monoclonal antibody that we designed specifically to the ROR1 protein, and we are going to validate this. We embarked [on] a clinical study using the antibody with CLL. That is a very exciting [development] because we're able to target this protein and the antibody that we developed to try and block the function of this ROR1 protein. As mentioned, ROR1 may be playing a functional role in driving the leukemia process. If we can inhibit that function, that may have a therapeutic advantage.
What we sought was to try and understand how ROR1 works, and we spent a lot of time doing this. ROR1 actually binds WNT5A, and that is one of the ligands that we identified for ROR1. We found that when you add ROR1 to leukemia cells, it really activates a signaling cascade that we have been mapping out. It activates another set of signal molecules that can play a very potent role in driving cancer progression.
There is a lot of biology on these signaling molecules. What we found was that the antibody that we developed for clinical trials can actually block that signaling. It's somewhat akin to ibrutinib, which blocks the signaling by the IG receptor, which serves as a growth or survival factor for leukemia. Therefore, ibrutinib can block the signal pathway for the antibody. This antibody against ROR1 can block the signaling pathway for ROR1. More recently, we have been looking at this from the standpoint of, how will this work with ibrutinib?
We have looked at patients with ibrutinib treatment, and we found that when they’re treated with ibrutinib, it does block the B-cell receptor signaling pathway. However, the pathway that we have identified for ROR1 is an overdrive; it's almost acting like a lifeline for the leukemia cells. It may count, in some measure, for why the leukemic cells persist for so long. It may account for why patients treated with ibrutinib—at least in the first year or two of therapy—hardly ever get a complete response (CR) in that the cells are being fueled by another survivor pathway.
We are very excited by the phase I study. We chose effects similar to ibrutinib, but with an antibody, we get some redistribution, reduction in lymph node size, and the important thing is that it seems to be very well tolerated. We have only given a few doses to patients just to test the safety. However, I am even more excited to combine this then, because when we combine it with ibrutinib, we notice that we are blocking not 1 signaling pathway, but another signaling pathway. It is sort of like taking the legs out from under a table. If you take out 1 of 4 legs, the table can still stand. But, if you take out 2 legs, the table falls over. That is what we are hoping to do in leukemia.
Anecdotally, we had some patients who were treated with this antibody and then they started ibrutinib afterward. We had some very dramatic reductions in leukemia cell count and in CRs that were occurring within 5 months of treatment. Giving these together is going to make a lot of sense, and we are hoping to launch a clinical trial to test that.
This is something that is exciting because this is attacking the leukemia by attacking its biology. This idea of attacking its biology may result in more complete and durable remissions and may potentially have patients treated with these drugs for a period—not have to be on lifelong therapy.
I like to call back to the terminology of “The Terminator” who, at the end, is crushed by this machine and is totally immobilized. But then it seeks to rewire itself and finds a detour route and is able to rise again. I think of these tumor cells like “The Terminator,” [there is] always a way around our efforts to thwart them.What is important is that we should be attentive to what does not work well. For high-risk patients—those patients who have deletions in 17p and possibly 11q deletion—it may be that chemotherapy or even chemoimmunotherapy may not be a good option. Emphatically, patients with deletions in 17p should not be given chemotherapy because it can actually make things worse for the patient.
I have been advocating that for some time. When we developed the RESONATE-2 study of comparing patients treated with either ibrutinib [or] chlorambucil, we excluded those patients who had deletion 17p because I felt it was unethical to give patients with 17p deletion chlorambucil. We excluded those patients from the trial because you would have some patients with 17p deletion possibly being randomized to receive chlorambucil. Before treatment, you should at least assess high-risk patients with fluorescence in situ hybridization (FISH) and look at the complexity of the genome by doing a meta-phase analysis.
There are ways of doing the meta-phase analysis to stimulate the cells and then you get a better-quality study. You can look for what we call complex karyotype. That has a certain prognostic significance, too, in terms of genomic stability. If you find a 17p deletion, one should really avoid giving chemotherapy to those patients. A newer targeted therapy such as ibrutinib, venetoclax (Venclexta), or other targeted therapies that may be coming down the pike may be better for patients than giving the same therapy that we know does not work very well.
That is very important to know if you can get a profile of the genetics of the patients. You should know that genetics change over time, particularly after therapy. I saw a patient in clinic who was treated with BR and did not have cells with 17p deletion as detected by FISH. A few years after he was treated, he relapsed and progressed. We looked at the genetics, and now, most of his cells have 17p deletion. If it has been tested before, you need to test it again and make sure that these high-risk genetic markers are not acquired in the course of the disease or in the course of treatment of the disease.I am very keen on the combination of ibrutinib plus this antibody against ROR1, which has been called cirmtuzumab. Clinical trials are beginning to test that. Another combination, which is attractive, is the combination of this with venetoclax, which is a BCL-2 inhibitor. This works with different mechanisms than ibrutinib, so you are, again, attacking 2 different mechanisms.
Venetoclax is a very powerful drug. The problem with venetoclax is that it’s almost like an ice cream headache—in that you get too much tumor lysis and you can get into serious trouble or even die, because too much of a good thing is a bad thing.
We have looked at ways of stratifying patients by their risk of having a bad tumor lysis, and you can imagine that based on an estimate of what size and tumor burden the patient has. If it is a high tumor burden, they are at risk for developing tumor lysis; if you have a low tumor burden, you have a low risk for tumor lysis. With ibrutinib and venetoclax, we are now trying to test the combination and even seeing whether treatment with ibrutinib may lower the tumor burden. That might allow for the lower risk of tumor lysis syndrome in patients who had therapy with venetoclax.
Other antibody combinations have been done with ibrutinib. Ibrutinib has been paired with rituximab, obinutuzumab, and BR. My take on the data is that it’s hard to know whether the addition of the antibody or chemoimmunotherapy, in the case of the HELIOS study, has improved the outcome of patients treated with ibrutinib by itself. That is an important question because if you can get some mileage out of getting something in addition to ibrutinib, then I’m all for it. If it doesn’t seem to make much difference whether you get the antibody or not, then it’s hard to justify the time and expense of getting that.
There are some things about ibrutinib that might limit the effectiveness of the antibody. A theoretical basis group at Ohio State University has been claiming for some time that ibrutinib may make antibodies, such as rituximab, less effective in clearing the leukemia, and they have argued against combinations. The jury is still out on that combination.This is under investigation. One does not like to go onto therapy that is considered lifelong; it does not have the prospects for longevity. You have to take anticancer drugs for the rest of your life, so more patients assume that their life is not going to be very long. Also, the cost of the drug and any complications from taking the drug then get put onto a lifelong picture and it’s sometimes not easy. If there are very small side effects with the drug and you having to take the drug forever, patients don’t like that.
The idea of trying to get deeper and more long-lasting remissions by combinations is something that we are all trying to achieve. If that can be done, then there is a prospect of having patients stop at a point in time. By looking at minimal residual disease (MRD), we can assess that in the blood or in the bone marrow. We have sensitive tools to look at that, with 1 CLL cell in 10,000, which is the current benchmark. Or if you use sequencing, it is 1 cell in 1 million. We are getting to the limit of the sensitivity because our sample size is only so big.
Let me put it this way. If you only collect 10 cells from a blood sample and you didn’t find the leukemic cells, all you can say is we didn’t find any in the 10 cells that we looked at. You don’t know if it’s 1 out of 20 that leukemic cells are there; it’s just [the] luck of the draw at that point. The ability to go deeper for MRD is going to be limited by how many cells we can safely acquire for that analysis. One in 10,000 cells or 1 in 10 million is what we have, and that is a pretty sensitive measure. If we can start achieving that level of clinical response, it argues well that some of these patients may have even deeper remissions and might be able to go off therapy and stay off of therapy for some time, if not be cured from their disease.