CRI Speaks with 2019 AACR-CRI Lloyd J. Old Award Recipient, Dr. Cornelis ‘Kees’ Melief

This Tuesday, April 2, Cornelis J. M. ‘Kees’ Melief, MD, PhD, will receive the 2019 AACR-CRI Lloyd J. Old Award in Cancer Immunology presented by the American Association for Cancer Research (AACR) and Cancer Research Institute (CRI) during the 2019 AACR Annual Meeting.

Dr. Melief, a member of the CRI Scientific Advisory Council and CLIP Grant Review Committee, is being honored for his work on the development of effective immunotherapies for virus-induced tumors, including a novel vaccine strategy that uses synthetic long peptides (SLP) to target cancers associated with infection by the human papilloma virus (HPV). This vaccine approach has been clinically effective in patients with pre-malignant vulvar lesions, head and neck cancer (in combination with checkpoint immunotherapy), and clinically promising in advanced cervical cancer (in combination with chemotherapy).

Currently, Melief is the chief scientific officer at ISA Pharmaceuticals, emeritus professor of tumor immunology at the Leiden University Medical Center in the Netherlands, and a member of the Royal Netherlands Academy of Arts and Sciences.

Recently, I spoke with Melief about his career accomplishments, what receiving the Lloyd J. Old Award means to him, and the future of cancer immunotherapy. The following conversation has been edited for clarity and length:

Arthur Brodsky, PhD:

Before we get into the science, I was wondering, what was your relationship with Lloyd Old like, and what does it mean for you to be honored with the award that bears his name?

Cornelis J. M. Melief, MD, PhD:

That’s really a great honor because we had numerous interactions in the past. We even had a collaborative, CRI-funded cancer vaccine project. Before these collaborations I used to visit the annual CRI scientific meetings in New York, which were held in a wonderful theater near Broadway. Attending were only around 100 people or so, which was at that time the entire world population of cancer immunology and immunotherapy experts believing in immunotherapy of cancer. With Lloyd Old I not only discussed science but also classical music, of which I also have very good memories.

He was an inspirational figure in the then rather small cancer immunotherapy community. So I think he was a father figure to many of his PhD students, postdocs, and colleagues over the years, and he was, of course, an excellent leader of the Scientific Advisory Council of the CRI and of the Ludwig Institute for Cancer Research.

Arthur Brodsky, PhD:

This award recognizes your many contributions in cancer immunology, and specifically in the area of cancer vaccines. Before we get into your work more, would you mind first explaining what vaccines are in the context of cancer?

Cornelis J. M. Melief, MD, PhD:

Of course, everybody knows vaccines as preventive vaccines, and they’re probably the greatest medical invention of all time in terms of gained lives that would have otherwise been lost. The ironic point is that many people don’t appreciate them that much anymore these days, as witnessed by the large number of people who are anti-vaccine activists, because the vaccines have been so successful that the horrors that were there when there were no vaccines have been out of the memory of people. People don’t appreciate anymore which dangers are lurking in the dark in the absence of preventive vaccines.

But that’s a completely different story. And so, where preventive vaccines work largely by the induction of antibodies, which then upon first contact with an invading microorganism immediately kill it, therapeutic vaccines need to work through cell-mediated immunity. So the type of vaccines that we and my colleagues have been developing in the area of cancer immunotherapy are particularly good at eliciting T cells against tumor-associated antigens.

The wonderful fact about T cells is that they are not diverted by freely floating antigens in the blood or in other body fluids but are highly specialized in the killing of abnormal cells like cancer cells. So they are highly goal-oriented in cancer cell recognition and killing.

Arthur Brodsky, PhD:

When it comes to the preventive vaccines you mentioned, they are usually against bacteria and viruses, which look a lot different from our normal cells, and it’s this foreignness that makes the immune system able to go after them. What is it about cancer cells, which arise from our own cells and so look comparatively normal to the immune system, that can help these vaccines be effective against them, too?

Cornelis J. M. Melief, MD, PhD:

I think there are several conditions for the success of cancer vaccines. One is the proper choice of good target molecules on the cancer cells. Of course, these could be viral targets because there are quite a few cancer viruses, but the second most attractive category that has come up in recent years are neoantigens caused by mutations in cancer cells.

But we should not forget about a third category that Lloyd Old was again very enchanted with, which is overexpressed antigens for which there is a T cell repertoire, such as selected cancer antigens, or in the case of melanoma cells, differentiation antigens.

Arthur Brodsky, PhD:

You helped develop the synthetic long peptide (SLP) vaccines that are used to educate the immune system about what cancer looks like. You have demonstrated their ability to help patients who have cancers that are associated with infection by the human papilloma virus (HPV), in particular cervical cancer and head and neck cancers. Could you share a little bit about the breakthroughs that you have helped make in this area?

Cornelis J. M. Melief, MD, PhD:

Yes. I was going to mention the second important condition for the success of a cancer vaccine, which is to have the proper vaccine platform. Indeed, I think there are now three successful cancer vaccine platforms: DNA, RNA, and synthetic long peptides.

Of these, I argue that SLPs are the closest to what actually needs to be presented by HLA molecules on antigen-presenting cells and cancer cells. So they are one step away from further trimming and loading onto HLA molecules, which is an essential condition for stimulation of cancer specific T cells. [Editor’s Note: These HLA molecules are part of the major histocompatibility complex, or MHC, structures that our cells use to display certain molecules for recognition by the immune system.]

I think the breakthrough came when we first showed in mice that synthetic longer peptides were much more effective than short peptides because short peptides get loaded onto all cells in a body that have MHC molecules, whereas synthetic long peptides (SLPs) need to go through an essential processing step that only dendritic cells can perform effectively. Then, if you add the proper adjuvants to activate those dendritic cells, then you get a really potent anti-tumor T cell response. (In other words, the size of these synthetic long peptides (SLPs) makes them much better at stimulating the appropriate immune responses against the cancer cells that express those targets.)

So we showed first in mice that these SLPs were much better than short peptides, also because the long peptides combine antigens that can be recognized by both CD4 and CD8 T cells. CD4 helper cells are an essential ingredient of a successful T cell response.

First of all, the CD4⁺ T cells are helper cells for killer cells. Helped killer cells performed much better than non-helped CD8⁺ killer cells. Secondly, and this is a point not sufficiently appreciated yet by many people, is that the CD4⁺ helper cells have considerable cancer-killing function by themselves, either directly or through macrophages.

Cornelis J.M. Melief, MD, PhD

Arthur Brodsky, PhD:

It’s nice that you ended on that note, because that kind of leads into my next question, which is that vaccines at their core are just kind of delivering what I call the ‘intel’ to the immune system about what the cancer cells might look like, but your work has also showed that it is not just enough to provide that information. It is also necessary to help motivate the immune system or stimulate it to act on that information.

Cornelis J. M. Melief, MD, PhD:

Yes, yes.

Arthur Brodsky, PhD:

So, what have we learned about the complexity of the immune system, and how has our understanding of the immune system shaped the way we are now designing vaccine strategies?

Cornelis J. M. Melief, MD, PhD:

This came from our work when we first showed after the mouse studies that the SLP vaccine against HPV was active by itself against premalignant disease like vulvar intraepithelial neoplasia caused by HPV 16, which is the most frequent bad guy among the HPVs. It’s the one most frequently causing cancers. HPV 16 is responsible for about half of all cervical cancers, 85 percent of HPV-positive head and neck cancers, and many premalignant diseases like vulvar intraepithelial neoplasia (VIN).

So we showed that the SLP vaccine worked by itself, but that in late stage cancer patients, the vaccine was much less active and did not cause an appreciably strong immune response by itself and also, therefore, was not associated with any measurable clinical activity.

So we started to ask the question, ‘what is different between late stage cancer and premalignant disease?’ And we found out that it was the hostile cancer microenvironment, which prevented the T cells from expanding. But also, there were large numbers of suppressive cells in the periphery, in the blood.

Then, we started investigating what would be the easiest way to get those T cells to work better, and we investigated first the effect of chemotherapy. This was actually quite a successful approach because the standard of care chemotherapy that late stage cervical cancer patients receive—which is a combination of a platinum compound called carboplatin together with paclitaxel—depletes immunosuppressive myeloid cells but does not deplete T cells, nor does it inhibit T cell function.

On the contrary, because of the depletion of these immunosuppressive myeloid cells in both the blood in the tumor, the T cells started to function much better, which we can already see in the blood after chemotherapy alone. We saw a rise in the response of the T cells to HPV 16 and to common recall antigens, which was quite weak, but demonstrable.

And after an additional dose of vaccine, which was timed to utilize the lowest point of the myeloid cell levels after chemotherapy, we saw a huge T cell response, which was much larger than we had ever seen without the time to chemotherapy. So we published this after a single dose of vaccine in 2016, and we have now conducted a much larger trial that we will also publish soon, in which this is associated with clinical benefit. [Editor’s Note: myeloid cells are innate immune cells that can influence the tumor microenvironment and determine whether or not anti-tumor immune responses are suppressed or supported.]

Cornelis J. M. Melief, M.D, PhD, highlighted the use of HPV-targeting vaccines on Sunday, March 31, 2019, at AACR19.

Arthur Brodsky, PhD:

That was cervical cancer patients, correct?

Cornelis J. M. Melief, MD, PhD:

Correct. And then we investigated another avenue of interaction between the most widely used immunotherapeutic agent today, which is anti-PD-1 [checkpoint immunotherapy], and we found out that there is also a highly likely synergy between anti-PD-1 and this SLP vaccine against HPV 16 in patients with head and neck cancer. That was a study we published together with Dr. Glisson from MD Anderson. In that study, we saw a highly promising doubling of the overall response rate from between 33 and 36 percent to 60 percent, and we saw a doubling of the median overall survival from 9 months to 18 months, with three patients still in an excellent clinical stage two and half years after treatment, which is very rare for this type of patient.

Arthur Brodsky, PhD:

Yes, that’s remarkable. It’s great that these vaccines are coming to fruition, but I don’t know how much the public appreciates just how much basic research and how many decades of work went into understanding the immune system and the potential of vaccines before they actually get to patients. So for you, how does it make you feel to know that all this work is paying off and that it is actually helping to achieve the ultimate goal, which is to save patients’ lives?

Cornelis J. M. Melief, MD, PhD:

It’s highly rewarding, because for years and years, we were toiling away in mouse models because we had insufficient understanding of the basic workings of the immune system, collaboration between different cell types, dendritic cells, CD4⁺ cells, CD8⁺ cells, natural killer cells, and so on.

Understanding this, plus understanding that T cell treatment by itself is often not sufficient—you know, what I just described, the need in late stage cervical cancer patients for combination treatment to alleviate the immunosuppression in tissue and in the blood in these cancer patients—this is also a need one sees in adoptive T cell therapy, where, as you can appreciate, the patients need co-treatment with doxorubicin, which is an immune-depleting treatment.

So the conditions in the patients’ tissues and in the circulation are often such that you either need to make room for the T cells or you need to deplete hostile elements to make the treatment work, and we’ve increasingly started to know what these hostile elements are. Of course, the entire immune system is made of checks and balances. The most likely situation where you can actually achieve good successes with a single dose of treatment is the CAR T cell treatment, but don’t forget that the CAR T cells have been equipped with the activating signaling domains, and therefore they become more or less independent of the checks and balances that I was describing.

But that also sometimes has severe side effects. So we have to learn how to harness these extremely powerful cells. Combination treatment with vaccines is one relatively safe way to harness the awesome powers of T cells.

Arthur Brodsky, PhD:

Yes, and as you alluded to, we have so many more options nowadays when it comes to different ways to trigger the immune system and induce the it to go after cancer. But also, as obviously you are well aware, there is a lot of work to do.

Those combinations that you mentioned are great first steps, and now it seems like in a lot of cancer types, it is moving in the direction of not only more personalized approaches such as vaccines and figuring out what each patient needs, but also tailoring the combinations to each individual patient depending on which checks and balances their immune system is using at the time.

Cornelis J. M. Melief, MD, PhD:

Right. Yeah, it is probably wise to have a set of very well-selected markers to interrogate the tissue before treatment and then select exactly that combination treatment that could benefit that particular patient. So it will become highly personalized combination treatment.

For example, we now know that there is immunosuppressive TGF-b (transforming growth factor-beta), which is a cytokine that, in cancers like highly fibrous pancreatic cancer, keeps the T cells out. So, fortunately, one of my colleagues has developed an anti-TGF-b agent that might work in that situation in conjunction with other therapies that increase the numbers of T cells.

So if you allow the T cells to go in, then increasing their numbers by vaccine also makes sense. I think that combination is an attractive option for patients that overexpress TGF-b.

Arthur Brodsky, PhD:

I’d like to touch again on the importance of basic research to help us build that foundation upon which we can make even better treatment strategies for patients. Moving forward, what do you see as the most pressing challenges in cancer immunotherapy, and how do you think basic research can help us address some of those?

Cornelis J. M. Melief, MD, PhD:

I think we need to do further work on optimizing adjuvants [that can help stimulate adaptive immune responses]. There is great work on adjuvants, but only a few adjuvants have really made it to the clinic. And I think there is this entire category of TLR (Toll-like receptor) ligands, each of which needs to be investigated in detail.

So the next area for investigation would be working on improved delivery, and there have been already wonderful exploits showing that delivery using nanoparticles is a good way to go. For example, RNA packaged in particles could be used to efficiently channel both the adjuvants and the antigen into the dendritic cells. This might be a good way to go.

There has been interesting work in which you can further optimize vaccine performance by cytokines, such as IL-15, IL-7, and IL-2. IL-2 by itself is probably too toxic, but there are now variants of IL-2 that lack the toxicity but have retained the T cell expansion capacity.

And then there are numerous monoclonal antibodies of the TNF (tumor necrosis factor) receptor family. We ourselves have worked on anti-CD40, which I still feel is an excellent anti-cancer therapeutic because it is one of the most powerful activators of dendritic cells.

Unfortunately, upon systemic administration, anti-CD40 is fairly toxic. But we have shown in a mouse model—and there’s no reason why this couldn’t be used for human development, and I know people are working on it—many of the most powerful monoclonal antibodies known as immunomodulators, including anti-PD-1, anti-CTLA-4, and anti-CD40, can be delivered with great success locally instead of systemically.

For example, we’ve shown that for both anti-CTLA-4 and anti-CD40 therapy, if you deliver it locally, in the area of lymph nodes downstream of where the vaccine was delivered, you get exactly the same effects as systemic delivery of larger doses of the drug, because locally, you can deliver less, and you can also deliver it in a slow-release system, so that it can only get exactly where it matters.

What we also showed in the case of anti-CD40 is that the rather severe liver toxicity that was seen upon systemic delivery was completely gone upon local delivery. Since a vaccine is also locally delivered in most cases, I think you can combine vaccines with local immunomodulators plus adjuvants and expect to see even better effects.

Arthur Brodsky, PhD:

That’s great to hear. So, just following up on that a little bit, since you just laid out some of the challenges and some of the opportunities, what are your hopes for how we might overcome these challenges, and what do you think the field might be able to accomplish in the next few years?

Cornelis J. M. Melief, MD, PhD:

I think particularly in the area of vaccines there is a great opportunity to combine vaccination against neoantigens and overexpressed self-antigens, as well as viral antigens, with immunomodulators.

We are also quite excited to exploit the effects of chemotherapy, because the current view of chemotherapy is a very cancer cell-centric view. People expect chemotherapy to kill as many cancer cells as possible, so the drug is dosed at the maximum tolerated dose, but we know that if you combine it with vaccines, as we showed in mice, you can decrease the dose to, for example, 40 percent of the maximum tolerated dose and see no side effects or hardly any side effects compared to a full dose, because now you’re treating the immune system and not so much the cancer cells. So a triple therapy of a low dose chemotherapy plus vaccine plus anti-PD-1 is something I would be very enthusiastic about.

Then if you can rapidly analyze DNA and RNA for mutant sequences that are immunogenic, or immune-stimulating, in the context of the HLA molecules of that particular patient, you can quickly develop a personalized vaccine. Then, if you can combine it with low dose chemotherapy and immunomodulators—or with other drugs that come out of the advanced, specialized tissue analysis, like for example the Immunoscore™ developed by Jérôme Galon—then we can tailor your therapy to the specific needs for that particular patient. Then, next to the personalized vaccines and the chemotherapy and/or anti-PD-1, you can also think of adding other drugs like anti-TGF-b or anti-LAG-3, depending on what you observe in the tissue at diagnosis.

Arthur Brodsky, PhD:

That gives me really great hope that we are finally moving away from the one-size-fits-all strategy that’s mostly been used to treat patients.

Cornelis J. M. Melief, MD, PhD:

Exactly. Absolutely.

Arthur Brodsky, PhD:

As for biomarkers—PD-L1, for example—it made sense to look at that as the biomarker initially for PD-1/PD-L1 immunotherapies, but now we realize that it is imperfect to the point that even the patients that don’t have PD-L1 expression can still respond to those treatments.

Cornelis J. M. Melief, MD, PhD:

That’s another point. For example, if you have a tumor that is cold and has no T cells in it, it is unlikely to express high levels of PD-L1, unless it is an autonomous expression. But if there are no T cells producing gamma interferon, then a vaccine could generate enough T cells that would go in and produce enough gamma interferon to now make that tumor PD-L1 positive.

But if the T cells are there but didn’t go in because of TGF-b, then you also have to add anti-TGF-b to make vaccine-induced T cells go in, and so on. It’s a dynamic situation, and as you said, we have to look at the complexity and deal with it, and instead of one-size-fits-all, to actually personalize it to some extent. In large-scale clinical practice, this is of course quite a high challenge, but as time goes on, I think we will learn to discern groups of patients that benefit from not too complex and not too expensive therapy.

So we also need to look at our capacity to afford the treatments, and I think a combination of vaccines and chemotherapy is not expensive. Anti-PD-1 is expensive, but the price will go down as time goes on and the competition increases.

Arthur Brodsky, PhD:

And if we can come up with the right strategies to cure patients in the beginning, I think that saves a lot of money, too, if you don’t have to keep retreating patients. Obviously the patients would prefer that as well.

Cornelis J. M. Melief, MD, PhD:

Right. Because this is a great challenge, that, although many patients respond to anti-PD-1, outside of melanoma, the number of cures is not really to our liking.

Arthur Brodsky, PhD:

Yes, still less than half.

Cornelis J. M. Melief, MD, PhD:

Right. Absolutely. So we need to do better than that, and we need deep and complete responses, eventually, by making it more effective.

Arthur Brodsky, PhD:

It’s so interesting, going back to your point about how chemotherapy is a very blunt tool. We looked at cancer cell growth, so we are going to try to stop that growth. And that is, in large part, the rationale behind chemotherapy. But now, we have come to appreciate how much the cancer cells can actually evolve.

I’d go back to the interferon gamma signaling that induces PD-L1. But as Toni Ribas showed, sometimes the cancer cells can just lose the genes in that pathway, so that they stop expressing class I MHC molecules and then are no longer targeted and eliminated by killer T cells, and then it’s kind of going back to the drawing board and thinking, ‘Okay, now that the cells have reacted in this way, then what? What’s the next step? What would be the next treatment?’ And obviously we don’t know that yet for a lot of patients, but…

Cornelis J. M. Melief, MD, PhD:

Well, if this patient’s tumor cells no longer have proper interferon signaling and that’s why they are therapy resistant, one way to get at that would be to see if they shed enough tumor antigen and get it loaded on macrophages. If those macrophages are then activated by tumor-specific CD4⁺ T cells that recognize the antigen on the macrophages, rather than on the tumor cells, then those macrophages might kill the tumor cells without any expression of interferon driven antigen. Those mechanisms have already been demonstrated in animal models.

So I think the cancer immunology community as a whole is too much focused on CD8⁺ killer cells, because the imagination is stirred by the killer cells. But macrophages, if properly modulated and made to produce reactive oxygen species and nitric oxide, are formidable killers. That’s what they do all the time against bacteria. But they can also be mobilized against cancer cells.

So I am a proponent of also looking at the full capacity of CD4⁺ T cells, not only as helpers but also as direct or indirect killers.

Arthur Brodsky, PhD:

Mm-hmm. Even when the backer phages attack the cancer cells, isn’t that induced by either interferon gamma or TNF?

Cornelis J. M. Melief, MD, PhD:

Right. Absolutely. Absolutely.

Arthur Brodsky, PhD:

And that’s produced by the T cells? The T cells would be one of the cells that produce the interferon gamma, right?

Cornelis J. M. Melief, MD, PhD:

Yeah. This is what we have observed in our mouse model of vaccination, that if we combine macrophage depletion with vaccination, it worked much less well, although the macrophages in the original tumor before vaccination were characterized as bad macrophages, M2 macrophages.

But the interferon gamma produced by T cells that went into the tumor after vaccination produced so much interferon gamma that now the bad macrophages were modulated into good ones. So then you don’t want to deplete them anymore. So there are drugs promoted for cancer treatment, like CSF-1 receptor inhibitor, that would deplete macrophages. But I think it’s better to modulate a bad cell into a good cell than to deplete it.

Arthur Brodsky, PhD:

Yeah, it could have unforeseen consequences.

Cornelis J. M. Melief, MD, PhD:

Right.

Arthur Brodsky, PhD:

With the field being very focused on CD8⁺ killer T cells, it’s a constant struggle between adopting a paradigm and then trying to maximize the utility of that, because going down that path, you can only focus your energies in so many places at once. And it makes me think of something like natural killer cells, which I know have been starting to be used with CAR constructs, but I think it’s clear that they can kill cancer cells, but it seems to me they have also been one of the cell types that have been neglected as far strategies to utilize them.

Cornelis J. M. Melief, MD, PhD:

Right. For example, in the context of monoclonal antibody therapy, natural killer cells have excellent Fc receptors, a class of receptors that enable immune cells to interact with other cells. So if they’re combined with a monoclonal antibody targeting the tumor cell, like anti-CD20 for B cell lymphomas, the NK cells as well as the macrophages can engage in ADCC, antibody-dependent cellular cytotoxicity. This can be further amplified, as we showed many years ago, with cytokines like IL-2. So the natural killer cells can be motivated to do more killing than they would do spontaneously by additional immunomodulators.

Arthur Brodsky, PhD:

I want to go back to your earlier work, where you talked about your HPV vaccine. You first showed that in humans, in the patients who didn’t have advanced cancer yet, those who had premalignant disease—

Cornelis J. M. Melief, MD, PhD:

Correct.

Arthur Brodsky, PhD:

Now, obviously any kind of new treatment usually first goes through clinical trials with the patients who are most advanced. Immunotherapy, too, started out being used on the most advanced patients, but now it’s also going into clinical trials for patients with earlier stage cancers. And as your work has shown and also Bob Schreiber with some of his mouse models, these vaccines can oftentimes be more effective when it was introduced earlier in the course of the disease.

So can you talk a little bit about that potential too? With the exception of indications where immunotherapy has been approved as frontline therapy, in some ways it’s used as a last resort, after chemotherapy or something else fails. But can you talk a little bit about the transition that’s going on in the field right now to help patients earlier?

Cornelis J. M. Melief, MD, PhD:

Yeah. This is an absolute revolution going on. So the use of, for example, PD-1 and CTLA-4 blockers—but in particular anti-PD-1 because it is less toxic—for adjuvant (post-surgery) treatment of high-risk melanomas is extremely attractive because it seems to prevent recurrences.

And one sees the promotion to earlier stage of disease with all of these immunotherapy drugs. So pretty soon, it will be the first choice of even treating patients very early on, and also in diseases like lung cancer. So it seems to me very attractive also in those situations. I think the combination of anti-PD-1 and chemotherapy being advocated in lung cancer, and approved for that, I think it has already proven to be superior to chemotherapy alone.

And we also know in the diseases that we are studying that, for first-line HPV positive head and neck cancer, anti-PD-1 is also likely to be approved quite soon.

Arthur Brodsky, PhD:

That’s great to hear. Always good to have more options for the patients. Well, it’s been great talking with you about all of these different immunotherapy-related topics. Thank you again for taking the time to speak with me Kees, and congratulations again on your award!

Cornelis J. M. Melief, MD, PhD:

Well, it’s my pleasure. The CRI has been, as I said, directly stimulating our research, including one project on HPV. I think it’s a lean and mean organization, and I have been very much enchanted with CRI over the years, and people like Jill [O’Donnell-Tormey, PhD, CRI’s chief executive officer and director of scientific affairs], whom I’ve known for years. I have had many fruitful interactions and she is arguably the most effective executive of any cancer research organization that I am aware of.

Drs. Jill O’Donnell-Tormey, Cornelis Melief, and Ellen Puré.

Arthur Brodsky, PhD:

That’s great to hear. I know we here at CRI feel the same. Looking forward to seeing you in Atlanta for AACR next week!

Cornelis J. M. Melief, MD, PhD:

Yes, same to you. Nice talking to you.